1.- EVALUATING THE USE OF L1 IN THE ENGLISH LANGUAGE CLASSROOM
By Richard Miles
The issue this paper is going to examine in more detail is whether or not the use of students’ L1 in the classroom by either, the teacher, the students, or both, hinders the learning of a second language (in this case English) or can facilitate it. Currently there is a loosely enforced English-only policy at the institution where this research was carried out and recently there has been debate as to its validity. This project in essence then is a form of action research where the findings could have a direct impact on this particular school and the way lessons are taught.
While there have been many theoretical arguments both for and against the use of L1 in the L2 classroom, (which will be looked at in more detail in Chapters 2.2 and 2.3) there has been little research carried out which has measured the exact effects of L1 use in the classroom. Perhaps the difficult nature of measuring and gathering evidence in an attempt to answer such a difficult question is the reason behind this. Not only must a valid and reliable way of measuring and assessing student learning be established, but at least two languages must be used correctly and clearly in the classroom as well.
How to measure the effects of using L1 on learning, poses a difficult and complicated question. One obvious way to measure L1’s effects on learning is by trying to control all the other variables, and then measuring the improvement of the respective students. If student learning can be attributed directly to the use of L1 then a strong case can be made in favor of using L1 in the language classroom. This is essentially the goal of this paper.
As in any research field, terminology can often
confuse and obscure the real issue. ‘Mother tongue’, ‘first language’ and
‘native language/tongue’, are essentially all the same though it is possible to
argue that there are instances when they mean different things. For example; an
Inuit person living in
Due to the specific nature of the subjects in this research experiment (all of them are Japanese) the aforementioned terms will be used interchangeably. None of the students speaks another language, and all consider Japanese as their first language, native language and mother tongue.
Proponents of an English-only policy will collectively be known as the Monolingual Approach. Those advocating the use of L1 in the classroom will be known as the Bilingual Approach. It is recognized of course that this may be oversimplifying, but for the sake of convenience, these terms will be used as they are, in this paper.
The primary goal of this paper then is to find evidence to support the theory that L1 can facilitate the learning of an L2, at least in this particular situation and to demonstrate that the use of L1 in the classroom does not hinder learning. Such a conclusion would reinforce the researcher’s own personal bias, which has come about through his own teaching experience.
A Historical View of the Issue
A look at the history of L1 use in the L2 classroom quickly reveals periodic but regular changes in how it is viewed (Auerbach, 1999, p12). Several hundred years ago bilingual teaching was the ‘norm’, with students learning through translation. The use of L1 to study L2 was almost universal and readily accepted, in part because language teaching placed an emphasis on the written word above the spoken word. In the 19th Century, this trend slowly reversed itself (towards a monolingual approach), in part due to a shift towards an emphasis on the spoken word. The impact of mass migration, colonialism and a large increase in research in the field, would further strengthen the Monolingual Approach in the 20th Century.
The mass migration of peoples to other countries, particularly from Europe to America was important because it forced educators to refocus their lessons, from smaller translation-oriented classes to bigger classes, and perhaps more importantly, from students with a common L1 to students with a mixed L1 (Hawks, 2001, p47). No longer could teachers rely on using L1 to help them. The only way to teach was to use the L2 as the medium of teaching.
Experiences garnered by the many teachers who went
abroad during the colonial teaching period would further help the monolingual
tenet to evolve (Phillipson, 1992, p186), as would British colonial and
neo-colonial policies (Hawks, 2001, p47). As English became the predominant
culture in the British colonies, those who were not a part of it, were forced
to assimilate if they wished to better their life or be a part of the ruling
elite. Those moving to
The rise of an English-only classroom for political and practical reasons (of the teacher’s, not the student’s) conversely brought about the exclusion of the student’s L1. Those caught using L1 were often punished or shamed for doing something wrong (Phillipson, 1992, p187). The idea of bilingual education was seen as unnatural or inefficient (Pennycook, 1994, p136). Perhaps furthering the desirability of an English-only policy was the fact that many teachers themselves were monolingual. They could not, nor did they perceive the need to speak the L1 of their students (Phillipson, 1992, p188). By enforcing an English-only policy, the teacher could assume control of the class, and would naturally be in a position of strength. On the other hand, by using L1 in the classroom, the teacher risked undermining him/herself, as the students being the better speakers, would control the communication.
The emphasis on monolingual teaching of English also inherently implied that the native speaker was the ideal teacher. This was closely tied not only to political agendas, but also to the economics of the global EFL field (Pennycook, 1994, p176). English speakers could control all the employment opportunities, by being seen as the ‘ideal teacher’.
