Variations of overt and covert language practices of educators in the North West Province: Case study of the use of Setswana and Sesotho at primary and secondary schools

This article presents overt and covert language practices of educators at primary and secondary schools, with special emphasis on the use of Setswana and Sesotho in three districts in the North West province of South Africa. Field research was conducted during which 990 out of 1 060 questionnaires were collected from participating educators during the third school term of 2013. Based on this broad data collection, the article attempts to identify mismatches or congruences between the overt (official language policy) and covert (daily language practice) of educators at school. Significant regional differences between the researched districts and between the researched language groups became evident. Language practices are determined by the regional circumstances as well as the linguistic setting. Consequently, the use of Setswana and Sesotho differs widely regarding sex, the overall linguistic setting at a school and the general language setting in the researched districts.


Introduction
For many African countries the education systems face a variety of challenges. There are general impediments like a shortage of qualified teachers and school buildings, lack of (quality) school materials, unfavourable student-teacher ratios, and many more. In addition to these serious and difficult challenges there are language-related problems as well. As many African nations are multilingual, the decision regarding the choice of language of instruction and languages to be offered as subjects in the education system remains problematic. The majority of African countries selected their former colonial language as Language of Learning and Teaching (LoLT), despite the fact that as far back as 1960 the United Nations Educational, Scientific and Cultural Organization (UNESCO) already declared in its 'Convention against Discrimination in Education' in Article 5 (c) that: It is essential to recognize the right of members of national minorities to carry on their own educational activities, including the maintenance of schools and, depending on the educational policy of each State, the use or the teaching of their own language, provided however: (i) That this right is not exercised in a manner which prevents the members of these minorities from understanding the culture and language of the community as a whole and from participating in its activities, or which prejudices national sovereignty; (ii) That the standard of education is not lower than the general standard laid down or approved by the competent authorities; and (iii) That attendance at such schools is optional.
Moreover, UNESCO declared the importance and relevance of mother tongue education in different documents, for example in the 'Recommendation of the development of adult education' in 1976 and in the 'Universal declaration on cultural diversity' in 2001, to mention but a few: The expert view is that mother tongue instruction should cover both the teaching of and the teaching through this language…It is obvious yet not generally recognized truism that learning in a language which is not one's own provides a double set of challenges, not only is there the challenge of learning a new language but also that of learning new knowledge contained in that language (UNESCO, 2003). Regarding this international position of UNESCO, the language policy of South Africa seems to be the perfect response, at least on paper. Post-apartheid South Africa is an example of one African multilingual society that recognised eleven official languages in its constitution of 1996 and mentioned many others, too (Constitution, 1996). However, the selection of these eleven languages was partly arbitrary, as other spoken languages in South Africa were not given the same legal status. Nevertheless, these eleven official languages represent the large majority of South Africa's indigenous languages. In addition to this legislative manifestation, speeches of politicians highlight the value and importance of multilingualism for South Africa. As former President, Thabo Mbeki (1999) so aptly described: In the smallest village of our country and in the biggest cities, our people can reflect on the reality of their lives in their mother tongues…Tesame moet ons verseker deur die Afrikaanse taal, soos met al ons ander Suid-Afrikaanse tale, die droom verwesenlik van 'n verenigde Suid-Afrikaanse nasie. 1 For the building blocks of this nation are all our languages This article presents overt and covert language practices of educators at primary and secondary schools, with special emphasis on the use of Setswana and Sesotho in three districts in the North West province of South Africa. Field research was conducted during which 990 out of 1 060 questionnaires were collected from participating educators during the third school term of 2013. Based on this broad data collection, the article attempts to identify mismatches or congruences between the overt (official language policy) and covert (daily language practice) of educators at school. Significant regional differences between the researched districts and between the researched language groups became evident. Language practices are determined by the regional circumstances as well as the linguistic setting. Consequently, the use of Setswana and Sesotho differs widely regarding sex, the overall linguistic setting at a school and the general language setting in the researched districts.
