The role of foreign and indigenous languages in primary schools : The case of Kenya

This article investigates the use of English and other African languages in Kenyan primary schools. English is a foreign language to the majority of Kenyans, although there are some who claim that it is a Kenyan language. English is however the official language of Kenya and, in terms of policy, the medium of instruction from Grade 4 onwards. Kiswahili, an indigenous language, is the national language in Kenya which is taught and examined as a compulsory subject from Grade 1 up to Grade 12. Kiswahili is also a subject at Kenyan universities. The other indigenous languages are only taught up to Grade 3 and only in rural schools. While the situation described above is the language policy in schools, the practice differs from the policy. Based on a comprehensive study conducted in 2006, this paper shows how English and the indigenous languages complement each other to facilitate teaching and learning in primary schools. It is shown that Mathematics and Science lessons in Grade 4 (when English becomes the medium of instruction) are conducted in bilingual English-Kiswahili code switching and in trilingual English-Kiswahili-native language code switching in urban schools and peri-urban and rural schools, respectively.


Introduction
It is clear that Kenya needs to be pragmatic about the medium of instruction and the role that foreign and indigenous languages play in the country's primary schools.With 42 spoken languages (including Kiswahili and English) within Kenya's borders and the requirement that a uniform national examination has to be written at the end of the primary school cycle, the language planners have faced a serious challenge for a long time.
Historically, English came to Kenya through British colonialism, while Kiswahili initially spread throughout the country as a trade language from the coast.Although English is now considered to be a Kenyan language by some (cf.Webb and Kembo-Sure 2000:13), to a large extent it has to be learnt formally.On the other hand, though Kiswahili is widely spoken and acquired on the basis of informal exposure to this language, Webb and Kembo-Sure (2000:9) argue that "Kiswahili itself is as alien to most rural people as is English, and even among those who claim to speak it, only a small proportion are fluent enough to engage in serious discussions".The other Kenyan languages are only used in restricted areas (their so-called "home areas"), largely for intra-ethnic interactions.Practically, therefore, Kenyan children do not have a single, uniform language that they all speak when they enrol in Grade 1.In addition, broadly speaking, there exist three linguistic zones in Kenya, namely (i) rural areas that are typically inhabited by ethnically homogenous communities and use local native languages, (ii) peri-urban areas, which, though rural, have ethnically diverse occupants and thus use Kiswahili as a language of wider communication, and (iii) urban areas that comprise large towns and cities with people from a multiplicity of ethnicities, in which Kiswahili and English are the languages of wider communication.
The official language policy in education prescribes the mother tongue as the medium of instruction in rural areas for the first three years of learning, whereas either Kiswahili or English is used in the urban schools.No clear policy exists for peri-urban schools, although they usually use Kiswahili to introduce education.During the first three years of schooling, English and Kiswahili are taught as subjects.However, from Grade 4 onwards, the teaching of and in mother tongues ceases.In Grade 4, Kiswahili continues as a subject whereas English becomes the medium of instruction as well as the language in which national examinations are written except (obviously) for examinations on Kiswahili.What this means is that the process of preparing the learners to sit for a uniform national examination (viz.the Kenya Certificate of Primary Education) starts in Grade 4, regardless of whether or not the learners are able to use English at that stage.Indeed, "at the end of the primary English course, all learners are expected to have a sufficient command of English in spoken and written forms to enable them to communicate fluently, follow subject courses and textbooks, and read for pleasure and information" (Kenya Institute of Education 2002:3).It is worth noting that although English is taught in classrooms from Grade 4, it is doubtful whether the learners converse in English when they interact among themselves on the playground and/or in class.
Thus they lack the requisite practice of speaking English.The Kenya National Examination Council's annual newsletters show that many candidates for the Kenya Certificate of Primary Examination fail to answer questions due to their inability to understand what is asked, i.e., due to a poor command of English (Kenya National Examinations Council 2004:62).Subsequently, the following questions can be asked: Are the learners practically ready to use English from Grade 4? What exactly goes on in the English-medium classes?As teachers are presumed to be role models to the learners, in what language(s) do they converse among themselves?How competent are they in English?As the newsletters of the Kenya National Examinations Council, especially the 2003 letter, show that learners fare better in the Kiswahili examination (both composition and objective) than in the English one, should Kiswahili not be used as the medium of instruction in primary schools?What is the learners' English language status in Grade 4? In what language(s) do the learners prefer to write?