The appearance of the Direct Method of teaching just over a hundred years ago also contributed greatly to the consolidation of the idea that all L1 languages should be excluded from the classroom (Harbord, 1992, p350) & (Pennycook, 1994, p169). The premise of the Direct Method was that second language learning mirrored first language acquisition: lots of oral interaction, little grammatical analysis and no translation. The Direct Method would soon be discredited when it failed in the public education system (Brown, 1994, p44), but it would have a lasting influence on ESL/EFL classrooms.
Also, pivotal in forwarding the argument that L1 should not be used in the classroom, was Krashen, who advocated maximum exposure to the target language. He stated that all the lesson or as much as possible should be in L2 (English in our case), and that there was a definite relationship between comprehensible input in L2 and proficiency (Krashen, 1985, p14). Crucially though, this perhaps implied that time spent using L1 would only detract from learning. He even suggested that the reason exposure was not
always successful in facilitating proficiency, was because learners had access to their L1 either in class, or out of it (Krashen, 1985, p14). Others such as Gatenby 1950 (in Phillipson, p185, 1992) agreed, by claiming that the language being studied should be the mode of communication in the lesson. This idea that the L2 lesson should be taught in L2, in order to maximize exposure, and thereby learning, is perhaps the key concept which monolingual supporters have based their approach on.
The Makere report in 1961 further reinforced the idea of using nothing but English in the classroom. There are five basic tenets originating from this report, which have been called into question, but which were taken as the ‘truth’, at the time. They are:
· That English should be taught in a monolingual classroom.
· The ideal teacher should be a native English speaker.
· The earlier English is taught the better.
· The more English used in the classroom during lessons, the better.
· If other languages are used, English standards will drop (Phillipson, 1992, p185).
Phillipson has described these as the ‘five fallacies’ of modern English language teaching (Phillipson, 1992, p185) but the implications of these tenets are far-reaching and their influence can be found almost everywhere English is taught, even today. For the purpose of this paper, the first tenet is obviously the most important one. Nevertheless, it becomes difficult to separate them, as they are all interrelated and when combined they strongly proclaim an English-only policy in the classroom. Tenet 1 proclaims English-only is what should be striven for and conversely Tenet 5 claims that the use of L1 will hurt learning. The more English used the better (Tenet 4) also directly implies the less L1 use the better. The native teacher tenet (Tenet 2) also implies native English speaking teachers are more valued than non native English teachers, further emphasizing the superiority of English and conversely, the inferiority of the student’s L1. Tenet 3 implies that it is better to learn an L2 when you are younger, through direct exposure to the language, rather than when you are older and can utilize your L1 knowledge directly to help in learning L2.
By the 1970s these five tenets would be incorporated into the Communicative Approach, which quickly came to dominate language teaching. Native English teachers teaching only in English and excluding the students L1 would become the goal for many
Communicative supporters. As a whole, the Communicative Approach firmly believed the idea that monolingual teaching with authentic communication in L2 was the best way to learn a language (Pennycook, 1994, p169). Many linguists insisted that the target language be used for all purposes in the classroom (Wringe, 1989, p9) even when the reasons for using it remained unclear (Hawks, 2001, p47). Communicative researchers not only believed in the use of L2 as the medium of teaching, but many others also believed that L1 use actually interfered with L2 learning and brought about ‘error transference’ (Pacek, 2003), thereby hindering learning. These errors from L1 interference would be formed into what is now known as the Contrastive Analysis Hypothesis (Brown, 1994, p193). It was thought that errors in L2 learning could be predicted by comparing and contrasting L2 with L1.
Some researchers claimed that the learning of an L2 followed the same principles as the learning of an L1 (Phillipson, 1992, p191), which further reinforced the idea of using only the target language to maximize exposure and consequently learning.
Even as recently as the 1990s, the English-only movement has been further solidified by the various versions of the national curriculum orders in the UK, which established the use of the target language (TL) as the means of communication in the classroom (Pachler & Field, 2001, p84).
Recently though support for an English-only policy has
been declining, and some researchers and teachers have begun to advocate a more
bilingual approach to teaching, which would incorporate the students’ L1 as a
learning tool. Others have even gone as far as saying the use of L1 in the
classroom is necessary (Schweers, 1999, p6). Countries, such as
Many researchers now believe that the search for a ‘best method’ is a futile effort (Lewis, 1993, p189), because there can never be one method that suits all (Nunan 1999, in Pracek, 2003). Many methods have their place, and many techniques have their place, depending on the differing circumstances of the teaching environment. By excluding the students’ L1, we are severely limiting the number of methods and techniques available to teachers.
After looking at the relative merits of the Monolingual versus Bilingual approach to teaching L2, this paper will attempt to answer a simple but significant question: Can the use of L1 in an L2 classroom facilitate learning?