working together, our unique idiomatic expressions that reveal the inner meanings of our experiences. This speech by Mbeki at a meeting of the Afrikaanse Taal en Kultuurvereniging (ATKV, Afrikaans Language and Culture Association) in August 1999 clearly shows the importance and connection of an equal language usage and a national identity. He highlighted that using and sharing each other's languages helps to erase borders between all eleven official languages and its users. Hence, the equal usage of these languages aims to bridge existing gaps resulting from the historical burden of the apartheid era. Similarly, the current President of South Africa, Jacob Zuma, mentioned at the South African Teachers' Union (SADTU) congress in October 2014 the importance of teaching indigenous languages. 'Our children should be proud of our 11 languages and should learn to speak as many as possible…It is up to educators to produce these new South Africans…from the ashes of our tragic past' (Zuma, 2014).
'One nation, many languages-Setshaba se le sengwe sa dipuo tse dintsi'. This slogan of the Pan South African Language Board (PanSALB) highlights the aim of the language policy for the Republic of South Africa (PanSALB, 2009). The South African Schools Act of 1996 gives the school governing bodies (SGBs) the opportunity to define their language policy (Department of Education, 1996). Article 6 entails the legal framework of the language policy and names the eleven official languages and Article 29 focuses on the right to education. Nevertheless, the historical burden of the apartheid era in South Africa remains high. Additionally, the increasing process of internationalisation and globalisation puts more pressure on all languages through the increasing global usage of English as language of economy, education, commerce and media. All this affects the individual language practice and language attitude of educators. Even though this manifestation within the constitution and the SGBs are good examples of an overt language policy, there are many aspects of covert practices at schools too. Despite the abovementioned legal framework, there exist many implicit, latent and unstated practices (Schiffman, 1996: 13;Ndhlovu, 2015: 68). On the one hand this article focuses (mainly) on overt language policies at public South African schools in the North West province. On the other hand the analysis of the responses to the questionnaires offers some insight to the involved educators regarding covert language practices.
The structure of this article is organised as follows: The second section gives details about the methodological approach and the research area. The third section offers a conceptual background of language policy. The focus of the fourth section is on the historical, judicial and societal circumstances of language policy in South Africa. In the fifth section, some results of the potential effects of (covert) language practices on language policy are described and analysed for the North West province. The article ends with a conclusion and an outlook on further research areas.

Methodology and research area
This article is part of an ongoing PhD project in Human Geography entitled 'Potentials of implementation of language policy in the education system in South Africa: A regional focus on Gauteng, Limpopo and North West Province'. A mixed-method approach is used for the PhD project (Tashakkori & Teddlie, 1998;Creswell & Clark, 2007;Atteslander, 2010). This approach offers a chance to get inside information about language attitudes through semi-structured interviews and data about (covert) language practices through questionnaires. This article focuses particularly on some of the results of the quantitative data collection in selected districts in the North West province (Figure 1). Furthermore the language policy documents of the schools were analysed to get information about the overt language policy. Data collection through questionnaires took place between June and September 2013, the third term of the South African school year. The 106 primary and secondary schools involved in the project were randomly selected. The only limitation, for practical reasons, was to select schools that had a minimum of 10 educators, so that at each school up to 10 educators answered the questionnaire. The author collected a total of 990 out of 1 060 possible questionnaires from the 106 schools. The questionnaires were delivered in the three districts of Bojanala, Dr Kenneth Kaunda and Ngaka Modiri Molema. All educators were informed about the aim, structure and content of the research. Additionally, they were briefed that the participation was voluntary and totally anonymous and that all data would be handled confidentially and only used by the researcher himself. To ensure learning and teaching was interrupted as little as possible by this research, the questionnaires were (mainly) collected some days after delivery. Furthermore, the absence of the researcher helped the educators to fill it in at their convenience (Cohen et al., 2011: 77-80). The pitfalls regarding the absence of the researcher was that he could not help if questions caused problems or if questions were misinterpreted or even if questions were filled in by someone else or not during optimal circumstances. To avoid these pitfalls as much as possible the researcher explained the questionnaire in detail to the principal. The researcher left his contact details so that the school would be able to contact him if questions or problems developed (Cohen et al., 2011: 404).
