University of Botswana, Botswana
Received date 29/02/2016; Accepted date 09/06/2016; Published date 11/06/2016
Visit for more related articles at Research & Reviews: Journal of Educational Studies
This article discusses code-switching as used as a strategy for teaching and learning in selected Botswana primary schools. Codeswitching from Setswana (Botswana’s national language), to English (official language) and vice versa, was observed in both rural and urban primary school classrooms where the language of instruction is supposed to be English in Standard Two and subsequent primary school levels. Through the use of the qualitative approach, this study investigated how the language-in-education-policy is implemented in primary schools. Teachers were the key participants observed at different times and teaching different lessons. Data were collected using open ended questionnaires, classroom observations and interviews. The findings indicated that code-switching from Setswana to English or English to Setswana in this study was used where it was not necessary with ethnic minority learners who did not understand Setswana and where most learners were proficient and competent in English. The findings are significant in that they could help the teachers to reflect on the code switching practices they do unnecessarily and see if they benefit the learners or not and probably refrain from such. The conclusion was that code-switching becomes a problem when teachers
Code-switching, Languages of instruction, Rural, Urban, Language-in-education policy, Language planning.
The goals of a language-in-education policy are often linked to unity and identity with the suppression or failure to recognize and support other linguistic varieties thus leading to refusal to grant a voice to the ethno cultural groups . Some policies prescribe the use of a national language in heterogeneous classrooms where learners speak a variety of first languages. This may not be practical and realistic. This situation may frustrate learners instead of facilitating their learning if the national language is not the learners’ first language. The use of English as an official language and medium of instruction has affected the outcomes of language-in-education policies and also led to failure and unsuccessful implementation of the language-in-education policies in some African polities and other states where English is a learnt as a foreign or second language [2-4].
In some cases when speakers address their deficiency in the target language they code-switch to their first language, a situation, that can occur on different circumstances. Scholars have researched on code-switching and why it is used in various classroom contexts. For example, Modupeola , points out that code-switching in the classrooms is done as a communication phenomenon, providing students with opportunities to speak and understand concepts, and facilitating flow of the classroom instruction, to start from the known to the unknown, clarifying and stressing meaning, and stress the importance of content. Isaac  sees code-switching as a necessity in the classroom, an aiding instruction and expressing pertinent issues and therefore perceives it as vital for enhancing learning in the classrooms. Lugoloobi-Nalunga , shares the same views as the latter that code-switching in a multilingual classroom in Sweden was done for clarification and emphasis and helped immigrants students to understand instruction much better as some immigrant learners were not well versed in English as one could think. Thompson looks at code-switching from the societal perspective that the reasons why code-switching occurs is because people want to say something in secret. For instance, they want to fit in a conversation and to convey their thoughts. Hence, Gal sees code-switching as performing a socio-linguistic function.
Adendorff  studied English-isiZulu code switching among isiZulu-speaking teachers and their learners. He reported that code switching from English to isiZulu during an English lesson was used by the teacher for academic reasons and to maintain “social relationships in the classroom”. Azlan and Nazasuman  point out that code-switching was done to enhance solidarity in the first language and also to convey ideas in specific situations. All these situations of code-switching were done for a good purpose, to promote and enhance learning in the classrooms and these scholars share the same sentiments as those already discussed above. However, Cook argues that code switching in classes which does not share the same mother tongue may create problems as some learners will feel left out in the teaching process, thus creating psychological problems and therefore defeating the whole purpose of learning.
Code-switching refers “to the mixing, by bilinguals (or multilinguals) of two or more languages in discourse, often with no change of interlocutor or topic...its occurrence occurs within the confines of a single sentence, constituent or even words”. Codeswitching is common in African states where speakers are exposed to a number of languages and a number of linguistic variations. The speakers use it for different purposes such as negotiating, challenging, or changing different conversational situations and even emphasizing a point. Also, code-switching can be used to indicate lack of competence and fluency in the other language used. However, code-switching is not only common in communities and or amongst speakers; it is a common phenomenon in the classrooms of different levels [8,10-12]. This paper argues that the code-switching observed from Botswana primary schools studied presented problems and therefore could not be used as a teaching strategy as it was intended by teachers. The researcher arrives at this conclusion because code-switching was done in classrooms where the teachers and learners did not have a common language in rural primary schools. In addition, code-switching was done where it was unnecessary in urban primary schools in classrooms where learners come from preschool already speaking English and in the presence of foreigners and immigrants. Both situations impede learning. Though not a new phenomenon in the global world, in Botswana context, it is weak strategy for teaching that can cause learners and teachers to be disgruntled. It needs a reflection and reconsideration in the context in which it is done since it leads to inaccessibility of information.