This experiment was conducted at the University of Kent, in England. More specifically, at Chaucer College, which is situated within the university, but is for first-year Japanese students only. All the students enter university in Tokyo, Japan, but spend their first year studying English in England, regardless of whether their major is English or a different subject such as Business, Science or Computing. All the students are aged 18-19 years old. They all speak Japanese as their first language, and almost all have studied English for 6 years at secondary school. All the students in our experiment were male.
Upon arrival in England they are tested. The test used every year is the KET test (the Cambridge Key English Test). This test is the first level of Cambridge exams in English for speakers of other languages (ESOL). Only the reading, writing and listening sections were used (the speaking section was replaced by our own speaking test). These scores determine which class they are placed in, ranging from level 1 to level 10, with 1 being the highest and 10 being the lowest level. The students stay for one year, living on campus and study from Monday to Friday, from 9:15-4:00pm. There are 7-12 students per class and in addition to studying English they also study business, history, geography, current issues, developmental studies and computing (all in English).
At the start of their year it would be fair to assess the levels of the students as ranging from false beginners to intermediate (our experiment deals with the former). Even the highest-level students would not be able to attend a regular English university, and the lower levels, can barely communicate or function in English. At the end of their year though, a few top students apply to the University of Kent, while the rest return home, to continue their studies.
The teachers are all very experienced teachers and all native speakers. Some have taught at the school for 10 years, while a few have joined recently. Some of them have also been to Japan, and can speak Japanese. While the official school policy is English-only, some teachers who are able to speak Japanese, and who are teaching low-level students, use Japanese in class, to varying extents. All the teachers are qualified teachers and many posses at least a graduate degree in either education, EFL/ESL teaching or in another field.
As the students live together, they unfortunately spend a lot of time speaking Japanese, and have little or no exposure to English outside the classroom, apart from simple shopping, commuting and eating out.
The First Experiment
There are actually two different experiments in this study. The first one will compare the progress of three classes, as they study English. They are the three bottom classes in the school, from here on in known as MG8, MG9 and MG10 (with MG8 being the highest and MG10 being the lowest class). They all studied English for the same amount of time, and used the same textbooks in England. On the first placement test (KET test), all their scores were very low but showed some slight differences. After five months, they were given a similar test (KET test), and their progress analysed and compared.
The main difference between the classes in terms of how they studied was that MG9's teacher could speak Japanese and did so periodically in class. He used Japanese to help explain things when they were not clear, to quickly clarify new vocabulary and grammar, and sometimes to explain instructions. While English was the predominant language used in the classroom, he described it as necessary to incorporate Japanese as a tool for teaching and learning.
The teachers in the other two classes were unable to speak Japanese. Of course this did not preclude the students from using it in class, but it did mean that Japanese was largely excluded from the teaching process. In MG8 Japanese was not allowed at anytime, even in student-to-student exchanges. Naturally this was difficult to rigidly enforce, but it was explained to the students that it was in their best interest to attempt to speak in L2 only and that this would lead to quicker improvement and learning on their part. The teacher teaching this class firmly believed in the Monolingual Approach so he enforced the rule as strictly as possible. There was a little resentment at first, but afterwards there was almost complete acceptance. Naturally there were a few occasions when students would lapse into L1. In MG10, however, Japanese spoken by the students was permitted to an extent although the teacher was unable to speak Japanese. Unfortunately the teacher was unable to verify exactly what the Japanese was used for.
The first part of the experiment compared the progress of these three classes, to see what, if any effect, the use of L1 by the teacher and students had on overall learning.
The second experiment (see Chapter 4.0) is an inter-class experiment. Four lessons were taught to MG9: two with an English-only policy and two using Japanese when necessary.
Chapter 3.3: The Expected Findings
It was expected that while all classes would hopefully show an improvement, MG9 would show a significantly greater improvement on the second KET test. In the second experiment it was hoped that the students would exhibit learning/more learning in the lessons where Japanese was available for them. If both these claims were to be validated, then the researcher could claim to have found evidence for his theory that the use of L1 can facilitate the learning of an L2 and should not be excluded from the classroom.
As was mentioned before, the students all took a placement test a few days after they arrived at the college. The main purpose of this test was that it provided teachers and the school with a way to allocate students into classes. Coincidentally, it served as a good baseline for our research purposes. While it is recognized that this was not necessarily the best way to go about establishing a baseline, this test has proven to be a reliable way of testing students entering the college. The KET test of English tests the reading, writing and listening skills respectively, in different sections. A separate oral test was also given so that the students could demonstrate their speaking skills as well. The oral score was not used for placement, but was recorded nevertheless. The other three sections of the test were averaged to give a score out of 100, which is recorded in Table 3.4.1. The oral scores were scores out of 20 and were done by individual teachers using strict criteria of questions and prompts, for which students were assigned points. Points were given for having progressed to certain stages, without help, having been able to answer questions, and talk about the prescribed topics and demonstrate viable understanding, as well as using correct English.