For the preparation of this questionnaire and to get some more in-depth information about language attitude of involved stakeholders, 27 semi-structured interviews were conducted between June and September 2012 (Kvale, 2007: 46;Creswell & Clark, 2009: 207-210;Alvesson, 2011: 52-53). Subsequent to that, a pre-test was done at eight randomly selected schools in Gauteng and the North West province early in 2013. This pre-test helped to adjust and modify some of the questions regarding formulations, the answer categories or the general order of the questions. With the help of the educators' comments or unanswered questions the questionnaire was modified (Oksenberg et al., 1991: 352). Due to the fact that far less than 15% of the educators took much longer than the pre-estimated time and only minor adjustments and changes were necessary, no further pre-test was done (Weinberg, 1983: 332).
The questionnaire consisted of 31 questions divided into three parts, the structure and layout being consistent with relevant literature (Porst, 2008;Raithel, 2008;Tuckman, 2012). The first five questions focused on the general situation and challenges of the South African education system. The aim was to get some information if and to which extent language policy related issues are challenges at the researched schools. The main part of the questionnaire covered various questions about usage of languages, advantages and challenges of using African languages or sundry statements regarding all official languages in South Africa. The last part of the questionnaire consisted of personal questions.

Conceptual framework of language policy
Haugens' (1959) publication, 'Planning for a standard language in modern Norway' is seen as the starting point of a theory of language policy. 2 In the first years the research focused on linguistic aspects. A main focus of research in the 1950s and 1960s was on the development of the standardisation of indigenous languages. From the 1970s onwards the research became more interdisciplinary. The 1970s language policy was no longer seen as a pure pragmatic and neutral decision, but rather as part of interacting social, economic and political decisions. Hence, language policy is no longer only researched by linguists, but also researchers in other disciplines. Today the whole scientific approach is broader and it includes social, historical, economic, judicial, political and geographical aspects as part of research in language policy issues (Ricento, 2006: 10-23). With the help of geographers, the spread of languages can be visualised through various maps on a global and national or regional and even local level (Fouberg et al., 2009: 176). Figure 2 shows the structure of language policy. Language practice, language attitude and language management are areas of language policy (Spolsky, 2004;Orman, 2008). In 1986, Neustupny introduced the term 'language management' to supersede 'language planning'. The idea behind it was to highlight that the whole language policy process is hardly straightforward, but is rather a variety of variables that influence language policy directly and indirectly (Spolsky, 2007: 202). Language practice indicates which language(s) or which variety of a language a person prefers to use in different social settings. Language attitude Acquisition Planning tries to find out why a specific person or specific social groups prefer using a language or some specific variety. For the last 40 years or so research regarding educational contexts about language attitude has grown significantly. There exist studies, for example, about language attitudes of educators, learners, minority groups or migrant learners (Webber, 1979: 217-232). Language management consists of status, corpus and acquisition planning. Status and corpus planning were introduced by Kloss in 1968 (Kloss, 1968: 69-85) and Cooper added the term acquisition planning in 1989 (Cooper, 1989).
Fishman is often described as the founding father of the sociology of languages, as he has analysed how languages interact within various societies (Fishman et al., 1968). Language practices vary between different social situations and even within one specific social situation, depending on who interacts with whom (Fishman, 1970: 68). In his study, Schiffman makes a distinction between the overt language policies in various legal documents and the covert language practices of individuals and groups at each and every institution or situation (Schiffman, 1996;.

