Language practices in Namibian primary schools

ABSTRACT

The main focus of the article is on teachers’ classroom language practices in eight primary schools in Oshana and Khomas school regions. Language ideologies and translanguaging practices were used as a theoretical framework to evaluate the language ecologies that were created by the language practices in the schools included in the research study.

The data consist of 140 responded questionnaires, 19 individual interviews and eight focus group discussions with teachers, and 47 observed lessons. Mixed methods were used to analyse the data. Questionnaire responses were analysed quantitatively except for open-ended questions. Qualitative content analysis was used to analyse the interviews and focus group discussions as well as the observed lessons.

The results show that despite teachers’ efforts to gear the classroom language practices towards a more multilingual language ecology by using different multilingual strategies, the prevailing hierarchical language ideologies, reinforced by the assessment that is done in English only, contribute to maintaining the dominant position of English compared to the local languages.

KEYWORDS:

• Language ideologies
• Namibian language education policy
• school language ecology
• teachers’ language practices
• translanguaging

Introduction

In postcolonial countries, the educational language policy (LP) often involves a European language and some local languages. They do not have the same social value, however, but constitute complex hierarchies and are therefore used in different domains in the society. This has an impact on societal language ideologies that shape educational practices and discourses. Language ideologies are therefore a useful concept in showing how these processes relate to one another (da Costa Cabral 2021). Teachers’ language practices are closely related to the notion of legitimate language (Bourdieu 1991), which means the language variants that are recognised as acceptable for use in public domains and associated with authority. The social status and function of languages are central concepts in the study of language ecology, as well, and thus related to the question of the legitimacy of languages and the hierarchies they construct. Language ecology has been defined as ‘study of interactions between any given language and its environment’ (Haugen 2001, 57) and it is interested in the ways languages interact in the minds of multilingual speakers as well as in multilingual societies.

Education is a key site of defining, producing and reproducing what is considered legitimate language (Heller and Martin-Jones 2001). As ideologies influence teachers’ pedagogical decisions, uncovering the ideologies embedded in their classroom practices would help training teachers to be more aware of them (Kiramba 2018, 293). Translanguaging pedagogy is a means of creating multilingual spaces and connecting multilingual students’ home and community language practices to what is considered legitimate language practices at school (Otheguy et al. 2015; García, Ibarra Johnson, and Seltzer 2017). It is used in this paper as a framework to evaluate the nature of the language ecologies that are created by the language practices.

As teachers are powerful LP implementers (García and Menken 2010), the focus in this paper is on their classroom practices. As regards teachers’ practices, certain variables have been suggested in previous research as being related to them. They include the subject (Uys et al. 2007; Iipinge and Banda 2020), teachers’ age and experience (Alisaari et al. 2019; Borg 2018) and the degree of linguistic diversity (Iipinge and Banda 2020) which are therefore compared in this study. Namibia represents a typical postcolonial language education policy with English as the main medium of instruction (MoI) and despite the limitations of this study, the results are therefore applicable to other similar contexts. The study contributes to understanding the relationship between the societal language ideologies and the language practices in primary schools especially.

The data consist of a questionnaire addressed to teachers, 19 individual semi-structured teacher interviews and eight focus group discussions, and 47 lesson observations conducted in eight government primary schools in Khomas and Oshana regions. The theoretical framework of the study is discussed in Section 2, and the data and method are presented in Section 3. The findings are presented in Section 4, followed by a discussion in Section 5 and concluding remarks in Section 6.

Literature review

Language ideologies influence teachers’ beliefs and practices. Silverstein (1979, 193) defines them as ‘sets of beliefs about language articulated by users as a rationalisation or justification of perceived language structure and use’. In other words, socially held beliefs about languages impact the way legitimate language is determined by rationalising and justifying the roles given to different languages. As Rampton (2006, 15) points out, ideological discourses construct ‘communities’ and ‘link languages with peoples’. This means that public domains are ideological constructs, as well, and language practices contribute to legitimise specific political arrangements (Gal and Woolard 2001). In this paper, language ideologies are understood to include the values and beliefs as well as the discourse about language use.

The notion of legitimate language is central in educational language policies, as it determines what language practices are valued and considered normal or appropriate according to the prevailing ideological orientations (Heller and Martin-Jones 2001). Weber and Horner (2012) distinguish five types of language ideologies, two of which are especially relevant in post-colonial contexts (cf. Iipinge and Banda 2020). First, ‘language hierarchy’ ideology is prevalent in a situation where one language dominates and the others follow in decreasing order of importance. Second, ‘one nation – one language’ ideology is the basic ideology on which European nation states have been constructed and that has also affected the language policies in postcolonial countries (Beck 1995; Norro 2022). The power position of the dominant language is further reinforced by associating it with a national territory and according it the status of official language.

Language ideologies and educational language practices are tied together as language ideologies and hierarchies are reproduced, but also contested in classroom interaction (Heller and Martin-Jones 2001; da Costa Cabral 2021). Teachers are constrained by the prevailing conditions, and often have to harmonise contradicting interests, but they also possess resources and space to execute their agency, and empirical study of their language practices help understand the way structure and agency are articulated (Heller and Martin-Jones 2001). Assessment policy has been argued to affect de facto language policies (Shohamy 2006) and is a constraint affecting educational language practices largely.