Here are the results for the three respective classes:
As can be seen from the data above, all the scores were below 30%, and ranged all the way down to 9%. There was never more than a 2 percentage point difference between any of the students, but the averages for each class were significantly different. MG8 averaged 26.12%, while MG9 averaged 20.31% and MG10 averaged about 13.35%, for an average difference of about 6-7% between classes. All the classes were separated into groups of 12, except MG9 and MG10, where the line was arbitrarily drawn between 17% and 16%. The total number of students in MG9 was only 8, while the total number of students in MG10 was only 7. It should be noted that one student in MG10 returned home for personal reasons after a few weeks, so his score was not included as no follow up score could be obtained.
The Follow up
The follow up test was conducted about 5 months after the baseline. All students sat the same exam, at the same time. The test was once again a KET test, although a different version. The main purpose of the exam was that it served as a midterm exam for the students, but for the purpose of this experiment, it serves as a good way to check on the improvements made by the students.
The results can be seen in Table 3.5.1., below and are presented along with the results from
It is possible to see a big overall improvement in the scores of each class, although a few students actually scored lower the second time in the oral exams for MG8. One student scored the same in MG10 on the oral exam, but all the students improved in MG9, on all the exams.
The Oral exams offer the strongest support for our theory that L1 use can help students learn and improve. The overall scores for MG9 (for which L1 was used) more than doubled, as did the scores of MG10 (in which L1 was permitted), but the scores for MG8 (where L1 was not permitted), while they did improve, increased approximately 33.5% only. Surprisingly MG10 actually improved more than MG9, but there is a logical explanation. One student did very well in MG10, far above the rest, meaning he could be considered a statistical anomaly and perhaps skewed some of our data. If his score is not included, then MG10’s overall improvement is about 98%, compared with MG9’s, which is 111%. If the highest student in MG10 is included, their overall improvement rate is 112%. There is only a small difference in the improvement between the two classes in which L1 (Japanese) was permitted. In MG8, where it was not permitted, the improvement was substantially lower at 33.5%.
Overall, MG9 averaged the highest in the oral exams at 10.93, followed by MG10 at 10.14, and then MG8 at 9.95. In the original test, MG8 were at the top with an average score of 7.45, followed by MG9 with 5.18 and then lastly by MG10 at 4.78. The data here is also favourable to our claim.
It is also interesting to note that in MG8 the scores of two students actually declined, and one stayed the same. In MG10 one student’s score also stayed the same. All the students improved in MG9, although, the highest scoring student in the original test, showed the lowest improvement overall, with an increase of only 5.55%.
On the KET exams, the results are more mixed, and generally not supportive of our argument. Again, all the classes showed great improvement, and all the students scored higher in the second test. The levels of improvement ranged from 29.6% to 123% in MG8, 65% to 247% in MG9, and from 62% to a 722% improvement in MG10.
The percentages for improvement overall are as follows; MG8 improved about 95.5%,
MG9 improved about 130%, and MG10 improved 189.9%. While we were expecting the improvement of MG8 to be the lowest, we were surprised to find that the improvement in MG10 (where L1 is permitted but not used by the teacher) was higher than the improvement in MG9 (where L1 is permitted and used by the teacher). When looking at the total overall averages though, MG8 were still better than MG9 and MG10 were still at the bottom.
The Second Experiment
While some of the evidence in the first experiment appears to support our thesis, there are still some problems establishing causality (this will be discussed in Chapter 5.4). So, in order to further strengthen the claim that the use of L1 in the classroom can facilitate learning, it was necessary to carry out a second experiment. Only one class (MG9) was selected for the second part of our experiment. It was felt that this class was the best choice as the teacher could speak Japanese, thereby removing a lot of potential obstacles from the design of the experiment. It was also felt that MG9 would be best suited for the experiment as they were used to using Japanese in the classroom and were also used to having periods when they were not allowed to use it. Unlike other classes, the experiment could be carried out in a natural environment, without the students being aware of it, and without having to adjust the established learning conditions. By having one teacher and one class, it was felt that the conditions could be controlled better, and any findings that were found would be more reliable and valid.
The Background Context
The experiment was carried out at the same school as the first experiment, but focused on MG9 only. As was mentioned before, in this class, the teacher was able to speak Japanese (L1), and utilized it when he deemed necessary. The students were also allowed to speak Japanese when they needed to (usually when their level of English prohibited them from expressing what they wished to say or ask). Primarily Japanese was used to help learning (explaining grammar, clarifying queries, translating new words, clearing up confusion), but inevitably it was sometimes used for casual daily conversation. However, there were several hours of lessons a week in which Japanese was prohibited, in an effort to challenge the students to speak more English. So, it was relatively easy to conduct the experiment within the normal confines of this classroom.