Historical, legal and societal framework of language policy in South Africa
During the apartheid era, South Africa had just two official languages, namely Afrikaans and English. As segregation of the society was the key idea of apartheid, this ideology strongly influenced the former South African education system. The segregation was alongside race and/or languages and highly favoured white learners. White learners were taught by better qualified teachers, in better equipped schools and in smaller classes. The learnereducator ratio at white schools was 18.7:1, whereas at black schools it was 41.2:1 (Rehklau, 2013: 307). Furthermore, white learners were mainly taught in their home language(s) (either Afrikaans or English). For black learners these languages were the second, third or even fourth language. 'English and Afrikaans were compulsory subjects for matriculation for all students, and African languages had a subordinate position in the apartheid education system. The situation in schools mirrored the situation in the broader society' (Gilmartin, 2004: 406-407).
Additionally, the 'Bantu Education Act' of 1953 strongly influenced mother tongue education ('moedertaalonderwys') up to today. This idea of mother tongue education was neither based on pedagogical nor neutral assumptions, but was purely ideologically motivated. The random division of the indigenous languages (Kaschula, 1999: 66;Chebanne, 2005: 162;Ndhlovu, 2008: 65-66) and the usage of indigenous languages as LoLT was to steady a separate development. The curriculum was only to prepare black students for their subservient role in South Africa (Heugh, 1999: 302).
As mentioned earlier the status planning in South Africa specified eleven languages as official languages in Article 6(1) in the South African Constitution. The selection and announcement of the eleven languages caused some criticism and debates, specifically regarding Sepedi. The 1993 Interim Constitution of South Africa mentions (in Article 3 (1)) eleven official languages, among which was Sesotho sa Leboa 3 (Constitution, 1993), which is Northern Sesotho.
In the current, legally valid Constitution of 1996, Sesotho sa Leboa was renamed Sepedi (Constitution, 1996). This caused some criticism from the Balovedu community who speak Khelovedu, a dialect of Sesotho sa Leboa. They demanded a more impartial term covering many speech varieties without privileging one. The Khelovedu speakers thus protested against the renaming of Sesotho sa Leboa as Sepedi. Nowadays they demand that Khelovedu should be recognised as another official language 4 (Constitutional and Legal Services Office, 2012). Hence, to avoid a new discussion about Article 6 a tacit compromise and consent was found by PanSALB. They agreed thenceforth to use only Sesotho sa Leboa as a term and never Sepedi (Perry, 2004: 514), however Sepedi remains in the Constitution. Nevertheless, the majority of its speakers refer to Sepedi and hardly name the language Sesotho sa Leboa. Hence, the naming of Northern Sotho is a good example of the mismatch between the governmental language policy statements and the perspective of its speakers. Regarding the situation at schools in South Africa, the South African Schools Act of 1996 regulated the language policy through the SGBs. Every SGB has the chance and duty to select which language(s) is/are used as medium(s) of instruction, as a subject and for communication with parents within the given judicial framework. Earlier research done in Gauteng and North West showed that the principals' language attitude had sometimes quite a significant influence on the decisions of the SGBs. They could not support or diminish the use of African languages without the principal's considerations (Kretzer, 2013: 223-225).
Apart from this judicial manifestation, there are many other language-directed initiatives and policy documents with a specific focus on the further development and promotion of African languages, for example PanSALB, the Language Task Group (LANGTAG), the Gerwel Committee Report (Gerwel, 2002) and the Soudien Report (Webb, 2013: 176). The Gerwel Committee Report of 2002 highlights the importance of multilingualism and equity of all official South African languages, but also concentrates on the importance of English. This twofold or ambivalent position is visible throughout the whole document. There are many positive statements regarding multilingualism and an equal support of all eleven official languages. But otherwise it sees 'South Africa as a leading country within the Anglophone world'. Additionally, English is defined as a 'major binding language among South Africans' (Gerwel, 2002). Other governmental documents deal with the nearly invisible usage of African languages as LoLT at Higher Education Institutions (HEI) (Department of Education, 2008: 94). It clearly states that even after many initiatives and policy documents the full and equal implementation of multilingualism at HEI remains symbolic. The Soudien Report points to a very interesting fact as it describes the influence of individual language attitudes of students in clear favour of English (HESA, 2010: 30-31). Despite these numerous policy initiatives, societal language attitudes greatly affect the language policy in South Africa in particular regarding the African languages. Some initiatives as mentioned earlier are in themselves ambivalent regarding African languages. Ndhlovu (2008: 65) states: Therefore, unlike in the apartheid era where racial or ethnic difference was the underpinning ideology of linguistic imperialism and cultural domination, the hidden language policy of post-apartheid South Africa is premised on perceived social or political class differences. The ruling political elites and the intellectuals appear to have an insatiable appetite for the English language to a point where the African official languages have been essentialised as identity markers for the less educated subaltern. Additionally, it is not just a general and vague societal language attitude in which English is preferred as an indicator of education and a sign of being part of the elite, but rather a component of the ruling African National Congress (ANC) itself. Due to various reasons the majority of the ANC prefers English and is not really supportive towards the mission of PanSALB. They see English as a non-ethnic and neutral language 5 , and supporting African languages would lead to segregation alongside these language groups. 'Indeed, the only party to have shown itself hostile to PanSALB is the ANC. This comports with the ANC's historical preference for English. Since its founding, the ANC has regarded "minority" languages as indicators of divisive tribalism' (Perry, 2004: 511).
