Access to tertiary education and the labour market for the vast majority of SA disadvantaged learners depends on them becoming fluent in a LoLT that is not their L1. This is because such learners have no choice but to write school-leaving examinations in a L2, which is usually English. In this quote, a principal ponders upon the conundrum that this reality poses:
When do I think it is a suitable time for learners to [begin being] be taught in a first additional language [L2]? That is a $100 question [sic]. Because we also [in addition to retaining the L1] need to prepare learners for the outside world. We have got to introduce foreign languages for the outside world. We have got to introduce foreign languages but the problem is when. I don’t have an answer now for that. Ja, I think that that is the biggest problem I always think about (Principal C, principal at FP L1 LoLT school).
It is a problem that the research attempted to solve by responding to the question: What model of LoLT would optimally facilitate learners’ mastery of a L2 LoLT while being conducive to their maximal learning?
To do this, the research explored the impact that LoLT policy had on learner performance, comparing the performance as well as the implementation of two LoLT models while asking a central question: How does LoLT affect learner performance?
The findings generated by asking this question are significant as they indicate a way forward beyond the LoLT predicament faced by SA education. Education in SA has been shown to fail the most disadvantaged of learners and the nature of LoLT policy in the country is believed to have contributed to this, significantly.
How the research responded to the problem in terms of conclusions that were drawn reflecting the learner performance and nature of implementation associated with each LoLT model researched follows.
Research conducted during Part A concluded that:
a) Learning occurs best using L1 LoLT. Grade 3 Xhosa learners in isiXhosa LoLT contexts performed better than Grade 3 Xhosa learners in English LoLT contexts, particularly in Mathematics.
b) The performance of learners in the early-exit transitional model of LoLT (those following the model referred to during the research as FP L1 LoLT) declined over time between Grade 3 and Grade 6, after the point of transition to a L2 LoLT in Grade 4.
c) The performance of learners within schools that were designated as following FP L2 LoLT performed better over time between Grade 3 and Grade 6 compared to the performance of learners in schools designated as following a FP L1 LoLT. This finding must be interpreted in conjunction with findings 5.2.2 (d) and (e), below.
d) Xhosa learners who had had more years of exposure to English as LoLT performed better later than Xhosa learners who had had fewer years’ exposure to the language as LoLT when taught and assessed in it.
Research conducted during Part B concluded that:
a) The formulation of LoLT policy by schools was generally irregular in its process because of the limited extent to which prevailing legislation was followed or relevant theory was consulted by schools.
b) The greatest challenge within both isiXhosa and English LoLT contexts was learners’ inability to understand and use the LoLT in order to learn. Teachers’ inability to understand and use the LoLT in order to teach was also found to be a challenge, but to a lesser extent than was the case for learners. As such, the LoLT was, in effect, a barrier to learning. This was particularly the case for where English was LoLT.
c) In an attempt to overcome the LoLT barrier to learning, teachers and learners within both LoLT contexts commonly employed mixed-language-use practices to communicate during lessons. The use of these practices deviated from the designated LoLT and in a technical sense constituted LoLT non-implementation.
d) The non-implementation of English as LoLT within FP L2 LoLT contexts existed to such a significant extent that the LoLT descriptor designating the LoLT model implemented within these contexts was inaccurate.
e) Schools that were designated as following a FP L2 LoLT model were found to be unintentionally implementing a form of dual-medium (two-way immersion) LoLT model because of the significantly high prevalence of mixed-language use during lessons in those schools.
f) Mixed language use in classrooms predominantly took the form of CSM practices, found present to a pervasive extent within both LoLT contexts – but especially within L2 LoLT contexts. The practices were defining features of these contexts and, as such, needed to be taken into account when adopting LoLT policy for the effect that they had on teaching and learning.
g) The LoLT transition occurring at Grade 4 within FP L1 LoLT schools, following an early-exit transitional model of LoLT, was problematic in terms of a decline in educational outcomes evident from the point of transition onwards.
h) Compared to where L2 LoLT was designated for use in classrooms, teachers and learners were better able to teach and learn using L1 LoLT. As such, teachers and learners were more comfortable in L1 LoLT contexts, and learner participation was higher.
i) The implementation of both English and isiXhosa as LoLT was hindered to a somewhat equal extent by factors existing inside and outside of the classroom.
j) The implementation of prevailing LoLT policy within disadvantaged contexts affected Xhosa learners more negatively than positively, in a pervasive manner.