This article attempts to answer these questions.The article reports on a comprehensive empirical study of Grade 4 classroom instruction in selected Kenyan primary schools that was conducted in 2006.The essence of the inquiry was to ascertain which practical language/communication problems teachers and learners face.In addition, the study identified the strategies that the teachers and learners employed to circumvent these problems.In this article, I argue that these strategies point to a practical solution for the problem of the medium of instruction in Kenyan primary schools.
In section 2, I review literature on medium of instruction, whereas section 3 sketches the conceptual framework of this research.The methodology of the study is outlined in section 4. Section 5 presents and analyses the data, and the discussion and conclusion are presented in section 6.

2.
Views from the literature on languages used as media of instruction For a long time, the issue of the medium of instruction, in especially primary schools, has persisted in spite of the 1953 UNESCO conference which resolved that it is prudent to introduce education in the child's mother tongue.This persistence is due to either the relics of colonial legacies or the local people's attitudes to certain languages (Okombo 1999;Rew 1999;Ogechi 2001;Kioko and Muthwii 2003).Since the conference, a plethora of studies focusing on the medium of instruction have been undertaken.This review focuses on the studies conducted in Africa.
It is generally acknowledged that language is central to the entire notion of education and without language education is hardly possible (Obanya 1999:19).It is therefore important that a country plans its educational language policy carefully.Multilingual Africa has tried five options (Obanya 1999:23), namely (i) promoting national languages and developing them for use in education; (ii) using indigenous languages for instruction in basic education, including teaching them as subjects at all other levels; (iii) using indigenous languages for instruction in the early years of formal education and teaching them as subjects at all other levels; (iv) teaching indigenous languages as subjects to a limited extent in basic education; and, finally, (v) an ever-lasting experimentation and argumentation on the use of indigenous languages in education.Each of these strategies has its problems, and Obanya (1999) concludes that Africa's language-in-education history is a sad one in the sense that educational policies have not reflected the sociolinguistic realities of the continent.He argues a case for the development of plans for step-by-step introduction of the official language as medium of instruction; that is, he advocates for mother tongue/primary language/first language as medium of instruction in the initial years after which the national and official languages (should these differ from the national languages) are introduced.Bamgbose (2000) and Prah (2002Prah ( , 2003) ) have for a long time followed a similar line of argumentation.According to Bamgbose (2000), African primary school learners face two major problems.First, some of their governments have erroneously decided to teach them in foreign languages that are alien to them.Second, most of their teachers are poor role models since they are not proficient in the European languages in which they teach.Thus, later in their lives, the learners have to unlearn their teachers' awkward language constructions.Bamgbose (2000) calls for more use of indigenous languages in primary education citing successful experiments in his native Nigeria.Prah (2002Prah ( , 2003) ) tends to concur with Bamgbose (2000).He contends that the quality of education offered in most of Africa is highly compromised and that graduates have difficulties finding jobs.He cites the language of instruction as the main cause, as most African learners are taught in foreign languages.According to him, all African learners, "from primary to tertiary level, should be educated in local languages, home languages, mother tongues" (Prah 2003:23).In line with other studies (cf.Delpit 1988;Hewlett 1995;Wong-Fillmore 1991;Cleghorn and Dube 1999), Prah (2003) seems to be arguing that when a second language is learned at the expense of the first, the learner fails to achieve.His approach appears rather radical.While other scholars supporting first language as medium of instruction argue that learning via the second language is facilitated when a learners' first language is allowed to further develop in as well as out of school (cf.Carey 1993;Cummins 1994;Wong-Fillmore 1985, 1991), Prah is emphatic that only learners' home languages should be used as medium of instruction, with other languages being subjects only: I am not suggesting that exclusively only African languages should be taught.The suggestion is that education should be conducted in the home languages of the people.