Language policy is always highly political and, especially in the case of South Africa, it was and mainly still is one of the characteristics to indicate power. Furthermore, language policy as part of the Bantu Education policy was a sign of separation and domination (Setati, 2008: 104). Therefore all new political regulations, agendas or initiatives focused mostly on equity. The Ministry of Education, as consistent with many other African and global trends, saw a direct correlation between the establishment of equity in their education system and the improvement and securing of quality as a result (Sayed & Ahmed, 2011: 107;HESA, 2014). Many international evaluation studies and much academic research highlighted the still existing correlation between socio-economic framework conditions and educational outcomes of South African schools (Fiske & Ladd, 2004;SACMEQ, 2010: 23-44;Van der Berg, 2008: 153). The abovementioned position of the ANC and the overall societal scepticism towards African languages and multilingualism in general reflect the idea of a close nexus between language uniformity and national development (Pool, 1972: 213). This idea from Pool is still relevant for political decisions, even if many other authors highlighted how important the use of mother tongue is for a qualitative and meaningful education in South Africa or all over Africa (Brock-Utne, 2001: 120;Ndhlovu, 2015: 169).

The potential influence of language practice and attitudes for the use of Setswana and Sesotho in the North-West province
As mentioned earlier the individual's language attitude and language practice can influence the implementation of language policy in schools. Therefore educators can be language policy actors for a more intensive use of African languages. Educators' daily (covert) language practice can be congruent and consistent with the overt language policy of the school, but it is not everywhere and always the case. Their language practice within the classroom and the general school environment can affect the language attitude of other educators, learners and parents. 'Teachers of nondominant languages also play a critical role, however, in the broader context of the school (i.e., the space outside of the classroom)…As policy actors, they simultaneously reproduce and challenge existing language ideologies in the school environment' (Brown, 2010: 298).
The aim of this section is to give some results of the data collection during the third term of 2013 in the North West province. In this research, 990 educators took part; 73% were females and 27% male. This gender segregation differs considerably between the different education phases. In 'Foundation Phase' just 2.5% were male educators, but this figure increased significantly through the phases. In 'Intermediate Phase' still just 27.4% were male educators, but in 'Senior Phase' (40.7%) and in the 'Further Education and Training' (FET) band (46.9%), it nearly equalled the female counterpart. A very high response rate of around 93% showed keen interest and high motivation to complete the questionnaire as the participation was voluntary and unpaid. Furthermore, only 3% did not finish their questionnaire, while around 40% wrote comments related to the questionnaire.