It was decided that in the first week a lesson would be conducted as usual, utilizing L1 when needed. The following week, another lesson was taught to them, without L1 usage being available. At the end of both weeks, the students were tested to see how much they had learned. The two-week cycle was then repeated in reverse order with two new lessons, so as to try and establish a measure of reliability if any positive (or negative) findings were found.
The Design of the Second Experiment
Due to logistical limitations, it was not possible to carry out an experiment that was as comprehensive as we would have liked. Ideally it would have been better if we could have separated the class into two groups, and used one as a control group, and one as an experimental group. Instead we had to use the same group and vary the lessons to make a comparison.
An attempt was made to account for as many variables as possible. All four lessons were carried out with the same group of students, and the same teacher. All the lessons were deemed to have been of a similar level of difficulty, and the same time was spent on each lesson in class. The tests and lessons administered were essentially the same as would have occurred in any other given week, during the course, so as to prevent any surprises and to prevent unfamiliarity from affecting the results obtained. The students were not told of the experiment because it was thought that prior knowledge of this information could influence the results. After the results were collected, the students were debriefed, and their consent was obtained.
As the lessons were part of a well-established routine, the students were aware of when they could use L1, and when they could not, so this was not a problem. In the lessons where Japanese was used, it was thought that as per usual, Japanese spoken by the students would be allowed when it was needed. Consequently, it was the students who determined when L1 was used and for what purpose it was used. However, all students abided by a prior class rule, that they were to use English unless it was not possible, in which case Japanese was allowed. The teacher would usually not initiate anything in Japanese, but would use Japanese if the students prompted him. His Japanese was recorded so that it could be checked and verified later by a native Japanese speaker to ensure that any Japanese he used, was used correctly and did not mislead students.
Each lesson had similar components, such as the initial learning of new vocabulary, which was necessary for mastering the key concepts in the main part of the lesson. After the meaning of the vocabulary was established, the vocabulary was demonstrated by the teacher and then manipulated by the students, with the teacher providing the appropriate feedback. When L1 was permitted, students would often translate the new word into Japanese, and check to see which part of speech it was. Then they would practice using it, helped by some demonstrations from the teacher. A class set of cards was also made, so that the information could be readily accessed by anyone who needed it. The same pattern was followed in the English-only lessons, except that they were not allowed to translate the word into L1, but had to define the new vocabulary in words they already knew (L2 only). The teacher obviously assisted with this, by providing plenty of examples, so that they could deduce the meaning of the word from the context of its usage.
In the next part of the lesson, examples of the key structure for the week were examined, explained and manipulated. Students were then required to perform certain tasks, which helped them learn the key concepts further. After the point had been understood to the teacher’s satisfaction, various tasks were given out to the students, where they had to manipulate and use the new structures and grammar, to communicate with each other, negotiate and exchange information. Efforts were made to insure as realistic activities as possible.
At the end of the week, students were required to do a test where they could demonstrate they knew the new vocabulary, could use it correctly, and could also demonstrate that they understood and could use the key concepts of the lesson for that week.
In order to provide evidence of learning, a baseline test was given to the class at the beginning of each week. The baseline tests were similar to the final tests the students were given, but obviously different in terms of the specific content, which was varied. A comparison of these two tests and the differences between them would provide the data showing how much learning had taken place.
The content of each lesson was based on four consecutive units in the course book. The titles of the lessons reflected the key grammar components of each lesson. The lessons were as follows:
· Superlatives (L1 use permitted)
· Past habits / used to (English only)
· Conjunctions (English only)
· Time clauses (L1 use permitted)
It should be noted that the method of teaching in these classes, could be criticized for not incorporating the latest findings in SLA (Second Language Acquisition) and although reflective learning and consciousness raising were encouraged, a lot of the lesson could be said to have followed PPP (presentation, practice, and production). However, it was felt that as this was the standard method for teaching in this school, and the students were familiar with it, it would be counterproductive to attempt anything radically new just for the sake of these experiments. The lessons in the experiment were more representative of the type of lessons that generally occur at this school, so if significant findings were found, they would be more meaningful.
Here are the results for the first lesson on ‘Superlatives’ (L1 permitted)
From the results we can see that the students were best prepared for Lesson 3, as they scored 61.83% on the baseline. They also did well on Lesson 1 in the baseline, by scoring 60.67%, but did far worse on Lessons 4, where they scored 41.1%, and on Lesson 2, where they scored 32.46%
After a week’s worth of lessons, they did considerably better overall. Lesson 1 showed the highest score at week’s end, with 89.18%, followed by Lesson 3 with an average score of 67.25%. The two lowest scoring lessons in the baseline showed significant improvement and almost surpassed the average of Lesson 3. Lesson 2 averaged 66.38%, and Lesson 4 averaged 64.1%.