In addition, other languages can be learnt as subjects geared towards the cultivation of other African languages.Such additional languages could include both the old colonial languages and some local or African languages.(Prah 2003:24) The above-mentioned scholars plead for more use of African languages in education.The foregoing studies' argument has been a main theme of studies and conferences in the recent past.For instance, the 1999 conference on Indigenous Languages for Education and Technology (held in Kisumu, Kenya) argued that African languages could successfully be used in education (cf.Rew 2002).South Africa's 11 official languages (Desai 2003), Oromo in Ethiopia, and Somali in Somalia (Griefenow-Mewis 2002) were show-cased as examples of using indigenous languages in education.At the same conference, Wolff (2002) argued that the African language question in education is a creation of Africans themselves, as Africans are not ready to "throw off the yolk of colonialism".Whereas some countries insist on using foreign languages in education, the practicability of this policy is questionable.Wolff (2002) cited experiences from Kenya where, as mentioned in section 1, officially the medium of instruction after Grade 4 is English, yet in most schools English and either Kiswahili or another African language are juxtaposed.
Malawi is another African country where the answer to the medium of instruction in schools has remained elusive.In a recent national symposium on Language Policy for Education (cf.Kamwendo, Mtenje and Sandhaus 1999), the need for more use of mother tongues in Malawian primary schools was emphasised.Since 1964, teaching in Malawi has been conducted in English and Chichewa, which are alien to many Malawi nationals.Such a practice is not only pedagogically untenable but also politically undemocratic (Moto 1999).
Even though South Africa and Tanzania are considered to fare well in terms of respecting both linguistic rights and the human right to education, there exist challenges as regards these two countries' language of instruction (cf.Brock-Utne, Desai and Qorro 2003).Desai (2003) shows that in South Africa, English is the medium of instruction after Grade 4; however, in the Western Cape Province, for instance, should teachers and learners share isiXhosa as mother tongue, they use this language for spoken classroom interaction in order to bring about more effective learning.In spite of that, all written tests are done in English; the effective learning made possible in isiXhosa is never tested in isiXhosa.
The case is different in public schools of Tanzania, where Kiswahili is the medium of instruction and examination in primary schools.Although Kiswahili is not a native language to all Tanzanians, it is widely spoken by over 90% of Tanzanians, including pre-school age learners (Rubanza 1999).Subsequently, what is learnt in Kiswahili in school is easily put into practice during the learners' daily activities.The teachers themselves are competent in Kiswahili, which they learnt in school and used as a medium of instruction at both elementary and post-secondary levels (i.e.teachers' colleges).However, there are now calls for the use of English, especially in the private English-medium primary schools (Rubagumya 2003).In this regard, Kadeghe (2003), for example, is of the opinion that English as a medium of instruction in secondary schools is not the cause of academic underachievement in Tanzanian learners, seeing that Kiswahili technical terms would have to be learnt in the same way English terms are learnt.Conflicting arguments such as those above create confusion and ambivalence.If this ambivalence on the part of scholars persists, then Kiswahili as the medium of instruction is unlikely to be introduced in Tanzanian secondary schools.
In Kenya, the problems surrounding choice of medium of instruction in primary schools seem to be escalating.Since 1967 when the New Primary Approach failed miserably, it has been official policy that mother tongues are used up to Grade 3 in rural areas and either English or Kiswahili in urban areas (Republic of Kenya 1981, 1999).However, as stated in section 1, the practice does not seem to reflect the policy.Thus the medium of instruction in Kenyan schools remains a challenge for education even in the 21 st century (Digolo 2005).As happened in the early 1960s when Government falsely believed that it would be beneficial to use English as medium of instruction from Primary One (as Grade 1 was called then), at present most schools also use English from Grade 1, contrary to the current official policy (Ogechi 2001).This practice is especially prevalent in so-called private academies, which purport to offer superior education to that offered through public schools.However, whether or not this practice is pragmatically effective was the subject of my 2006 investigation on which this article reports.In some instances, the use of code switching has been reported to occur in Kenyan primary school classrooms (Cleghorn, Merrit and Abagi 1989;Jwan and Ogechi 2004).Jwan and Ogechi's (2004) study is interesting, as it involved an informal primary level centre for former street children who do not share a mother tongue.It was found that by introducing education through code switching, the learners had a smooth transition into the English only medium of instruction later.For this informal education centre, code switching was the practical "medium of instruction" in the early stages of children's educational careers.Might this "medium of instruction" be the practical one for formal primary school?If so, how can it be first taught as a language (i.e., as a subject) and then later used as a medium of instruction?That is, given that language teaching at elementary levels is prescriptive, is it possible to teach the grammar of code switching -or should schools opt for an African language only?