The significance of code-switching as a strategy or a problem discussed in this paper is that, it is a policy problem and therefore requires policy makers to critically reflect on the implementation of the language-in-education policy, assess its risks and challenges especially where the prescribed languages of instruction disadvantage learners. The theme also challenges the Ministry of Education and Skills Development to take into consideration the languages spoken by teachers before they post them to rural primary schools where learners speak different home languages from those of the teachers. Again, since teachers specialize in subjects during training, a policy that supports subject specialization in primary schools could be implemented. Further, if the teachers are doing the code-switching unconsciously they could also reflect on this practice and monitor if it is good or it does more harm. This reflection might help teachers to avoid the use of such a practice if it is not beneficial to the learners. Code-switching in both cases of primary schools studied has been found to be a problem.
Background to the study
Botswana is a landlocked country located in Southern Africa. It shares boarders with Zambia on the North, Namibia on the West, South Africa on the South and Zimbabwe on the North East. The country has a radius of 580 000 square kilometres with a population of 2 024 904 . Botswana is a multilingual and multicultural nation with 28 languages spoken in the country . Despite its linguistic diversity Botswana recognizes only two languages of instruction: Setswana which is the national language and English as a foreign language. Although researchers have unearthed the challenges of such a policy, the legacy of using two languages of instruction has been carried on before independence. For example, before independence there was no policy in place but English was given a high status. However, this changed after independence when it was argued that national language should be recognized for national unity and identity. In this regard, Setswana was used from Standard One to Three as a medium of instruction while English was taught as a subject and would take over as a medium of instruction in Standard Four. At this juncture teachers were not conversant in English hence there was a lot of code-switching in the classrooms. In fact some teachers taught subjects that were supposed to be taught in English in Setswana.
The first president of Botswana assigned the first commission to look into the education system. This National Commission in Education 1977, recommended changes in the policy and it put emphasis on giving Setswana a national status over other indigenous languages; hence the policy recommended that Setswana be used as a medium of instruction from Standard One to Four while English takes over at Standard Five and subsequent levels. Examinations were supposed to be tailored to suit the policy and hence learners were supposed to write examinations in Setswana language at Standard Four.
However, there was an argument that the policy delayed learners to start learning in English, the language of the examinations. This led to another commission which was set up to review the policy to match issues of globalization and meet the challenges of the twentieth century . The Revised National Policy on Education, 1993 recommended that Setswana be used as a medium of instruction only at Standard One and English takes over at Standard Two. These policies had negative implications. First, the policy designers overlooked the fact that Botswana is multilingual. Second, the policy designers disregarded the fact that some Batswana do not speak Setswana as a first language and therefore using it as a medium of instruction would present problems. Importantly, all policies have been silent on the use of indigenous languages in the education system. However, before independence, in the North East, Ikalanga was used as a medium of instruction and this was stopped after independence to foster unity and identity by using only Setswana (the national language). Today, indigenous languages have not been extended to lower primary school like it has happened in other countries such as Tanzania and Somali .
To investigate how the language-in-education policy was implemented, this study adopted the qualitative approach using the various data collection methods such as classroom observations, open ended questionnaires, interviews and field notes. The data collection methods were triangulated to allow for corroboration and also that divergent evidence could increase the trustworthiness of the data. The objectives of the study were to investigate how the language-in-education policy in Botswana primary schools is implemented in ethnically and linguistically complex classrooms; examine the implementation strategies used to implement the policy; establish the challenges encountered in the process of implementation and find out the solutions to the problems encountered.
This study was conducted in six districts of Botswana out of the nine main districts. The six districts included Ngamiland, Kgalagadi, Kweneng, Central, Southern District and North East. The distance between the districts was also an important consideration. The districts were far apart from each other. This was helpful in allowing data not to be influenced by their closeness. Again, the districts were chosen because of their linguistic diversity to see how a policy that recognizes only two languages of instruction is implemented in situations where learners do not speak the target languages.
In each district, one school was sampled randomly taking into account its linguistic and cultural diversity. In rural areas the aim was to choose schools where learners go to school speaking different home languages from the school languages. For comparison, in urban areas, the idea was to choose schools with learners who speak Setswana as a first language and those with immigrants and foreigners.