Translanguaging pedagogy is based on multilinguals’ natural fluid language practices and includes for instance the use of all learners’ languages in partner discussions and group work, multilingual materials, word and cognate walls (García, Ibarra Johnson, and Seltzer 2017; Menken and Sánchez 2019). Translanguaging practices, when legitimated in the classroom, may help connect home and school language practices and create a multilingual language ecology. According to many theories of translanguaging (e.g. García and Kleyn 2016), there is a fundamental difference between code switching and translanguaging, as the former is said to presuppose that multilinguals have separate systems of named languages whereas the latter refers to one single repertoire multilinguals draw on. However, Cummins (2021) argues for a crosslinguistic theory of translanguaging that affirms the reality of ‘languages’ in a multilingual’s cognitive system and sees both code switching and translation as instances of translanguaging. The latter is adopted in this paper.

The dominant position of English in the Namibian school system and its effects on the educational language practices have been investigated in several studies (Uys et al. 2007; Banda, Mostert, and Wikan 2012; Van Der Walt 2015). Simasiku, Kasanda, and Smit (2015) and Iipinge and Banda (2020) collected their data in secondary schools in Caprivi and Omusati regions respectively, and discovered teachers having mixed opinions on code switching in class. Those in favour of it believed it enhances teaching, learning and student participation whereas those against it said it to be against the LP and have negative consequences for students who are assessed in English only. Sometimes a lack of terminology in local languages hinders the use of languages other than English (LOTE). For instance, Kaphesi (2003, 277) observed mathematics teachers in Malawi having problems in finding equivalent mathematical terms between Chichewa and English.

In previous studies, differences between language teachers and content subject teachers have been observed (Uys et al. 2007; Iipinge and Banda 2020). Comparisons between different subject teachers and the qualified and non-qualified to teach in mother tongue were therefore made in this study to see if their practices vary. The high number of different languages has also been stated as an argument against the use of LOTE in class (see e.g. Iipinge and Banda 2020), so teachers’ practices in this study were compared between the linguistically highly diverse Khomas region and the less diverse Oshana region. As teachers’ own experiences have been shown to affect their beliefs and practices (Alisaari et al. 2019; Borg 2018), their age and teaching experience were also considered when comparing their practices. The findings of this study are compared with previous research to discover whether they corroborate or differ from them, especially as regards the use of LOTE and the effects of assessment policies.

The research questions this paper endeavours to answer are:

1. What self-reported classroom language practices do teachers have and how are they related to certain background factors?

2. What observed practices do teachers have and how do they either correspond or differ from the self-reported practices?

3. What kind of language ecology do teachers’ practices create?

The context of the study is described in the beginning of Section 3 before presenting the data and method.

Data and method

Context of the study

The national language education policy in Namibia allocates the status of school languages to 14 languages and stipulates the use of the pupil’s home language or a predominant local language as MoI in junior primary (grades 0–3). Grade 4 is a transitional year, and English is the only MoI in senior primary (grades 4–7). It is still allowed to use other languages ‘in a supportive role’ to ensure pupils understand new concepts (National Curriculum of Basic Education 2016). Schools may obtain permission to offer English medium instruction even in junior primary, and the Ministry of Education’s statistics show that almost 25% of grade 1–3 learners are enrolled in English medium schools (Education Statistics 2019).

The two regions where the data was collected, Khomas in the capital area and Oshana in the north differ in their linguistic diversity, Khomas being more diverse than Oshana, where the local Oshiwambo languages (Kwanyama, Kwambi and Ndonga) are the most widely spoken. Recently, there has been movement towards northern economic centres, which has diversified the linguistic ecology of the region, especially in urban areas. One of the schools in Oshana and three of the schools in Khomas offer local languages (Khoekhoegowab, Oshindonga, Oshikwanyama, Otjiherero and Afrikaans) as MoI in junior primary, the rest only offer English medium. All eight schools cater pupils from low- to middle-income families.

Data collection

The data consist of a questionnaire (Appendix 1), 19 semi-structured individual interviews and eight focus group discussions with 3–6 participants, all conducted in English¹, as well as lesson observations. The questionnaire and the interviews are used to obtain data on teachers’ self-reported classroom practices in order to answer research question 1. One to three individual interviews and one focus group discussion were conducted in each school. Of all interviewees, 78% were female and 22% male. The questionnaire and the interviews also included data about teachers’ beliefs, but they have been analysed and discussed in Norro (2021, forthcoming).

Lesson observations offer data on teachers’ observed practices and insight into the relationship between their self-reported and observed practices (research question 2). As consent to video recordings of the lessons was not obtained from all parents and pupils, the author took notes during the lessons using an observation frame, and the observation notes form the data for research question 2.

The data collection took place in two separate periods. The first one was in the beginning of 2020 school year when data was collected in two schools in Khomas. It was however interrupted because of the Covid-19 lockdown. A second data collection was made in September–November 2021 in two more schools in Khomas and in four schools in Oshana. The permissions to conduct research were obtained from the Ministry of Education’s Permanent Secretary, from the National Commission for Research, Science and Technology (NCRST) and from the Regional Directors in Khomas and in Oshana. The preliminary ethical review by the ethical committee of the researcher’s affiliated university² was obtained as well.

The school principals were contacted to obtain their consent for the research. The participants received information on the preservation of their anonymity, the focus of the research and the secure storage of the data in an information letter. The principals and the teachers involved gave their documented voluntary consent for the questionnaire and the interviews. The sampling was done based on the availability and voluntary participation of the teachers.

The questionnaire was distributed to all the teachers, the response rate being 51%. The questionnaire was available online as well, and five teachers took it there. The total number of questionnaires included in the data is 140. In the questionnaire, there were questions about teachers’ background that are presented in Table 1.

Table 1. Background variables of the participants (questionnaire).