In terms of improvement, Lessons 2 and 4 showed the greatest amount of statistical improvement, no doubt because their initial scores were so low. In Lesson 2 the average score increased by 104%. In Lesson 4, the average score increased by about 56%. Lessons 1 and 4 also exhibited increases in the average score, but as their initial scores were higher, the increase was relatively lower. In Lesson 1, the average score increased about 47%, whereas in Lesson 3 the average score increased about only 9%.
With the completion of the research and the collection of the data, we can now turn our attention to analyzing the results. Our analysis will focus on answering two questions. The first is whether the use of L1 in the classroom hindered learning. The second is whether the use of L1 in the classroom helped students more than the prohibition of it. Overall it would be fair to categorize our conclusions as positive, in terms of validating our research thesis. Results indicate that the use of L1 in the classroom does not seem to hinder the learning of an L2, and in fact seems to facilitate it in some situations. Results also show that English-only can help learning and lead to improvements. However, upon closer examination there seems to be evidence that the use of L1 in the classroom does actually help learning and lead to more improvement among the students than English-only. It should be noted though, that there were some problems with this study, which may have influenced the results.
Analysis of Experiment 1
The results from the first experiment suggest that the use of L1 in the language classroom does not hinder learning. All the students in MG9 improved considerably in all aspects of their L2 as evidenced by the KET test and the oral exam. Some improved more than others, but the fact that they all improved (quite considerably in most cases) would seem to contradict the notion that using some L1 in the classroom hinders the learning of an L2. Of course it could be argued that the students in MG8 and MG10 also improved, without the use of L1 by their teachers. This is not completely unexpected though, as the students studied intensely for five months, and were living in an environment where they needed L2.
The key to the results in Experiment 1 is that while all the classes improved, MG9 improved more than MG8 who were a higher class (based on the initial tests). This improvement is most evident in the oral section. Not only did MG9 (where L1 was used) improve more than the other two classes, but they also scored higher overall than the other two classes. These results are probably the strongest evidence we found in Experiment 1, to support our original thesis. The use of L1 in MG9 certainly did not hinder their progress, and actually appears to have helped it.
The reason for this might lie less in a direct relationship between learning and L1 use,
but more in a sense of confidence that developed between the students and the teacher. A quick survey among other, non language teachers, who taught all three classes, found that they were almost unanimous in the opinion that MG9 exhibited far greater confidence when speaking, compared to the other two classes. The students themselves agreed and many said that perhaps the biggest reason for their apparent confidence was the relaxed atmosphere in the classroom. They knew that if they made a mistake or got stuck and couldn’t express themselves, they could always fall back on L1, knowing the teacher could understand easily and would help them. Consequently, they also didn’t worry about making mistakes as much, and considered expressing themselves the primary objective of speaking. Interestingly enough, students in MG8 seemed to agree that they became hesitant when speaking in L2, knowing they were fairly low-level, and that if they were not understood or got into difficulty, they could not get help via L1. The students in MG9 reported that by sometimes conversing with the teacher in L1 (especially near the beginning of the course), they felt they could get to know the teacher better, whereby they could relax more, and feel more confident when attempting to speak in L2, knowing the teacher and trusting he would support them when necessary.
It would seem likely then that although the use of L1 in the classroom did not perhaps directly contribute to the superior improvement in speaking of MG9, it did serve a purpose, which led to the aforementioned improvement. Conversely, the prohibition of L1 in the classroom, lead to lasting insecurities among the students of MG8, which contributed to their slow improvement with regards to speaking.
In terms of the other test, which was administered (the KET test), the results are less supportive of our position. MG9 did not exhibit any substantially greater improvement than either MG8 or MG10. However, they still exhibited a very good improvement rate, which further diminishes the notion that L1 use can hinder development. Clearly learning was not hindered in this case, by the use of L1 in the classroom.
The results of Experiment 1 show that not only does the use of L1 not hinder the development and improvement of an L2, it can actually facilitate it, by installing a sense of confidence among the students, based on a trust in the teacher to support them when needed.
Analysis of Experiment 2
The results for Experiment 2 could be described in the same way as Experiment 1. Overall it would be fair to say that they were fairly supportive of our original thesis. In all four lessons, the average score for the class improved (possibly as a result of testing so soon after having taught the structures). Only in a couple of instances, did a student’s individual score decrease from the baseline. Regardless of whether the lesson was English-only or a bilingual lesson, the students as a whole improved, further supporting our initial theory that the use of L1 does not hinder the learning of an L2.