Key participants of this study were implementers of the policy such as teachers and school management. The reason for choosing teachers was to tap first-hand information on their classroom experiences during the implementation process. The school management are the internal and immediate supervisors of teachers and therefore they were in a position to provide information at supervisory level on the implementation of the policy and the strategies in place to implement it.
In each school, the levels targeted were Standards One, Two, Four and Seven. There were reasons for choosing the targeted levels. Standard One is where Setswana is used as a medium of instruction for the teaching and learning process while English is taught as a subject. Standard Two is where English as a foreign or second language is introduced as a medium of instruction and used in subsequent levels. Standard Four is where learners write their National Attainment Tests in English. Learners have to apply themselves to gauge whether they have mastered English for the past four school calendar years. Lastly, Standard Seven is a level where learners write their final Primary School Leaving Examinations using the medium of instruction English. Therefore, all these levels were important for this study to draw conclusions on the implementation of the policy at different levels of primary schooling.
Data were analyzed under each key research objective to make sure that all the objectives were answered. Themes and subthemes were developed from each key research question and interpreted and substantiated with teachers’ voices. This is where the concept of this paper, code-switching, was developed from the implementation strategies used by teachers in the classrooms.
In this section two types of primary schools’ results are presented: rural (schools A-D) and urban (schools E-F). Teachers reported that they code-switch from Setswana to English or vice versa during lessons for learners to understand. This was also evident during classroom observations where code switching from Setswana to English and English to Setswana was common in all the primary schools studied. Code-switching is supposed to provide a linguistic advantage rather than an obstruction to communication.
Rural primary schools
In the context of the rural primary schools studied (A-D) inter-sentential and intra sentential code-switching were observed. These two types of code-switching were done at clause or sentence boundaries or within clauses and sentences. In the classrooms, the researcher observed that code-switching from English to Setswana and vice versa was done in different occasions:
Firstly, code switching was done in an attempt to explain concepts and elaborate on examples; In a Standard Two classroom, the teacher would say, “compound words ke go tshwaraganya mafoko...” meaning that (compounds words meaning to compound two words to make one...)
Secondly, code-switching was done to reprimand learners who were not concentrating. For example, the teachers would say, “nnang fa fatshe, you are making noise... (sit down, you are making noise) hei… what are you doing there...? boelang kwa mannong a lona...(What are you doing there, go back to your seats) ... stop bullying others... tla kwano ke go betse... (stop bullying others, come over let me discipline/beat you up).
Thirdly, it was used to elicit answers from learners to break the silence in the classrooms. For example, teachers would say “...tlhe batho buang... (Please, say something)...are bueng... (Lets’ talk)...ee bua... (Yes, say something).” Important to note is that learners were not allowed to code switch when they responded.
Teachers gave reasons for their code switching. In school B, the teacher said, “I use Setswana to teach English because at term two learners are beginning to understand some Setswana words and therefore introducing English as a medium of instruction will only drive them away from school. I introduce it slowly because there are a lot of signs that they cannot follow the language of instructions, so why use a language the learners fail to understand?” True as the statement could sound, the learners did not understand neither languages used in code-switching. In most cases when the teacher uttered these phrases, learners did not respond accordingly.
In another English lesson in a rural primary school a teacher introduced a topic as such:
Code switched version (both English and Setswana)
“Our topic today is compound words. Can you say compound words? (Learners sing in chorus compound words). Compounds words ke go tshwaraganya mafoko a mabedi mme o le bitsa gape o le kwala jaaka o kare le lengwefela. Le a utlwa! Mafoko ao ke cattlepost, policeman, headman and blackboard jalojalo...”
“Our topic today is compound words. Can you say compound words? (Learners repeat in chorus compound words). Compound words means to compound two words together and pronounce them as one. Do you understand! Some of the examples are: cattle-post, policeman, headman and blackboard and so on...”
The code-switched version in the text above indicates that a lot of Setswana words were used to define compounds words which may interfere with learners understanding. Hence, where learners had to get the gist of the explanation, another language which the learners had difficulty in understanding was used. Learners had a problem understanding Setswana in Standard One classes and there was no indication that things would be different at Standard Two. Again, at Standard Two this was a difficult topic to deal with. The teacher did not clearly explain to the learners what compounds words are. The researcher believes that she only knew the examples of the compound words but could not clearly show the learners what is compounded. In this case, even drawings or pictures could have assisted in helping learners to understand. The problem could also be compounded by the fact that some learners may not have seen a policeman where they stay because they stay in cattle-posts where there are no policemen.