Regarding the use of Setswana and Sesotho in the North West province there are significant differences. Setswana is used in 70% of the involved schools as LoLT or as a subject or offered as both. In contrast, Sesotho is only offered at 7.7% of the researched schools. Very clear regional concentrations became obvious. Only at one school in Ngaka Modiri Molema district is Sesotho offered as LoLT, and one school in Bojanala district offers Sesotho as LoLT and as a subject. All other schools which use Sesotho in their overt language policy are based in Dr Kenneth Kaunda district. Only in Dr Kenneth Kaunda district less Setswana is offered as LoLT or as a subject. These results are not surprising as this represents the linguistic setting of the province. Nevertheless, Setswana is offered in all districts as the main language. The partly diminished use of Setswana in Dr Kenneth Kaunda can be explained by the high usage of Afrikaans in schools in the southern district of the North West province compared to the other two researched districts. Nevertheless, Setswana is used in all researched districts. Another interesting difference between these three districts is the fact that Setswana is more often only offered as a subject in Dr Kenneth Kaunda. The analysis of the language policy documents at these schools further showed that many schools that did not have a written language policy document were currently re-drafting it or had misplaced it.
Coming to aspects of language attitude, some very interesting findings emerged. In the questionnaire, educators were asked: 'Until which grade (including) do you think should the home language of the learners be the medium of instruction?' 990 educators took part in the survey, but not everyone answered all questions or some questions were only for specific educators. This explains the huge variations between the responses to the questions posed to the different language groups. The 301 educators who work at a school where Afrikaans is offered as LoLT, as a subject or as both in the overt language policy had quite a positive language attitude towards Afrikaans. The majority (38.5%) advocated a usage of Afrikaans until Grade 12. Another peak was at Grade 3 (17.3%), and 11.3% wished for no use of Afrikaans as LoLT at all. The situation was nearly the same for the 76 educators who work at a school where Sesotho is used in the overt language policy. With 34.2%, a slightly smaller number of educators preferred the use of Sesotho throughout the whole education system, whereas the figures for Grade 3 (15.8%) and the non-use of Sesotho (14.5%) differed slightly too. It is very important to keep in mind that only 7.7% or 76 educators work at a school where Sesotho is offered as LoLT or as a subject, so that these statistics rely on only a few schools and educators. Concerning Setswana, the results look much more heterogeneous and not so much in favour of a longer usage of Setswana at primary and secondary schools. 678 educators worked at schools where Setswana is either offered as LoLT or as a subject. Only 24.5% preferred the use of Setswana for the entire schooling until Grade 12, and 26% were in favour of the usage of Setswana as LoLT until the end of Grade 3. Together with the educators who wished for no usage or a usage of Setswana as LoLT until Grade 1 or Grade 2, all in all 51.8% wanted only the usage of Setswana as LoLT until the end of 'Foundation Phase'. This means more than 50% who work at a school where Setswana is offered want to keep the status quo or even less Setswana. In contrast, only around 30% of educators at an Afrikaans medium school answered the same way. These are quite interesting results, giving a glimpse of the language attitude towards language(s) used at schools of the involved educators.
Regarding the gender analysis, no significant differences were observed. All in all, 945 educators (690 females and 255 males) answered and the findings are reflected in Table  1. The differences are negligible as they vary between only 1 or 2% up to 4% between the various grades. The only slightly higher difference in responses between female and male educators is observed in the answer categories of Grade 3, Grade 7 and Grade 12. Male educators responded more often for Grades 5, 10 or 11 (see Table 1). Hence, the answers of female educators peak more at the end of specific periods of learners, like the Foundation Phase (Grade 3), end of primary school (Grade 7) or end of schooling (Grade 12), unlike their male colleagues.
These results were not surprising as some general studies have shown that educators claim to use only English as LoLT, yet they also use the indigenous languages for oral communication. However, all written communication, either on the blackboard or in textbooks, takes place in English, as the bulk of materials is only written in English (Nadeau, 2009: 65). Therefore further questions were asked focusing on language practice(s) and language attitude(s).