Upon further examination it can be seen that the students seemed to have a good grasp of two of the lessons, before they were taught, both averaging about 60% on the baseline test. The other two were obviously more difficult as the students averaged about 32% and 41% on the baseline test. Fortunately for the researcher this unexpected finding provided a unique opportunity. By coincidence rather than design, one of the more difficult lessons was English-only, and one wasn’t. The same is true of the two easier lessons. This gives us the additional chance of being able to further compare the differences between English-only lessons and lessons where L1 is used.
Firstly, the results from the easier lessons will be analysed. Lesson 1 was on Superlatives, and the students averaged 60.67% on the baseline test. Compare that to Lesson 3 on Conjunctions, where the students averaged fractionally higher, with 61.83%. The difference between the baseline tests was insignificant. In Lesson 1, L1 was used/permitted during the course of the week, whereas in Lesson 3 it was not. At the end of each week final tests were administered and the results were quite different. In Lesson 1, where L1 was used, the student average improved from 60.67% to 89.18%, but in Lesson 3 where L1 was prohibited the student average only increased from 61.83% to 67.25%. These figures indicate possible support for our argument that L1 use in the classroom can facilitate the learning of an L2.
It is also interesting to look at the figures of some individual students. In Lesson 1 the weakest student (Student F) improved from 39.47% on the baseline to 85.5% on the final test. His score more than doubled. In fact every student in the class showed an increase of at least 13%, or more. Compare this with Lesson 3, and we can see that while some students did improve, Student G showed no improvement, and Students A, B, F and H actually scored lower on the second test.
Initially these figures also seem to support our thesis claim. Unfortunately there are some problems with these figures with regards to comparing different lessons (see Chapter 5.4). There is also the possibility that the second test in Lesson 3 was substantially more difficult than the baseline test, thereby causing the scores of several students to drop, and not properly reflecting the true learning that had occurred.
We can also compare Lessons 2 and 4 for further data regarding our thesis. Lesson 2 was on habits/ used to, and students found this difficult, as evidenced by their very low baseline scores; 32.46%. Lesson 4 was on time clauses, and the students faired marginally better, by averaging 41.1% on the baseline. In Lesson 2 L1 was not permitted but this did not prevent the students from improving greatly over the course of the week, and scoring 66.38% on the final test. All the students improved their scores, although two students still did fairly poorly on the final test, showing that while they had learned something, they had not fully grasped the full concepts of the lesson.
In Lesson 4, L1 was permitted and used by the students and the teacher alike. The average score was only 41.1% on the baseline, showing the students had a great deal of trouble with the use of time clauses. After a week of lessons in which L1 was permitted, they showed considerable improvement, as in other lessons. They averaged 64.1%. The results of this lesson compared with the results of Lesson 2 provide some rather interesting findings. It is often assumed that the use of the student’s L1 in the classroom is most appropriate in low-level classes, where there is little or no understanding. In lessons 2 and 4, our low level class exhibited low understanding of the lesson material in the baseline test, particularly in Lesson 2. It was hypothesized that as they seemed to have more of a grasp of Lesson 4, and that they could use L1 in the lesson to help them understand, they would naturally improve more significantly, in Lesson 4. The results show otherwise, and this is a serious blow to our research thesis. The students did improve considerably in Lesson 4, further providing support for our argument that L1 use does not hinder learning. However, the second part of our thesis is not supported by the comparisons between Lessons 2 and 4. In Lesson 2 where L1 was not permitted, the students improved substantially more than they did in Lesson 4 where L1 was permitted. Even though they scored lower in Lesson 2 on the baseline test, they actually scored higher on the final test, than they did in Lesson 4. This could suggest that perhaps Lesson 2 was easier to understand, or that possibly L1 use is not very important when trying to understand a difficult lesson.
If we look at the results for Experiment 2 together though, they are still favourable in terms of supporting our thesis. In both the lessons where L1 was used, the students all improved considerably, compared to only one of the two lessons where L1 was prohibited. The use of L1 clearly did not hinder learning and a strong argument could be made that it actually helped learning.
An Evaluation of the Experiments
While there were problems with this study, its design worked fairly well, for the most part. Because of the circumstances in which it was carried out, it was not possible to design the experiments exactly as we would have liked. For example, it would have been better had we been able to separate MG9 into a control group and an experimental group. A retest a few weeks later would also have been beneficial.
Still there, were a lot of strengths regarding the design of these experiments. The homogenous group of subjects meant that many possible variables were controlled. The similarity in level of the three groups in the first experiment also meant that comparisons could be made fairly easily, as did the fact that the same textbooks were used in each class, the same amount of time was spent on lessons, and the same tests were taken at the beginning and end of the experiment stage.