In one of the rural schools, a teacher was using Setswana medium to teach a subject which was supposed to be taught in English. One would mistake the English lesson for a Setswana one. When asked why this is happening this teacher said, “I have been teaching standard one class for twenty-five years where the medium of teaching has been Setswana and I am used to it.” This was also confirmed by the School Management that some of the teachers have been teaching the same level for many years where the medium of instruction is Setswana and therefore they are not proficient in English. The reason for teaching the same level for years was that the teacher knows how to handle children who are being initiated into the education system. An example is given below on a Standard Two science lesson she taught. This was the way part of the lesson was conducted:
The code-switched version – (mostly Setswana)
Re tsile go ithuta ka dikarolo tsa nonyane. Jaaka lo bona nonyane e na le dikarolo tse di farologanyeng jaaka tsa motho: molomo, diphuka, molala, tlhogo le menoto. Ntshupegetsang dikarolo tse mo nonyaneng e ke e beileng fa.” “...please stand up.”
“We are going to learn about parts of a bird. As you can see, a bird has different parts just like a human being such as: beak, wings, neck, head and feet. Can you show me the parts of a bird from the drawing that I put up here? ” “... Please stand up.”
After the above introduction there was silence in class. The possible reasons could be that, the learners did not understand both languages used. Secondly, this was a Standard Two lesson where the medium of instruction is English. This could be an indication of the teacher’s lack of proficiency and she imposes it on the learners.
Urban primary schools
In urban primary schools such as schools (E-F), the findings indicate that code-switching was done unnecessarily. Firstly, code switching was done even though most learners understood English from their preschool background where in most cases the medium of instruction is English. Secondly, there were foreigners and immigrants in the classrooms such that if teachers code-switched to Setswana they would not understand. The researcher must admit that this was a tricky situation because it was not necessary to code-switch. In school E, one of the research participant teachers argued that, “some of the concepts in their syllabus are difficult and therefore it is necessary to explain to the learners in Setswana.” For example in a Mathematics lesson, this is what the researcher observed:
The code-switched version (mostly Setswana)
Letshwao le ke mang le? Heelang, kare letshwao le ke mang? Le tlile go le dirisa jang lo sa le itse? Le bidiwa plus sign (+) kana go tlhakanya.
What is the name of this sign? I am asking you what the name of this sign is. How are you going to use it if you do not know it? It is called plus sign (+) or addition.
From the example above, the lesson was introduced in Setswana and some learners could only guess by looking at the plus sign that the teacher is introducing a topic in addition. This can only confuse learners even if they know the plus sign. From the instruction the teacher expected learners to understand the instruction in Setswana as she was asking them how they are going to use the plus sign if they do not know it. This could be a sign of lack of proficiency in English by the teacher. She had no reason to introduce Mathematics in Setswana when the medium of instruction is English and more so the learners are proficient in English.
It is important to note that some of the children in the urban schools speak English as their first language and to explain concepts in the Setswana will only impede their understanding. In one of the urban classes, one of the teacher participant argued that “yes I am teaching English but some of the topics are difficult especially the grammatical ones, I am not so good in teaching grammar.”
For example, this is how this teacher introduced parts of a speech in a Standard Four class:
The code-switched version (English and Setswana)
This is a difficult topic that we are going to learn today and you need to concentrate. Can you read after me:
Nouns – maina
Pronouns – maemedi
Verbs – madiri
Qualificatives – ditlhaodi.
Mphang dikai tsa dikarolo tse tsa puo... a re simololeng ka maina. (learners give some examples while others were passively listening)... A re tseyeng dibuka tsa rona and use the parts of speech in sentences... A re simololeng ka di – noun.
This is a difficult topic that we are going to learn today and you need to concentrate. Can you read after me: Nouns, Pronouns, Verbs and Qualificatives?
Give me examples of the different parts of speech... let us start with nouns (learners give examples while others were listening passively)... Take out your exercise books and use the parts of speech in sentences starting with the nouns.
The learners in this class were fluent in English because they have preschool background, parents talk to them in English; the class also includes foreigners and immigrants. It was not necessary for the teacher to use Setswana to teach an English lesson. Although some of the learners were listening passively and did not give examples, the teacher did not take the trouble to check if they understood or not. The learners who were responding were learners who speak Setswana as a first or second language and they were very few.