In addition to the aforementioned questions about the overt language policy at the schools and to get a deeper impression of the language attitude and language practice of the involved educators, more questions focused on the real classroom situation. Some of the questions were related more to the covert language policy. This means the questionnaire tried to find out more about specific communication situations as they differ widely between different social contexts or the purpose of the communication. Hence, three different questions in the questionnaire asked about language practices at classroom level. Furthermore, the questions could help to get a deeper insight on a larger scale into how code-switching is used in classrooms in the North West province. However, educators do not only teach through code-switching, but rather use some kind of code-mixing, which means they speak two or more languages within the same sentence (Ndhlovu, 2015: 168). Research about code-switching in the classroom situation mainly focuses on mathematics lessons (Setati & Adler, 2000: 243-269), whereas this article gives an overview of code-switching and code-mixing in general and throughout all phases.
Therefore the three questions focused on different social situations at schools. The first one asked about language(s) used in oral communication in a lesson, the second one enquired about written communication and the third one about the communication with the learner's parents. Regarding the usage of Afrikaans for oral communication just 17.8% responded with 'all the time', whereas 55.8% responded with 'never' and another 20.3% used Afrikaans 'sometimes'. Compared with the overt language policyregarding LoLT and languages offered as subjects-these numbers are more or less congruent. An analysis of Sesotho showed totally different numbers as just 3.2% used it 'all the time' and 69.8% never spoke Sesotho in a lesson and another 22.9% used it just 'sometimes'. Regarding the categories 'all the time' and 'never', this covert language policy was also congruent with the overt and official language policies at researched schools as only few schools offer Sesotho due to the linguistic composition of the schools. The situation for Setswana is quite diverse and complex. The majority of educators (38.3%) responded with 'all the time' and just 15.1% selected 'never' and another 21.5% 'sometimes'. These results are still quite consistent with the overt language policy documents of the researched schools. The main difference between the usage of Setswana on the one side, and Sesotho and Afrikaans on the other, was therefore that over 25% of the educators used Setswana 'regularly' and 'often' within the classrooms. This shows that the real use of Setswana regarding oral instructions or explanations was much higher than the official overt language policy at primary and secondary public schools. In this context there are very significant differences between female and male educators. 277 female educators used Setswana 'all the time' for oral communication (44.7%), whereas only 43 of their male counterparts used it 'all the time' (19.2%) as shown in Table  2. These differences do not appear in the case of Sesotho at all as eight male educators used Sesotho 'all the time' (4.1%), which is slightly more often (see Table 3) than their When it comes to the usage of the above analysed languages in written as opposed to oral communication the whole situation appears totally different for Sesotho and Setswana on the one hand and Afrikaans on the other. In written communication Afrikaans was used by 18% of educators 'all the time', so it remained stable compared to the oral communication. The only change was that only 10% used it 'sometimes' and a sharp increase to 66.9% appeared for educators who never used Afrikaans. The situation was partly similar for Sesotho and Setswana. More or less the same number of educators used Sesotho (2.7%) and Setswana (39.3%) 'all the time' in written communication, as they did in oral communication. The main difference between oral and written communication was the significant shift from the categories 'often ', 'regularly' and 'sometimes' to 'never' in the case of written communication. This leads to a sharp increase to 87.8% for Sesotho, and for Setswana to 36.9% for educators who 'never' used them for written communication.