Background variables
Region		Age		Teaching experience		Subject		Junior / senior primary		Pupils’ home languages		Qualified to teach in home language
Khomas	69	Under 40	88	0–5 years	45	Class teaching	22	Junior	22	1	47	Yes	83
Oshana	71	Over 40	49	6–15 years	34	Languages	36	Senior	87	2 or 3	52	No	54
				16- years	36	Social studies	15			more than 3	41
						Mathematics & Natural science	31

For the analysis, the subject combinations that the respondents had given in the questionnaire were grouped into four subject groups, that is, class teaching, mathematics and natural science, social studies and languages (including L1 and L2) according to the first subject that they mentioned. Though they may in reality teach several subjects, it was assumed that they mostly teach or identify with the first one.

Altogether 149 lessons were observed, 76 in Khomas and 73 in Oshana, 47 of which were used in the analysis. The sampling was based on the subject groups presented above. Grade four had been selected as the focus of the observations because it is the transitional year between mother tongue and English instruction. Therefore, class teaching was not included in the analysis, and does not figure in Table 2. Moreover, English lessons were the only language classes that were selected because of the restricted number of other language lessons available. Observed lessons were included in the analysis if they belonged to one of these subject groups, if the teacher in question had also been interviewed individually, and if there were at least four observed lessons with the same teacher. This sampling resulted in the data presented in Table 2. Teachers’ names are pseudonyms.

Table 2. Observed lessons included in the analysis.

Subject	School	Teacher	Number of lessons	Topics
English second language	Khomas, S1	Martha	2 × 80 min + 2 × 40 min	Sports; Alphabetical order and use of dictionaries; Preparing a listening comprehension
	Khomas, S2	Helena	2 × 40 min + 1 × 80 min	Animals; Vocabulary and writing
	Oshana, S1	Lydia	4 × 40 min	Types and parts of sentences
	Oshana, S3	Irma	4 × 40 min	Poems and elements of poems; Articles
Social studies	Khomas, S2	Sarah	1 × 80 min + 4 × 40 min	Drawing a map of the classroom
	Khomas, S3	Amanda	2 × 80 min + 2 × 40 min	Communication and modern technology
	Oshana, S4	Rachel	3 × 40 min + 1 × 80 min	Communication and modern technology
Mathematics	Oshana, S1	Henry	5 × 40 min	Fractions
Natural science and health education	Khomas, S2	Toini	4 × 40 min	Personal health and healthy food
	Oshana, S1	Laeticia	1 × 80 min + 3 × 40 min	Matter and non-matter; Living and non-living matter
	Oshana, S3	Clara	2 × 40 min + 1 × 80 min	Air and wind; Revision and test
	Oshana, S4	Vanessa	4 × 40 min	Properties of the states of matter; Test

The times in Table 2 follow the official lessons of 40 min, though the teachers did not always respect the time limits set in the time table. Four to five lessons per teacher have been included, covering one or two topic sequences.

Data analysis

Mixed methods were used for the analysis. The questionnaire responses were analysed quantitatively, except for open-ended questions. There were 17 statements about teaching practices that are frequently used in multilingual education, especially translanguaging pedagogy (García, Ibarra Johnson, and Seltzer 2017; Menken and Sánchez 2019). The respondents were supposed to choose between often/sometimes/never, according to how often they use these strategies. For the analysis, the options were given the values of 1 (never), 2 (sometimes) and 3 (often). Mean scores and standard deviation were calculated for the classroom practices and a factor analysis was conducted in order to group them. The closer the mean score for a practice is to 1, the less the respondents use it. If the score is between 2 and 3, they use it sometimes or often. The estimation method used for the factor analysis was maximum likelihood and the rotation method was oblimin method. If rotated factor loading was above 0.4 and final communality was above 0.25 the items were taken into final phase. Comparisons between different background factors and language practices were made to discover which variables are related. Association between two categorical variables were evaluated with Fisher's Exact Test (Hess and Hess 2017, 878–879). If not stated otherwise, the association was assumed to be statistically significant if p < 0.05. When several variables were compared simultaneously, the Bonferroni correction (Tasdan and Yeniau 2016, 256) was used to overcome the multiple comparison problem.

Individual interviews (19) and focus group discussions (8) were analysed as one data set and are both referred to in the text as ‘interviews’. For the analysis, orthographic transcription of the interviews was conducted. It is considered accurate enough for thematic analysis (Braun and Clarke 2006, 8). The interviewees were given a running code in order to anonymize them. The individual interviewees are referred to using a code formed of the initial of the region, the number of the school and the number of the interview, for instance O,S1,1 (Oshana, school 1, interview 1). The focus group participants are referred to by the initial of the region, the number of the school and a number of the participant according to the order they initiate their participation in the interview, for example K, FGD1, P1 (Khomas, focus group discussion 1, participant 1).

The interviews were analysed using qualitative content analysis (Hsieh and Shannon 2005; Elo and Kyngäs 2008; Schreier 2012). It is considered a suitable method for analysing rich data that requires interpretation, such as interviews and focus groups (Schreier 2012). The coding frame was built mainly inductively (Hsieh and Shannon 2005), though the research questions were used to determine which parts of the data were relevant (Schreier 2012, 81). Units of coding were determined using a thematic criterion (Schreier 2012, 136), that is, choosing a passage about a certain topic as unit of coding. After an initial open coding (Elo and Kyngäs 2008, 109–110), categories were formed and grouped into a coding frame. The main coding was then done using the coding frame, which was adjusted to fit all the data. A second round of coding was made after ten days, and the percentage of agreement (90%) was calculated between the two rounds of coding (Schreier 2012, 203–204) to evaluate the reliability and stability of the analysis. Some subcategories were then adjusted before the final coding. The findings of the questionnaire are reported with tables and discussed in the text against the theoretical background and previous research. The findings in the interviews are reported in continuous text and illustrated by excerpts from the interviews.