Code-switching was done out of desperation
The findings indicated that code-switching defeated the purpose of teaching and learning where learners did not understand the two target languages used, especially in rural primary schools. The impression the researcher got in the rural primary schools about code-switching and code-mixing to Setswana is that it was done by teachers for various reasons which include: a) desperation because there was no common language to use between the learners and teachers; b) to cater for teachers’ inadequacy in the language of instruction English; c) something had to be done in class to keep the classrooms noisy; d) teachers had a mandate to fulfill the syllabus and it had to be done one way or the other despite the odds; and e) it did not matter whether learners understood the lesson or not as long as the teacher can note something in her scheme book that she taught and the lesson went well. The researcher is saying this because it would not be reasonable for teachers to code-switch even though they know that learners are not benefitting from the exercise. The whole exercise of code-switching did not serve a purpose in rural primary schools.
To contribute to the issue of code switching and code mixing Baker claims that code-switching occurs for a reason. According to Baker it could be used for emphasis and clarification of a point, to attract or retain attention of the listener or to quote somebody. The classroom practices in rural primary schools fall short of Baker’s claim because teachers and learners do not speak the same languages. Code switching could work when teachers and learners are bilinguals of the two languages and it is done in the languages learners understand for emphasis, elaboration and explanation of concepts for better understanding. In light of what occurred in the classrooms, contrary to what a normal situation would present, teachers used code switching to Setswana because they wanted learners to contribute and participate in class even though the language they were expected to answer in would not offer much self-expression. Again, code-switching to Setswana also gave an impression that teachers would do anything to help learners to understand even if it means using a language they do not understand! If the teaching and learning is done as such, it means that creativity and originality are neglected. However, in these rural primary school cases teachers’ lack of proficiency in English cannot be ruled out.
Looking at it from another perspective, Crystal outlined the following reasons why speakers can code-switch such as to: compensate for deficiency, express solidarity with a particular social group used and to exclude others from a conversation who does not speak the second language. The argument is that code-switching should provide continuity in speech rather than presenting interference in language. In the case of rural primary schools the difference is that the speakers do not understand each other and hence, cause interference in language rather enhancing understanding. Kembo  argues that the major tasks of formal education is to develop pupils’ cognitive abilities, that is, their memory, their ability to generalize, to grasp relationships such as cause and effect, to predict the consequences of events, to grasp the essential message of a speech or a book and to evaluate situations. Kembo further points out that the school must develop its pupils’ social skills, that is their ability to work together with other people, to communicate with them, and to support those who need assistance. To develop all these skills requires a great deal of understanding, and according to Kembo  these skills can be developed only in a language that both the pupils and their teachers know very well, which is generally the mother tongue. Therefore, it would have been reasonable to code-switch between the learners’ first languages and the target language. It would make sense if the learners transferred some of the language traits to the target language in the process of code switching to develop their skills in the second language. This was not the case in rural primary schools.
Code-switching done due to teachers’ lack of proficiency in English
Further, code-switching to Setswana was done in schools where learners understand both English and Setswana, that is, when it was unnecessary to code-switch in schools E and F and it could be that the teachers lack proficiency in English. The classrooms were characterized by learners who were already fluent in English and also learners from foreign countries who come to school with a background of English and therefore, the code-switching to Setswana was not necessary. Important to note is that some of the learners in these classes understand Setswana. Looking at the complexity of the classrooms, a) the code-switch indicated teachers’ lack of proficiency in English; b) it excluded the foreign and immigrant learners; c) it also caused confusion on a few learners who understand some bit of Setswana because they took long to respond and try to understand some of the words in Setswana first; d) if the code-switching was a way of introducing foreigners and immigrants to Setswana, this was not necessary because the medium of instruction is English. Again, Setswana is used as a medium of instruction at Standard One only. According to the researcher, the code-switching caused confusion and undermined the proficiency in English the learners already have. Again, it delayed the learners in acquiring high cognitive abilities because of being taught in two languages simultaneously in an unsystematic manner.