In the case of Setswana, 39.3% used it always, whereas 36.9% never wrote anything in Setswana on the blackboard or in textbooks. During oral communication, only 15.1% never spoke Setswana to the learners. As mentioned earlier in written communication, the female educators used Setswana far more often (44.8%) 'all the time' than their male colleagues (22.8%). This difference was non-existent for Sesotho and Afrikaans. All in all this change between both spheres of communication is not surprising as English and Afrikaans educators can choose between a wide array of different kinds of publications, whereas these are much more limited for Sesotho and Setswana. Perry (2008:65) explains that: Providing texts of any sort in each of the languages present in most of the southern African nations appears to be a financial and logistical nightmare. However, like school conditions, textbook availability cannot be seen as a merely structural problem; it is one that is closely connected to socio-economic and political issues. The analysis of the communication with the learner's parents showed the heterogeneous linguistic setting of the schools involved, which is much more complex than any analysis of the official and overt language policy documents of the schools would assume. Furthermore, only a few language policy documents mentioned a specific language for communication with the learner's parents, whereas the majority only specified the LoLT, the taught subjects or the overall languages used. English (21.8%) and Afrikaans (15.6%) were only used sparsely 'all the time' and mainly only used 'sometimes' (36.4%) or in the case of Afrikaans mainly (62.5%) 'never' used. The situation for Setswana appeared totally different. The majority of educators (58.6%) spoke Setswana throughout the whole communication with the parents and only a small number (16.6%) 'never' used Setswana. A similar shift to a more intense usage applied for Sesotho, although on a significantly smaller scale, as still only 6.5% used Sesotho during meetings and communications with the learners' parents. As indicated earlier, Setswana was more frequently spoken by female educators with the learners' parents. These differences are less significant than within the classroom situation. 61.7% female educators and 48.9% male educators used Setswana 'all the time'. The situation for Sesotho and Afrikaans remained more or less the same.
In this context a regional analysis is very interesting. Afrikaans was mainly used in Dr Kenneth Kaunda district and hardly in Bojanala district. This language practice appeared with minor changes in all three communication spheres (oral, written, with learner's parents). This is not very surprising as the linguistic composition between Dr Kenneth Kaunda and Bojanala differs widely. In the Bojanala and Ngaka Modiri Molema district educators only used sporadic Sesotho 'all    Total  Male  108  56  5  7  26  202  Female  322  100  13  23  112  570  Total  430  156  18  30  138  772 the time', and its usage was rather much more concentrated in the Dr Kenneth Kaunda district. Setswana was used mainly in Bojanala and Ngaka Modiri Molema districts and to a lesser extent in Dr Kenneth Kaunda.

Conclusion
The main objective of this research article was to describe and analyse the overt language policy and the (covert) language practice of educators at primary and secondary public schools in the North West province. The majority of past research indicated the non-equal use of English and Afrikaans on the one hand and the African languages on the other. These results are partly confirmed, but also partly refuted. The three main findings of this article are that there are differences between the overt and covert language policy at public schools in the North West province; the language practices of educators are much more complex and heterogeneous than the official language policy of a specific school and there are significant regional differences between the three researched districts regarding language practice in the North West province.
To analyse the language policy at public schools in the North West province it is necessary to look deeper and not just to collect and analyse the official language policy documents at schools or to research selectively few schools. Rather it is important to involve many educators to get a deeper understanding of daily language practices at schools. Hence, broad data collection enables the opportunity to see a gap or mismatch between the overt and official language policy and the daily (covert) language practices of educators at schools, or if these two are always congruent with each other. In this context, it is also important to differentiate between different spheres and areas of communication. The research made differences visible regarding oral and written communication or different areas of communication, either the classroom situation or the contact with the learners' parents. In addition, significant regional differences became obvious. Schools in the Dr Kenneth Kaunda district are more multilingual and heterogeneous than in Bojanala and Ngaka Modiri Molema districts. Afrikaans and Sesotho are much more widely used by educators in the southern district. Nevertheless, Setswana is the dominant language at schools in all three districts.
As mentioned before, this article gives a glimpse of the questionnaire used for an ongoing PhD project, which covers Gauteng and Limpopo province as well. Additional quantitative research data collected in 2014 in Gauteng and in Limpopo in 2015 will provide more detailed information about the differences between the regional linguistic settings and the involved (African) languages. Further data collection and deeper analysis will help to get more information about the regional variations and the influence that age, sex, work experience, language groups and the phase or the subject an educator is teaching may have on individual language attitude and practice. This offers the opportunity to get a broad picture of parallels and differences regarding the overt language policy and the (covert) language practice between three very different provinces in South Africa. Hence, further research should include the learners and parents too. This expands the research to non-formal settings and limits it not to just the formal setting at schools within classrooms. Research of language usage of the learners during recess and/or the language use of parents with their children would help to broaden the understanding of the language usage of Setswana and Sesotho in a more holistic approach within educational institutions.