Conceptual framework
The search for a pragmatic medium of instruction is anchored in the concepts of human rights and globalisation.In this case, the right to education and linguistic human rights in the era of globalization form the conceptual arguments of the present discourse.The human right to education is enshrined in the United Nations General Assembly, which proclaimed the Universal Declaration of Human Rights on December 10, 1948.Article 26 (1) states: Everyone has the right to education.Education shall be free, at least in the elementary and fundamental stages.Elementary education shall be compulsory.Technical and professional education shall be made generally available and higher education shall be accessible to all on the basis of merit.Article 24 All language communities have the right to decide to what extent their language is to be present, as a vehicular language and as an object of study, at all levels of education within their territory: preschool, primary, secondary, technical and vocational, university, and adult education.
Article 27 All language communities are entitled to an education, which will enable their members to acquire knowledge of any languages related to their own cultural tradition, such as literary or sacred languages, which were formerly languages of the community.
Article 29 1.Everyone is entitled to receive education in the language specific to the territory where s/he resides.2. This right does not exclude the right to acquire oral and written knowledge of any language, which may be of use to him/her as an instrument of communication with any other language communities.
The arguments of these articles are no doubt noble.However, if they are implemented wholesale, they might end up isolating the implementing country from the rest of the world in the era of globalization.According to Punchi (2001:363) globalization entails "shared governance", "shared fortunes" and a sense of "community".It aims at making business more efficient, more discreet and more profitable through opening capital markets for transnational enterprises to operate freely in any country in the world (ibid).This therefore entails preparing individuals who are capable of using those modern information systems necessary to work on a global scale in a global medium of communication (Punchi 2001:365).Indeed, Skutnabb-Kangas (2001:372) summarises this argument as follows: If globalization is the integration and organizing of economic activities that transcend national boundaries connecting the business communities across the world on real time, a powerful global language to communicate in is also required.English seems to have acquired the status of the most dominant global language today.
The World Bank and International Monetary Fund as the two most powerful globalizers and purveyors of bank loans could easily support the argument, especially in education.As Phillipson (1999:199-200) argued: At the heart of all education is language policy.English is at the heart of contemporary globalization process, a dominant language … It is quite possible that the forces behind globalization would prefer the world to be monolingual.
So far, it is clear that there is conflict between human rights and globalization when the former are rigidly applied, especially in the poor countries which have to receive decisions made by more affluent countries -the real movers of globalization.However, there are contradictions between the agreed norm (i.e.human rights) and application of the same, especially in these more affluent countries, many of which prefer English only to respecting linguistic minorities (Skutnabb-Kangas 2001:211-212).Consider Skutnabb-Kangas' conclusion: It seems to me, then, that we need to discuss globalization in terms of common human rights implementation everywhere, rather than the West presenting requirements that other countries are asked to fulfil.Of course, a discussion about extending [human rights], about how to support (and pressurise) more countries into ratifying and implementing basic [human rights] has to continue, but without the western hypocrisy that at times characterises some of it.(Skutnabb-Kangas 2001:212-213) With this in mind, the present study argues that every child on Kenyan soil (be s/he a native Kenyan or a refugee) should have access to primary education and that such education should be presented in the language in which that child is practically proficient up until such time that s/he can use another language that gives her/him access to upward socio-economic mobility.

Methodology
The data analysed here were collected in three of the total of nine primary schools in three Kenyan provinces (Rift Valley, Nyanza and Western Province) which took part in a comprehensive project on medium of instruction in Kenyan primary schools.For the comprehensive project, each province provided three schools that had urban, peri-urban and rural characteristics, respectively.It was assumed that English and/or Kiswahili were the media of instruction from Grade 1 in urban schools.The peri-urban schools were assumed to use Kiswahili, whereas the rural schools were assumed to use a local mother tongue as the medium of instruction.