As Chomsky  puts it, code-switching could be done for language development when the speaker switches due to an inability to express her/himself, it gives the speaker an opportunity for language development. As one of the School Management pointed out, teachers might be hiding behind code-switching because they are not proficient in English. The School Management argued that, ‘some teachers are not proficient in English and this leaves doubts on how language teaching is done if teachers are not able to do so with appropriate language.’ Again, Nyati-Ramahobo and Orr reported from their study on primary education and language teaching in Botswana that teachers not only code-switch for the benefit of the pupils but also because of their own problem in expressing themselves in the target language. Consistent with this argument is what Crystal points out that, “One of the reasons for code-switching is the notion that a speaker may not be able to express him/herself in one language so switches to the other to compensate for the deficiency.” A result of a deficiency in English is echoed in the report of the Revised National Policy on Education of 1993 which states that:
... A major reason why children do not perform well at Primary School Leaving Examinations is that their mastery of English is poor. A number of reasons may account for this... many primary school teachers are themselves not very competent in the use of English...
The implication in teachers code-switching to Setswana during lessons is not only contrary to what the Revised National Policy on Education  states, but it also impacts negatively on learners’ language development as one school head in school D put it. The Revised National Policy on Education  requires learners to write their examinations in English. Therefore, the poor performance in rural primary schools could be partly attributed to the code-switching. In addition, the Revised National Policy on Education  states that:
...not only is English the language of instruction at higher levels of education, but also it is the official language, the medium of communication in business and in the legal system. Its mastery therefore brings with it advantages both within the education system and in the world of work. The reality of the situation is that English will continue to flourish and hold a special place in society.
The above quotation is an indication of how important English is regarded globally and therefore its mastery by both teachers and learners is necessary. However, Kyeyune raises a question about such situations regarding the use of English as a vehicle of classroom communication by teachers: “what course is open to us for ensuring effective practice and consequently desirable standards of using English as a means of transmitting knowledge, skills, attitudes and values?” The implementation of English requires change and therefore, it is necessary for the teachers to bring about this change and it can be through being trained or in-serviced to enable them to implement the bilingual programs with minimal problems. Therefore, the code-switching in the classrooms is an indication that teachers themselves are not proficient in the language and could only cause a delay in acquisition in learners of the target language. Further, the code-switching brought about confusion because although teachers were codeswitching between English and Setswana, the learners did not understand Setswana.
Webb and Sure [19,20] argue that the reasons for poor knowledge of English are clear: a restricted exposure to English (especially in rural areas), lack of motivation to learn English, generally inadequately trained teachers, inappropriate secondlanguage teaching methodologies, inadequate educational facilities and learning materials and overcrowded classrooms. Kembo  puts emphasis on training that plays an important role in the moulding of teachers. Kembo  posits that in many parts of Africa, it is assumed that teachers at primary school level can teach all subjects. This misconception is costly to both the pupils and society. The success of a second or a foreign language teaching depends on, the quality of teacher training because they should be the soul facilitators of learning.
Williams  observed that there is lack of proficiency in English, the language of instruction, among Zambian primary school children. This in turn prevents them from learning effectively. It has been observed that many teachers still use the various mother tongues for teaching but give notes in English and expect pupils to write examinations in English, this delays their acquisition of English.
In the case of urban primary schools the teachers seem to have problems, their first language is interfering with their second language. This could be the reason why they code-switch to Setswana even if most learners understand English. After a closer look at both cases, a question could be asked whether teachers should be blamed for circumstances they find themselves codeswitching in. For the rural primary schools, the policy makers can be blamed for designing a policy that does not accommodate the use of learners’ first languages especially at the initial stages of learning. Again, it is the decision makers who post teachers to schools where both the teachers and learners do not have a common language. The teachers find themselves victims of circumstances they have no control over. The policy itself is deficient. It needs to be addressed in order for student population to learn in an even plain field.
Further, in urban primary schools, the schools decided to follow an English only policy due to caliber of learners they admit in their schools. However, allocation of teachers is overlooked because even teachers who are not proficient in English are allowed to teach at levels that require the English medium. But with the limited human resources they have, the school management does not have any option but to make use of what they have even if there are problems. This issue should make the educational authorities reflect on training of teachers. The time is now ripe for primary school teachers to teach what they are good at and what they have mastered; not to teach all subjects even if they are not good at them.