The field notes from the observed lessons were similarly submitted to content analysis, using the findings of the factor analysis described above as a coding frame in order to triangulate the self-reported practices with observed practices and to find out whether they corroborate or contradict each other (Creswell and Plano Clark 2007, 63). As a result of an initial coding, the coding frame was completed with some data-driven subcategories. The percentage of agreement between the two main coding rounds was 93%, and the differing codings were adjusted in the final coding. The results were tabulated and analysed to reveal any consistencies or discrepancies compared to the interview and questionnaire data. The findings of the analyses of all three datasets is presented in Section 4.

Findings

The questionnaire responses and the interviews were analysed to find out what language practices the participants have and what factors are related to the choice and frequency of the practices (research question 1). Language practices include teachers’ preference for a certain MoI, English or home language, and their classroom practices.

Teachers’ medium of instruction preferences

The background factors show some tendencies of teachers preferring a different MoI according to whether they are qualified to teach in home language, the subject they teach and the number of their pupils’ home languages.

Altogether 83 respondents (60%) told they are qualified to teach in their home language. However, 103 respondents (75%) preferred teaching in English. Almost all respondents (98%) not qualified to teach in home language prefer teaching in English. However, half (49%) of the qualified also prefer teaching in English, and 41% in their home language (10% in both). Thus, being qualified to teach in home language does not automatically mean that the teacher prefers home language as MoI. Teachers gave several reasons for preferring English as MoI in the open-ended question. The most common are cited in excerpt 1.

Ex. 1 Because it is a medium of instruction.

English is the language, which I was taught in, from primary school right through to university.

It makes it easier to explain to learners of different backgrounds.

One important reason to prefer English as MoI was teachers’ own experience as students (see the second comment in excerpt 1). Many told they preferred English simply because it is the official MoI (cf. the first comment in excerpt 1), so they align to the official policy without contesting it. Practical reasons, such as the lack of material or appropriate vocabulary in local languages were also mentioned, as well as linguistically diverse groups (the third comment) where using LOTE was considered a non-suitable practice.

Class teachers and language teachers preferred teaching in home language more than the other two subject groups. Amongst both class teachers and language teachers there are those who actually teach in home language, and this may affect the results. The most striking difference (p = 0.0008) was observed between mathematics teachers and language teachers. Almost all (94%) mathematics and natural science teachers prefer teaching in English whereas only 52% of language teachers prefer English, 40% home language, and 9% both.

Ex. 2 I teach math and science, yeah I find it difficult sometimes when I’m teaching a topic and the learners don’t understand it, and I I as a teacher don’t have the right words in my mother tongue to ex-/ to tell the learners so that they can understand what I’m trying to teach them, so I’m just forced to teach in English. (O, FGD1, P1)

The lack of mathematical and scientific terms may explain the reluctance of the teachers of these subjects to use home languages as MoI (cf. Kaphesi 2003, 277). As excerpt 2 shows, even if they wanted to use them, the lack or ignorance of the appropriate terms forms an obstacle.

The degree of diversity of pupils’ home languages is related to teachers’ MoI preferences. The number of pupils’ home languages varied from one to ten, and for the analysis, they were divided into three groups: one, 2–3, or more than three languages. The findings are presented in Table 3.

Table 3. Number of pupils’ home languages and teachers’ language preferences.

Number of home languages	Both	English	Home language	Total
1	2% (1)	60% (28)	38% (18)	47
2 or 3	8% (4)	65% (32)	27% (13)	49
More than 3	8% (3)	85% (34)	8% (3)	40
Total	6% (8)	69% (94)	25% (34)	n = 136
			Fisher's Exact Test	p = 0.007

As can be seen in Table 3, the proportion of those who prefer teaching in English grows as the number of pupils’ home languages increases (p = 0.007). When there is one major home language, 60% prefer English. When there are 2–3 home languages, 65% prefer English, and when the number of home languages is more than three, 85% prefer English. In the interviews, a recurring statement was that it is impossible to use home languages in class when the pupils come from various language backgrounds (cf. Iipinge and Banda 2020).

Classroom practices

The questionnaire included questions about 17 classroom practices that are frequently used in multilingual education, especially translanguaging pedagogy (García, Ibarra Johnson, and Seltzer 2017; Menken and Sánchez 2019). The frequencies are presented in Table 4.

Table 4. Classroom practices.

	Mean score	Standard deviation	Factor loading	Communality
Factor 1 Multilingual practices
I introduce a new topic in the home language	1.4	0.7	0.85	0.69
I let the learners use their home languages when doing group work	1.6	0.7	0.74	0.54
I explain the core terms and concepts in the learners’ home language	1.9	0.8	0.69	0.46
I use some words, e.g. greetings, in the learners’ home languages in class to make them feel comfortable	1.9	0.8	0.68	0.49
I accept that learners use other languages than English in assessments	1.4	0.6	0.55	0.29
I accept that learners use their home languages or mix codes in class	1.9	0.7	0.53	0.33
I make the learners’ home languages visible in class (e.g. posters / word walls / drawings)	1.8	0.9	0.48	0.25
I switch the language if I notice that the learners do not understand	2.1	0.6	0.46	0.32
Factor 2 Monolingual practices
I help learners correct their oral utterances by asking them to repeat and by providing clues to the correct form	2.7	0.5	0.65	0.48
I use charts, pictures, drawings and other visual support	2.7	0.5	0.55	0.29
I reformulate learners’ erroneous utterances in correct form	2.6	0.6	0.52	0.32
I split new content into smaller chunks	2.4	0.6	0.51	0.27

Two factors could be found in the analysis, namely multilingual practices that are based on the use of LOTE to scaffold learning or to make the pupils feel more comfortable, and monolingual practices that include visual support, helping learners use academic English and splitting the content into smaller chunks. As can be seen in Table 4, monolingual practices are used more frequently than the multilingual practices. This may be explained by the fact that they are usual teaching strategies in any context. However, the smaller frequency of the multilingual practices may also indicate the reluctance of the teachers to use them because they are not stipulated by the official policy, except for the allowance (cf. The National Curriculum of Basic Education 2016, 29) of switching the language when necessary.