The data comprise Grade 4 Mathematics and Science lessons.This is the grade in which English officially becomes the medium of instruction in all categories of schools.Mathematics and Science were selected because they are the basic primary school subjects as per educational policy in Kenya.
The data used in this article, were collected by the researcher and an assistant.The researcher trained as a high school teacher during his undergraduate studies.He has university level teaching and research experience spanning over fifteen years in sociolinguistics including language education.The research assistant holds a doctoral degree in linguistics and is a lecturer of linguistics.Though not trained as a teacher, she has over seven years of university level teaching and research experience in linguistics.
The lessons were recorded and, during the recording time, the researcher/assistant observed and took notes, focusing on the language(s) used by both teachers and learners, both orally and in writing.Only trained and qualified primary school teachers taught the lessons studied.Any language errors (orally and in writing) were noted.
After each recording, teachers were requested to provide samples of the learners' writing in the form of essays.These were either class work or mid-term tests.In addition, learners were later asked to write an essay in a language in which they were comfortable to express themselves.The audio-recordings were transcribed and analysed to identify which medium/media of instruction had been used.

Urban schools
Results from a primary school in Eldoret town in the Rift Valley Province are analysed here.This school is located in an estate also settled by Sudanese refugees and as such has both Kenyan and Sudanese learners.Lesson delivery was done largely in English.The teachers did most of the talking, with learners making minimal oral contribution.Although teachers had been requested to ask learners questions that demand more than a single-word response, the learners mostly either gave brief responses or remained silent.The few learners who did respond actively, did so in ungrammatical English.
In addition, the learners were asked to write an essay either on Myself, My school or My worst dream.They were encouraged to write in (a) language(s) they felt comfortable to write in.The results obtained from the learners' essays are shown in Table 1.Table 1 shows that the majority of the learners wrote their essays either in English only or in Kiswahili only, while a small percentage (6.76%)used only Dholuo, Kikuyu or some Sudanese language.However, one learner each used code switching involving English and Somali, English and Sudanese or Kiswahili and English.Another learner made use of code switching between Kiswahili, English and Kikuyu.Most learners (66%; n = 39) wrote in English.The motivation to do so could be the fact that English is the official language of the school and the learners are encouraged to write in it to avoid punishment.However, what is interesting is how well some expressed themselves in English in terms of ease of formulating their own ideas, spelling, sentence and paragraph structure, as well as punctuation.
Most essays were hardly 10 lines long, as can be seen in ( 1) and ( 2) below.Most essays also consisted of numbered so-called "sentences", as in (2).We learnt that teachers encourage the numbering of sentences during guided essay writing in class work and in mid-term essay tests.I say "sentences" above, as these were actually often incomprehensible phrases (cf.(2)).
Spelling was generally poor, with some words spelt according to Kiswahili conventions, e.g.ticha (= 'teacher'), misi (= 'Miss') and veri (= 'very').Punctuation was not yet known, as some learners unconventionally used both small and capital letters, as can be seen in ( 1) where the sentences do not end in full-stops.The grammar was poor, with many ungrammatical phrases, such as class my ticha (my class teacher).The pupils' mastery of English could be said to be poor.This could be the reason why some learners began writing in English but then switched to Kiswahili.
(1) I go school My school is [name deleted by author] Primari School My school is good class ticha is Misi wafula my ticha is good ticha i am go to church is good church my school is goodveri veli good we go to schooli am good ticha in my school is ticha in yoa house my brothe is stanzanti 6 schoolis good ticha (2) I dreamed the dram 1.My dream worst 2. My worst self 3. I How spent school 4. Best my fhend 5. My self best school 6.My the How However, one case stood out in terms of her performance in Grade 4.She was among those who wrote their essays fully in Kiswahili.It should be noted that none of the essays written in Kiswahili was less than half a page, as opposed to the maximum ten lines of the English-only essays.Furthermore, none of the Kiswahili essays had numbered sentences/lines; they were in the form of continuous prose.The essay of the Grade 4 learner in question, a second language speaker of Kiswahili, covered a page of an exercise book.The concepts of sentences, paragraph and punctuation were not fully mastered, and her spelling indicated a mother tongue influence, e.g.akanibika instead of akanipiga (= 's/he beat me'), usingisi instead of usingizi (= 'sleep'), asubui instead of asubuhi (= 'morning') and akaniumisa instead of akaniumiza (= 's/he hurt me').On balance, however, one could easily follow the writer's story line, because she expressed herself clearly, as can be seen in ( 3).