Code-switching and the language-in-education policy
The irrelevant and unnecessary code-switching could also be attributed to a deficient language-in-education policy. The policy is not inclusive of other languages. Teachers are victims of a policy that is unresponsive and recognizes only two languages out of twenty-eight that exist in the country. In some of the classrooms teachers did not speak Setswana as their first languages and code switched to their own languages. Teachers who speak the national language as a first or second language were at crossroads; they had no common language to communicate with learners who did not have Setswana as a first language. They were aware that their delivery is ineffective and therefore not reaching the learners. In this case code-switching may not be effective because it has to be done by bilinguals who understand each other. The code-switching that occurred in rural primary schools is unidentified and unclassified. It is neither assimilative nor pluralistic. It cannot change the linguistics behaviour of learners. The code-switching discussed in this paper is an issue that needs further scrutiny if the education system has all its student population at heart. The policy could be reviewed in order to achieve the educational international goals. The policy makers should consider the plurilingualism perspective and multilingual education goals. According to Lugoloobi-Nalunga  a multilingual and inclusive policy should allow students and teachers to code-switch where necessary. But the current policy allows teachers who speak Setswana as a first or second language to code-switch while others are left out.
Code-switching, teacher training and the realities in primary schools
The unnecessary code-switching to Setswana, that is, where all learners understand English could also be a result of a mismatch between teacher training and the actual practice in primary schools. For example, teachers specialize in two subjects during training, and yet they teach all subjects in the field including the ones they have not been trained for. In the field, they are stationed in one class that they monitor, supervise and teach throughout the year or years. If the teachers did not specialize in language subjects, then their delivery in language subjects could result in unnecessary code-switching as it is the case in this study. They can code-switch where learners already speak English as a first language, in the presence of immigrants and foreigners and where learners come to school speaking English from preschool. In this case the code-switching is not done for language development, it is done because teachers are desperate to teach but have a deficiency that they do not know how to handle.
Again posting teachers to primary schools where there is no common language between them and learners makes the problem more complex. The teachers and learners do not share any common language, culture, history or background and both have a difficulty in using their background as a foundation for teaching and learning as the constructivism learning theory prescribes. Code-switching is regarded as a communicative phenomenon and therefore it should facilitate the flow of classroom instruction, and help the learners to move from the known to unknown . If the code-switching retards learning it can bring misery, hopelessness and loss of interest in learning amongst learners. This may only create problems as some learners may feel rejected by the education system which is supposed to nurture their learning. The situation demands a systematic way of implementing what the teachers trained for; it requires change and change for the benefit of the learners and the teachers to deliver quality service.
Code-switching observed in this study has implication for both rural and urban primary school learners. In rural primary schools it means that learners are not benefitting from the exercise since both teachers and learners do not have a common language, therefore teachers code-switch between the languages they speak and not spoken by the learners.
In urban primary schools, the code-switching still has implications because, it is done unnecessarily, learners speak and understand English, and therefore it can delay learners’ acquisition of English.
It is important that teachers understand the role of code-switching so that they do it accordingly to emphasize or clarify content or provide clear instruction in the language the learners understand. The way it was done in this study is haphazard and can lead to frustration instead of promoting learning.
The unnecessary code-switching also has implications for the policy, since independence, the educational policy has recognized only two languages and silent about the indigenous languages, this means that we need an inclusive policy where this code-switching could be done in learners languages, because it looks like it is part of teaching and learning process.
With evidence provided in both cases of rural and urban primary schools, code-switching has proved to be a problem in the teaching and learning process. The reasons vary for both rural and urban primary schools. It is evident that code-switching in rural primary schools in Botswana is done out of desperation due to communication problems between teachers and learners and it becomes a problem rather than a solution to communication problems. The teachers decide to code-switch but using their own languages that are not understood by the learners to try and reach them. In this case we cannot leave out the language planning because the planners could have been aware that in some cases Setswana is not a first language to some Batswana and therefore learning it could present problems. Again, we need to question the posting teachers to schools where learners speak different home languages from the teachers. Also, the issue of teachers teaching all subjects after specializing in training leaves them desperate because they have to teach subjects they do not understand or are not interested in and also using the language they are not proficient in.
In urban primary schools it was clear that code-switching is done because teachers were incompetent in English because most learners had a preschool background where the medium of instruction was English. Some pupils could have learnt English for three years before doing primary school. Also, in the same classrooms there were foreigners and immigrants who were fluent in English. Therefore, the code-switching to Setswana became a problem because it made communication between teachers and learners a problem too. Further, in urban primary schools, the School Management decided that they teach in English only from Standard One because the learners come to school already speaking English and therefore breaking away to use Setswana as a medium of instruction and then switching to English caused confusion. Therefore, in the two cases of rural and urban primary schools the code-switching done was inappropriate and unnecessary; this only made the teaching and learning process complex.