Switching the language was more common in Oshana than in Khomas. A distinct difference between the two regions, Khomas and Oshana, can be observed when comparing the diversity of pupils’ home languages. There is one major language spoken by the pupils in 20% of the responses in Khomas and 46% in Oshana. In Khomas, 2–3 home languages are spoken in 36% of the cases and in 39% in Oshana. Over 3 home languages are spoken in 43% of the cases in Khomas, but in only 16% of the cases in Oshana. The interviews show however that there is a difference between urban and rural settings.

Ex. 3 I think learners who are from town, they are much better than those ones from the village. Yeah, because in town … they use to meet people who are communicating in different languages, they use to watch televisions, but at the village, people, we use to communicate in our mother tongue. And you cannot find somebody, speaking to you … in English. (O, S3, Int. 3)

As the teacher’s comment in excerpt 3 shows, linguistic diversity is higher in urban than in rural communities, and pupils living in rural settlements are less exposed to English outside school (cf. Probyn 2021). According to excerpt 3, this affects their performance in school.

Statistically significant differences (p < 0.003) related to region were observed in statements 2.(a) I switch the language if I notice that the learners do not understand, 2.(d) I explain the core terms and concepts in the learners’ home language, and 2.(h) I make the learners’ home languages visible in class. The findings are presented in Table 5.

Table 5. Language practices in class according to region.

2. (a) I switch the language if I notice that the learners do not understand
Region	Often	Sometimes	Never	Total
Khomas	26% (18)	49% (34)	25% (17)	69
Oshana	20% (14)	76% (53)	4% (3)	70
Total				n = 139
	Fisher's Exact Test			0.0005
2. (d) I explain the core terms and concepts in the learners’ home language
Khomas	22% (15)	34% (23)	43% (29)	67
Oshana	22% (15)	62% (43)	16% (11)	69
Total				n = 136
	Fisher's Exact Test			0.0007
2. (h) I make the learners’ home languages visible in class (e.g. posters / word walls / drawings)
Khomas	19% (13)	9% (6)	72% (48)	67
Oshana	41% (28)	22% (15)	37% (25)	68
Total				n = 135
	Fisher's Exact Test			0.0002

Teachers switch the language more often in Oshana, where the classes are linguistically more homogenous than in Khomas. In Oshana, 96% say they switch the language often or sometimes whereas 76% do so in Khomas (p = 0.0005). Teachers also explain core terms and concepts in learners’ home languages more in Oshana, where 84% report doing so often or sometimes, in Khomas 56%. The reason for this arose in the interviews, as shown in excerpt 4.

Ex. 4 It means if you explain to this … group of learners in Oshiwambo, what about the other language. (K, S2, Int.1)

Teachers do not know all the learners’ languages and feel that if they use one of the local languages, they do not treat all learners equally. In the interviews, there are however examples of teachers using more than one local language to explain a concept and negotiating with the pupils about the right term in different languages.

Ex. 5 Like in the Afrikaans class now … so when you’re teaching it you need to go back … you translate now what do you call it in English … And if they don’t know it in English, then you go to their mother tongue … just to bring the idea to the learner … then you go to Afrikaans, so you first go to their mother tongue, to English, and then to Afrikaans. (K, FGD 1, P1)

As excerpt 5 shows, negotiation may be done in several languages to finally reach the goal, and learners are actively involved in this process. Sometimes teachers might also switch to the predominant local language, even if it is not the home language for all learners.

Ex. 6 All learners at the school are taught Oshikwanyama and English … so basically, even the Otjiherero speakers are taught Oshiwambo, so I think that’s when the other teachers are are comfortable in trying to explain … those in Oshiwambo. (O, FGD4, P3)

As excerpt 6 shows, the predominant local languages may be used to clarify the content if all learners learn them as subjects, depending on the LP of the school.

Teachers in different regions also have different practices as regards making the pupils’ home languages visible in class. In Oshana, 64% do so often or sometimes, in Khomas only 28%. Apparently, this practice is considered more appropriate when the majority of the pupils speak the predominant local language.

As regards classroom practices, significant differences were observed between junior and senior primary, but not between the different subject groups.

Ex.7 In junior primary, we use it most of the time. Especially the first three to four months … like in my class I try to use most languages that I can speak. (K, FGD1, P2)

In junior primary, home languages are used more than in senior primary, even when the official MoI is English, like in excerpt 7, because the pupils’ proficiency in English is usually lower than later on. When the MoI changes into English on the fourth grade, teachers use LOTE less, though it is still permitted to use them, but ‘in a supportive role only’ (National Curriculum for Basic Education 2016, 29). The interviewees also brought up their concern for assessment that is done in English only.

Ex. 8 But I have to go back to English … because the questions papers and the tests and everything … and the assessment is set in English … so, the mother tongue is just there for them to make them aware of what’s going on, but later on we have to switch back. (O, FGD1, P4)

Excerpt 8 shows that teachers want to make sure they do not use LOTE too much and that the learners always get the content in English eventually, because the assessment is made in English only. They feel they can use LOTE only when it is necessary for learning, when they see that pupils do not understand the lesson in English.