[English gloss: One day I dreamt that a thief came, beat me up and injured me seriously.I cried so much but I was just dreaming.I cried out, "Mother!Mother!"My mother rushed to where I was and found me asleep.She went back to sleep knowing that I was dreaming.When morning came, I woke up and washed my face and went out to take tea.My mother asked me, "Why were you crying at night"?I was not aware I had cried at night.I asked, "At what time"?My mother said, "Perhaps you were dreaming".I said, "Maybe I was dreaming".]

Peri-urban schools
The peri-urban schools in the Rift Valley and Western provinces are situated in what were "white highlands" during the colonial era.People from different ethnic backgrounds whose language of wider communication is Kiswahili presently inhabit these "white highlands".
Of the sampled essays that learners wrote as class exercises or mid-term tests, many comprised numbered sentences.Spelling was a problem, and clarity and ease of expression were seriously lacking.
The most interesting observations were however made during the classroom interaction.Talk in the Mathematics lesson was mainly teacher-centred, and where learners did respond, they used one-word sentences only.However, learners usually only responded when they were allowed to respond in a language they felt comfortable to use.Teachers also resorted to switching between Kiswahili and English during lessons.For instance, during a Science lesson on flowering plants, concepts were mentioned in English but the explanation was offered in Kiswahili.It was observed that a Kiswahili equivalent was given to almost all that was said in English and that almost no learner could respond to questions asked in English, as can be seen from ( 4) and ( 5).As can also be seen in ( 5), in some instances, the learners' responses were in Kiswahili although the official medium of instruction was English.It therefore appears that learners do not know English well enough to allow classroom activities to take place in English only.For instance, the class was asked to enumerate types of tubers.One learner named cassava, which the teacher promptly wrote on the blackboard.The next learner to name a tuber said muhogo (= 'cassava') -a sign that the learner did not know that cassava is the English equivalent of muhogo.
Instructions from the teacher which required an immediate response from learners had to be given in Kiswahili as the lesson conclusion given in (6) testified: What was also apparent is that even if the learners responded in Kiswahili, the teacher wrote the answers and notes on the blackboard in English.The learners later copied these notes.One wonders whether or not the learners were able to read and understand the notes in English.

Rural schools
The lessons in rural schools were also conducted in both English and Kiswahili, amongst others because the teachers at times had to translate questions asked in English into Kiswahili.In addition, the learners answered the teacher's English questions in Kiswahili.In fact, whenever the teacher noticed reluctance to respond, s/he permitted the learners to also answer in Kiswahili.Once this permission was granted, more learners raised their hands.However, when the learners responded in English, their answers were in awkward English, as illustrated in (7) through responses of three different learners: As indicated earlier, learners learn awkward English from their teachers who themselves do not speak English with mother tongue speaker proficiency.In fact, one teacher's English blackboard notes contained many errors, which the learners in her class copied verbatim.Examples of such errors are provided in ( 8 This was seen not only in Science and Mathematics lessons but also in English lessons.As indicated earlier, the languages of wider communication in rural areas are the local languages; learners encounter English and Kiswahili in school for the first time.Of importance here is that most teachers in the rural schools are also from the local communities.Thus they too use English and Kiswahili only in school when teaching.Most of them were observed using their local languages outside of the classrooms.It could be that their restricted use of English and Kiswahili negatively influences their use of these languages while teaching.