As with most research studies though, there were problems, which have to a certain extent, undermined the findings discussed in Chapters 5.1 and 5.2. In Chapter 5.4 these weaknesses will be discussed and presented along with suggestions on how to improve this kind of experiment in Chapter 5.5.
Problems with this Study
There are two major weaknesses with this study. They are: the failure to establish causality in Experiment 1, and the comparison of possibly incomparable statistics in Experiment 2.
Causality is obviously vital for any research study. Even with favourable results, such as in Experiment 1, the inability to clearly demonstrate causality hurts any claim that can be made.
One reason why causality was weakened was the possible presence of other variables. While many variables such as time, lesson content, student ability, nationality, and tests were accounted for, a few variables were not, such as: teaching methods, teacher personalities or even student factors. All of these could have skewed the results in Experiment 1. For example: it was widely thought among the teaching staff that MG9 were on the whole, a more outgoing and talkative class. Perhaps this is why they scored higher on the oral tests, but not in the KET tests. It is also possible that MG9 and MG10 improved more in comparison with MG8, because they were smaller classes, thereby allowing each student more talking time in class. Because variables such as these cannot be ruled out, it is more difficult to argue that the use of L1 in the learning process was the factor or cause, which contributed to MG9’s higher scores in the oral exams.
The second major weakness with this study is the reliance on statistics as evidence of learning. Tests and test scores are of course not always the best indicators of learning. There is also the problem with interpreting statistics. Comparing the improvement of classes based on percentages can be problematic, and in Experiment 2 it is possible to argue that the figures obtained are not necessarily indicative of learning and possibly not comparable. For instance, it is difficult to say that 1 percentage point on one test/lesson equals a percentage point on a completely different test/lesson.
Also damaging to our theory is that MG10 improved more than MG9 on the KET test. Although they were expected to improve more than MG8, the fact that they outperformed MG9 was surprising. It is possible to argue that they were also allowed to use Japanese in the classroom, and that this actually supports our theory, but of course the teacher did not use Japanese, so this actually undermines our theory, and perhaps gives support to the idea that L1 does not need to be used in the classroom.
In order to attain more convincing findings than the ones collected in this paper, there are several areas, which need improving. The most obvious would be to create a control group. This would allow the same lessons taught by the same teachers, to similar students, to be evaluated. By controlling almost every variable and only altering whether the group is able to use Japanese or not, causality could be attained. This would also allow the experimenter to compare near identical lessons, instead of similar lessons, which are potentially problematic. Naturally it would be important to make sure that the experimenting did not come at the expense of student’s education.
To further strengthen claims, it would have been better to expand this experiment to include not only low-level students, and also to go beyond using only Japanese students. It is quite possible that nationality or student ability could have played a part in the findings. With increased subject numbers and a more varied subject pool, it would be more possible to generalize findings.
Another area that we would have liked to investigate more had we had the resources and time would have been to try and ascertain in what specific situations L1 can help, within a lesson.
The results of this experiment have partially confirmed the original thesis. Due to a personal bias, and an increase in recent research supporting this bias, we investigated the use of L1 in the English language classroom. Before the experiment it was hypothesized that L1 use in the classroom does not hinder learning, as many have claimed, but that it actually helps learning. The purpose of this research was to try and prove these two theories and find evidence to validate our claims, so that it could be put into use at the college where this research was carried out. Indeed, as a result of this research serious consideration has been given to the school’s English-only policy. Early indications are that it will be abolished.
The experiment actually involved two different experiments. In the first, three low-level first-year university classes were compared. One class did not permit the use of L1 in the classroom, another did permit it, and the third actually utilized L1. The classes were similar in many respects and this enabled us to compare their progress over a five-month period. The results showed that in the class where L1 was utilized, the students showed a significantly higher improvement in the area of speaking. The reason suggested here is that confidence was the determining factor, and that L1 use helped to foster this confidence.
In the second part of our experiment, one class was focused on. Four separate lessons were taught to this class, two utilizing L1 and two which did not utilize L1. Results were mixed in this experiment. The first part of our thesis was favourably supported, as the classes utilizing L1 improved significantly; thereby showing that L1 use had not hindered learning. There was mixed evidence regarding the second part of our thesis though. In one comparison, the class using L1 outperformed the one, which did not use it. However in another comparison, the reverse was true, casting some doubt on the validity of the first findings. The mixed results were perhaps due to the design of the experiment, which relied on comparing different lessons, and assuming they were equal. If these problems could be addressed next time, more valid and reliable findings could be obtained.
Overall though, the findings in this experiment could be classified as positive. While there were problems, the findings were generally favourable and supportive of our original thesis, that L1 use in the English classroom does not hinder the learning of an
L2, and can actually facilitate it.