The use of classroom language practices varied very little according to the teaching experience or the age of the respondents. Younger teachers seemed to help learners correct their oral utterances more (statement 2.n, p = 0.002), whereas teaching experience was related to statement 2.(g) I use some words, e.g. greetings, in the learners’ home languages in class to make them feel comfortable. The less teachers had experience in the field, the more they used these expressions (p = 0.002).

Ex. 9 I did not meet er linguistic training … But I had met that one when I did the other course … sociology … Like how to fit in different communities that speak different languages and have different cultural backgrounds. (K, FGD 4, P1)

As excerpt 9 shows, there are elements in teacher training modules that are aimed at raising the cultural and linguistic awareness. This might encourage younger teachers who have done their training recently to use more greetings in pupils’ home languages. Training about multilingual teaching methods seems to be absent, however, a common situation in teacher training in sub-Saharan Africa (Clegg and Afitska 2011).

Partner and group discussions using home languages are a common practice in translanguaging classes (García and Kleyn 2016; Menken and Sánchez 2019). The data show a special use of peer assistance in Namibian classrooms, though this practice was not included in Factor 1 (see Table 4) because of its low communality estimation (0,14). Among the questionnaire respondents, 80% told they ask pupils to translate to their peers often or sometimes (see Table 4). The same practice was mentioned several times in the interviews, and observed in Ugandan classrooms by Altinyelken (2010). Teachers may ask pupils to clarify difficult passages in a text or explain a mathematical problem to their peers who share the same home language. Some teachers allow learners use their home languages in group discussions. New pupils coming from rural schools with no proficiency in English create a challenging situation, as shown in excerpt 10.

Ex. 10 Last term I got … somebody from a rural school. And then when … she came here, she could not speak or understand. Then … I gave her a caretaker, one of the … Herero speaking … like herself. Then she used to to explain to her every time in her own mother tongue. (K, S3, Int.3.)

Assisting newcomers by pairing them with someone sharing the same language background is a common strategy. It helps the new learners keep up with the instruction while acquiring English.

Observed practices

The observed lessons were analysed in order to compare and contrast them with teachers’ self-reported practices discussed in sub-sections 4.1 and 4.2., and to answer research question 2. The analysis revealed both correspondences and discrepancies.

As regards the multilingual practices (see Table 4), the lesson observations corroborate the difference between the two regions concerning teachers’ willingness to explain core terms and concepts in pupils’ home language and to switch the language when they notice that the pupils do not understand (cf. Table 5). These practices were relatively common in Oshana, but very little used in Khomas. In Oshana, teachers in all subject groups used LOTE to explain core concepts, whereas there were no instances of this practice observed in Khomas. Code switching was much more frequent in Oshana, as well. It was used both in content teaching and for so called classroom management, such as giving instructions or disciplining. In Khomas, one teacher switched the code a few times, but mainly to perform the class management tasks or with individual pupils.

The observation data differed however from teachers’ self-reported practices in that mathematics and natural science teachers in Oshana explained core concepts in home language and switched the code even more than the teachers of other subjects. They either gave key terms in Oshiwambo or elicited them from the pupils. They also used Oshiwambo in discussions and explanations. This deviates from the results of the questionnaire and the interviews where teachers of these subjects preferred English as MoI or found it difficult to use the local languages because of the lack of terms in them. It is in line, though, with Ashikuti’s (2019, 118) findings, according to which 65,5% of the teachers in her data believed that Namibian languages can be used to teach mathematics and environmental studies.

There were clear differences between individual teachers in using LOTE in their teaching. In Oshana, two of the three observed natural science teachers included in the analysis used them frequently whereas one did not use them at all. The observations corroborate what she said in the interview (excerpt 11)

Ex. 11 Teacher: Er when I’m teaching I use English yeah I use to try by all means to explain it to them, yeah, to make it easy for them to understand.

Interviewer: And what what languages do you use then? So is it is it only English that you that you use to explain?

Teacher: Oh yeah. (O, S3, Int.3)

As shown in excerpt 11, the teacher says she uses only English as MoI. In her school, the LP had been changed from mother tongue medium to English medium a few years ago. It seems this teacher has internalised the institutional policy and interpreted it as requiring strict monolingual practices, unlike some other teachers in the same school.

As part of the observations were conducted towards the end of the school year, the data contain an example of a teacher mediating to help her students understand the exam questions, similarly to the teacher in Prinsloo and Krause (2019). She both translated some of the English terms in the exam paper and answered pupils’ questions in Oshiwambo to clarify what is required of them.

As regards monolingual practices, the lesson observations corroborate the extensive use of visual support, flash cards, posters, demonstrations, even videos, as reported in the questionnaire responses and the interviews.

Discussion

The purpose of this paper was to explore what self-reported classroom language practices Namibian primary school teachers have and how they are related to certain background factors (research question 1), what observed practices they have and to what extent they correspond to their self-reported practices (research question 2), and what kind of language ecology these practices create (research question 3). To begin with, the majority of the questionnaire respondents (75%) prefer teaching in English. This is in line with Iipinge and Banda (2020, 18) who state that most Namibian teachers prefer teaching in English. This seems to reflect the language hierarchy ideology (cf. Weber and Horner 2012) prevalent in the society that gives English legitimacy (Bourdieu 1991) above the other languages as the only official language and the MoI.