Discussion and conclusion
In this study, my aim was to ascertain whether or not English as a medium of instruction is practical and appropriate in Grade 4. Based on the discussion above of three typical schoolsone urban, one peri-urban and one rural -it has emerged that the official medium of instruction (English only) for Kenyan Grade 4s is problematic: Learners are not ready to use English at the beginning of Grade 4. (It could be that they are ready to do so later in the year, but this seems unlikely.)It appears that learners in the urban primary schools are ahead of their peri-urban and rural school peers in the mastery of English.Furthermore, learners in the latter two categories appear enthusiastic to answer questions when teachers translate or rephrase their questions in Kiswahili (but not when the questions are posed in English), which indicates that these learners are not apathetic as regards classroom participation but that their command of English is not good enough to allow such participation.However, the fact that some urban primary school learners also wrote their essays in languages other than English indicates that they too are not (fully) comfortable with English-only at the start of Grade 4. Teachers, especially those in rural and peri-urban schools, have difficulties using grammatical English.This could be the case because (1) for a long time, most candidates admitted to teacher training college did not obtain good marks at school.Their average mark for the Kenya Certificate of Secondary Education (KCSE) was a D+ (Ministry of Education, Science and Technology 2001: 22; National Assembly 2009: 59), indicating that the candidates are not academically strong, and (2) the teachers in these areas operate in an environment where English does not feature in their everyday communication, except in class.Thus they lose their proficiency in English, if they possess any at all.
The observations show that there is no uniform language that was prevalent in the classrooms observed.Teachers appear to use code switching between English and Kiswahili as an effective strategy across the three categories of schools.However, code switching is a synchronic phenomenon and as such cannot be a medium of instruction.
The data analysed points to three possible recommendations for policy makers to consider.Firstly, the use of English at the start of Grade 4 is an infringement of both linguistic rights and the human right to education.Learners are not ready to use it and some teachers are poor role models of both spoken and written English.Secondly, the fact that there is an enthusiastic response when teachers code switch to Kiswahili in urban, peri-urban and rural schools shows that, if more of an effort is made, Kiswahili could be an effective medium of instruction and could be used in setting and writing a uniform national examination.Indeed, compared with those essays written in English, the learners' Kiswahili essays are more grammatical, comprehensive and advanced in terms of narrative structure.Finally, there is a need to decentralise examinations.English is being imposed on every child because a national examination has to be written.However, if the examinations were decentralised so that regions which share a medium of instruction are allowed to do their examinations in that language, then English, Kiswahili or the mother tongue could easily be used in regions where they are the dominant languages.
These recommendations should however not be implemented in such a manner that the end result is monolingualism.While I support a home language as the medium of instruction, language policy and the implementation thereof should actively encourage bilingualism, regardless of which model is adopted for use in schools.
a fruit?No pupil raises a hand to respond Teacher: Ninasema, ni wangapi wanajua matunda?(= 'I am saying, how many of you know what a fruit is?') Entire class raises their hands (5) Teacher: What does a root do?Only one out of 54 learners raised a hand Teacher: Mizizi inasaidia nini?(= 'What do roots do?')More than half the class raised their hands Learner: Inasaidia isianguke.(= 'It assists (the plants) not to fall') Kuna swali?OK.Kama hakuna swali, andika hiyo.(= 'Any questions?OK.If there are none, write what is on the blackboard')

measures to eliminate obstacles limiting effective access to education
The United Nations reaffirmed the right to education in a report of the Commission on HumanRights (2004), where it urged all States as follows (emphasis mine): a) To give full effect to the right to education and to guarantee that this right is recognised and exercised without discrimination of any kind.b) To take all appropriate , notably for girls, including pregnant girls, children

living in rural areas, children belonging to minority groups, indigenous children, migrant
children, refugee children, internally displaced children, children affected by armed conflicts, children with disabilities, children affected by infectious diseases, including HIV/AIDS, sexually exploited children, children deprived of their liberty, children living in the street, working children and orphaned children.c) To take all legislative measures to prohibit explicitly discrimination in education on the basis of race, colour, descent, national, ethnic or social origin, sex, language, religion, political or other opinion, property, disability, birth or other status which has the purpose or effect of nullifying or impairing equality of treatment in education.
This human right to education is only made possible if linguistic human rights are respected.The linguistic human rights in education are emphasised in Section II of the United Nations' 1996 Barcelona Universal Declaration on Linguistic Human Rights (United Nations 1998:27-28):

Table 1 .
Language(s) used in learners' essays -rural school