There were, however, other reasons based on the constraints of the teaching context and teachers’ past experiences as students. First, the linguistic diversity in classes was mentioned several times. In senior primary, pupils are usually grouped regardless of their linguistic background, and the more there are different home languages, the more teachers prefer using English. Teachers’ own experiences as students (cf. Alisaari et al. 2019; Borg 2018) and the lack of materials and appropriate vocabulary (especially in mathematics and natural sciences, cf. Kaphesi 2003, 277) were other reasons for preferring English as MoI.

Teachers in Oshana used LOTE more to explain and clarify the content than their colleagues in Khomas. They also made the pupils’ home languages visible in class more often. This difference is probably due to the higher linguistic diversity in Khomas, though there is also a difference between urban and rural schools, the latter being less diverse and their pupils having less contact with English outside school. This is typical for African linguistic ecologies, as pointed out by for example Probyn (2021, 161) as regards South Africa.

Pupils’ age and proficiency in English, as well as the official LP affect the language practices in class (cf. Clegg and Afitska 2011; Iipinge and Banda 2020). In junior primary, using home languages is necessary to support the adaptation and learning of the small pupils, whereas in senior primary, where the official MoI is English, teachers reported using them more in the beginning of the school year than later in grade 4. However, no significant differences between the classroom practices of different subject teachers based on the questionnaire data were found.

Assessment plays an important role in guiding teachers’ decisions about switching the language of instruction (cf. Cleghorn and Rollnick 2002; Shohamy 2006). They emphasise the importance of not overdoing it and the need to prepare the learners for the exams that are all written in English. The same arguments were advanced by teachers in Iipinge and Banda (2020) study by those teachers who did not code switch in their classes. However, it can be inferred that switching the language when it is necessary is nowadays a common and approved practice, at least in primary schools.

Based on questionnaire and interview data on teachers’ practices, it seems that the official policy, pupils’ age and proficiency in English, assessment and the degree of the linguistic diversity of the region affect teachers’ classroom practices the most.

Research question 2 concerned teachers’ observed classroom language practices. They corroborated teachers’ self-reported practices in that multilingual practices, such as code switching and explaining core concepts in home language, were much more frequent in Oshana than in Khomas. It seems that the more homogenous the class is, the easier teachers find using LOTE when teaching. Observation data also corroborated teachers’ self-reported extensive use of visual support, similarly to Altinyelken (2010), regardless of the region or the subject.

The observation data differ however from the self-reported practices as regards the use of LOTE in mathematics and natural science lessons. Whereas teachers expressed having difficulties in using them because of the lack of subject specific terminology in local languages, the observed teachers in Oshana used them extensively. It seems some teachers find a way of using LOTE if they think it may benefit their pupils. Individual differences exist, however, as shown in the data, which corroborates the importance of teachers’ beliefs about LP in guiding their decisions about which languages to use in class (cf. Norro 2021). Addressing LP issues in teacher education is therefore necessary.

The findings reveal a language ecology (research question 3) that is typical for postcolonial school systems (cf. da Costa Cabral 2021). Though the official policy includes several languages, instead of having the same value, English and the local languages constitute a hierarchy where English is on top and the local languages subordinate. The most common reason teachers gave for preferring English was that it is the official language and MoI. In other words, they considered it the only legitimate language for educational purposes after junior primary (cf. Bourdieu 1991). This hierarchical language ideology is maintained by assessment that takes place in English only, though it is possible that teachers use their agency to mediate through multilingual practices even in exams, as shown in the observed lessons of this study, as well as that of Prinsloo and Krause (2019). After junior primary, the use of local languages is legitimated only in situations where it is absolutely necessary for understanding and learning. Teachers’ multilingual practices modify the language ecology towards a somewhat more multilingual direction, but they are constrained by the societal language ideologies and assessment policies.

Conclusion

Against the multilingual reality of the society, the monolingual requirements set on pupils and teachers as regards classroom practices seem odd. Though code switching has been stripped of the prohibitive negative stigma, it is still considered as a disruption of monolingual language use, and the monolingual assessment policy does not allow a constant multilingual language ecology in classes, as teachers feel the need to switch back to English as soon as possible. Legitimising multilingual practices in schools requires a change in the language ideologies in the society in order to remove the constraints set by the monolingual expectations.

Legitimising flexible language practices would benefit learners in multilingual classrooms. Materials and scientific terminology in local languages should also be developed. Translanguaging pedagogy could offer a way of leveraging all students’ languages in multilingual classes, as it can be used even when teachers do not speak all the learners’ languages. The language ecology in schools could be made more multilingual if translanguaging spaces were created and multilingual pedagogies used more systematically. Including pedagogical methodologies that are suitable for L2 medium instruction in both initial and in-service teacher training would contribute to legitimising multilingual pedagogy and practices in Namibian schools.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by Suomen Kulttuurirahasto.

Notes

1 The same questionnaire was used in a previous study on teachers’ beliefs (Norro 2021) with a smaller sample.

2 The Ethics Committee for Human Sciences at the University of Turku, Humanities and Social Sciences Division 5/2019.

Appendices

Appendix 1

QUESTIONNAIRE (abbreviated)

I Background information

Tick the right alternative.

1. Age:  19-–24 yrs 25–-30 yrs 31–40 yrs

       41–50 yrs 51–55 yrs over 55 yrs

2. Years in service:  under 3 yrs 3–5 yrs  6–10 yrs

            11–15 yrs 16–20 yrs over 20 yrs

3. Academic and professional qualifications:

   less than Grade 12

   Grade 12 or 1–2 years’ tertiary

   more than 2 years’ tertiary

   Do you have formal teacher training?  yes no

4. School region:

5. What grade(s) do you teach this academic year?

6. What grade(s) have you taught during the last three years?

7. If you are a subject teacher, what subject(s) do you teach?

8. What language(s) do you most identify with yourself? (What would you call your mother tongue/first language?)

   Khoekhoegowab Ju’/hoansi    Oshiwambo

   Otjiherero     Afrikaans       Silozi

   Setswana     Thimbukushu    Rumanyo

   English      German      Rukwangali

   Portuguese      Other:

   Comments (if any):

   Are you qualified to teach in your mother tongue?  yes no

9. In which language do you find it most comfortable to teach?

   English home language

   Why?

   What are the main home languages of your pupils? (You can tick several boxes.)

   Khoekhoegowab Ju’/hoansi   Oshiwambo

   Otjiherero     Afrikaans       Silozi

   Setswana    Thimbukushu  Rumanyo

   English         German       Rukwangali

   Portuguese  Other:

   Comments (if any):

   What other languages do your pupils speak at home? (You can tick several boxes.)

   Khoekhoegowab Ju’/hoansi   Oshiwambo

   Otjiherero     Afrikaans      Silozi

   Setswana    Thimbukushu  Rumanyo

   English     German      Rukwangali

   Portuguese  Other:

   With whom do they speak these languages?

II Language policy and teaching practices

1. Do you agree or disagree with the following statements? Choose the amount of agreement that best corresponds your opinion.

   1 = I strongly disagree

   2 = I disagree

   3 = I don’t know

   4 = I agree

   5 = I strongly agree

   1. Pupils learn best when they are taught in their home language.   1 2 3 4 5

   2. English should be the medium of instruction from Grade 1.    1 2 3 4 5

   3. Instruction in the home language should be extended beyond

      Grade 3.          1 2 3 4 5

   4. Both English and the home languages should be used as medium

      of instruction throughout the Primary cycle (Grades 1-6)   1 2 3 4 5

   5. The learning outcomes do not depend on the language of instruction.

      Other factors are more decisive.       1 2 3 4 5

   6. My pupils have difficulties in understanding when they are taught

      in English.         1 2 3 4 5

   7. The existence of many languages in Namibia is a problem

      in education.         1 2 3 4 5

   8. Every child has the right to be educated in his/her own language.  1 2 3 4 5

   9. The many different languages of Namibia are a resource in

      education.  1 2 3 4 5

2. In what ways do you facilitate learning in class? Circle often/sometimes/never according to your teaching practices. You can comment on your answers in the end.

   1. I switch the language if I notice that my pupils do not

      understand.        often/sometimes /never

   2. I introduce a new topic in the home language.   often/sometimes/never

   3. I prepare (myself or with the pupils) vocabularies to help

      them understand a new text in English.    often/sometimes /never

   4. I explain the core terms and concepts in the pupils’

      home language.       often/sometimes/never

   5. I let the pupils use their home languages when doing

      group work.       often/sometimes/never

   6. I ask other pupils to translate to their peers.    often/sometimes/never

   7. I use some words, e.g. greetings, in the pupils’

      home languages in class to make them feel comfortable. often/sometimes/never

   8. I make the pupils’ home languages visible in class

      (e.g. posters/word walls/drawings).    often/sometimes/never

   9. I use charts, pictures, drawings and other visual support. often/sometimes/never

   10. I edit English texts by removing extraneous information

       and/or by replacing difficult words by easier ones.   often/sometimes/never

   11. I provide handouts containing some of the language the

       pupils will need when completing the task.    often/sometimes/never

   12. I split new content into smaller chunks.    often/sometimes/never

   13. I ask questions to direct the attention to the essential

       concepts to be learned.      often/sometimes/never

   14. I help my pupils correct their oral utterances by asking

       them to repeat and by providing clues to the correct form.  often/sometimes/never

   15. I reformulate my pupils’ erroneous utterances in correct

       form.         often/sometimes/never

   16. I accept that my pupils use their home languages or

       mix codes in class.       often/sometimes/never

   17. I accept that my pupils use other languages than

       English in assessments.      often/sometimes/never

   18. Can you give examples of your best practices in class? You can also comment on your answers to questions a-q.

III In-service training

1. What in-service training have you attended?

2. What did you find the most valuable in the training(s) you have attended?

3. If you were to attend in-service training in the future, what areas would you find the most important? Tick the five most important.

   • Discipline in class

     Mother-tongue instruction

     Use of different elicitation techniques

     English grammar

     Explaining concepts

     Use of non-verbal support

     Preparing teaching material

     Mastery of subject content

     Use of home languages as resource

     Assessment

     Other

     If you chose “Other”, please specify what:

IV Comments

Is there anything you would like to add or are there any comments you have on this questionnaire?

Thank you for your participation!

Appendix 2

Events numbers:

1. Classroom routines (greetings etc.)

2. Teacher giving instructions

3. Teacher introducing a topic

4. Teacher explaining

5. Use of prompts

6. Vocabulary given to or made by pupils

7. Handouts with some language to help complete the task

8. Pupils reading (silently or aloud)

9. (Creative) writing

10. Copying

11. Individual work

12. Group work/pair work

13. Pupils sharing their ideas or giving examples

14. Teacher asking questions to check understanding or revising

15. Teacher asking questions to orient the attention of the students

16. Pupils asking questions

17. Discussion

18. Teacher translating / code switching / code mixing

19. Peer translation

20. Use of visual support

21. Game, play, other action

22. Correcting language errors (grammar, vocabulary, pronunciation)

23. Pupils repeating (after the teacher)

24. Assessment

25. Maintaining discipline

26. Other