5.3.1 LoLT policy for SA’s future
From the conclusions generated during the course of this research, the following was apparent as being applicable to disadvantaged bilingual contexts of schooling in SA.
The use of weak bilingual LoLT policy such as the straight-for-English and early-exit transitional models of LoLT are not advisable ways of facilitating L2 LoLT proficiency in learners. In addition to being unconducive to facilitating bilingualism within learners and therefore incongruent with SA constitutional and legislative guiding principles on language in education, research has found that these models typically failed to produce academic L2 proficiency in learners, resulted in poor learner performance, and were generally detrimental to the education and development of learners in various respects. Instead, it is advised that stronger bilingual LoLT policy be followed when determining an appropriate LoLT model for the facilitation of learner L2 LoLT proficiency.
In its characteristically substantial use of both L1 and L2 as LoLTs, one such model option resembles a model that the research found to be associated with superior learner performance when compared to that of the early-exit transitional LoLT model. By definition, this implementation of strong bilingual LoLT policy is referred to as a two-way immersion (dual-medium) model, and involves the use of both L1 and L2 LoLTs in a dual-medium style.
While the research does not recommend that a textbook definitition of dual-medium model of LoLT be implemented within SA disadvantaged contexts of schooling, it does recommend that:
1. Learners should be provided with greater exposure to English as a LoLT before the inevitable transition to the sole use of the language as a LoLT.
This is recommended so as to prepare learners better for the L2 LoLT transition, a transition that research has shown most learners are unable to cope with and never recover from. It was found by the research that Xhosa learners who had had greater exposure to English as a LoLT in Grades 1–3 performed better during assessments written in the language at Grade 6 than learners who had not had this exposure.
2. The inevitable point of transition from L1 to the L2 LoLT should occur much later than is typically the case within the research context, ideally at Grade 7 or 8.
This type of LoLT model is typically referred to as a late-exit transitional model. SA’s history provides an illustration of the positive effect on learner performance that this model’s later transition to L2 LoLT has. The school-leaving examination pass rate for African-language-speaking learners fell drastically from the point where L1 LoLT use was reduced from 8 to 4 years in 1976, that is, from the time when a change was made from late-exit to early-exit LoLT implementation. In this year (1976), the pass rate was 83,7%, while by 1992 it had fallen to 44% (Heugh, 2003).
3. A dual-medium late-exit transitional LoLT model with gradually increasing L2 LoLT/gradually decreasing L1 LoLT use from Grade 1 up to the point of transition should be advocated for. The L2 should then be used as sole LoLT for the rest of schooling toward best preparing learners for school-leaving examinations and tertiary education in the L2. The L1 should be taught as a subject during this time.
The research concluded that this was most likely to be the optimal LoLT policy for disadvantaged SA learners. It is a combination of the only three ways that research has found such learners can successfully develop their L1, learn a L2, and perform satisfactorily in other subjects (Heugh, 2006, p. 64):
a) L1 used as the LoLT throughout school, with L2 taught as a subject;
b) L1 and L2 used as LoLTs side by side throughout school; and
c) L1 used as LoLT and L2 taught as a subject for +/– 8 years, with a very late-exit transitional to L2 as LoLT in the ninth year of schooling.
In accompaniment of this LoLT policy recommendation, several smaller but no less-significant policy recommendations are made towards effectively supporting LoLT policy implementation and learner performance within disadvantaged SA schools. These are discussed below.
5.3.2 Allowing and supporting school LoLT policy implementation
Several peripheral LoLT policy implementation recommendations were indicated by the research as needing to be addressed if LoLT implementation is to achieve optimal educational outcomes. They are presented in order of significance as evident in the findings:
a) The use of isiXhosa as a LoLT must be further promoted in education if the language is to contribute significantly to improving learner performance in parts of SA where the language is L1. The challenges faced by the language as an effective LoLT because of its linguistic nature, a breakdown in the African oral tradition, the scarcity of isiXhosa literature as well as teaching and learning resources, compounded by competition from English, are considerable. Areas in need of prioritisation include further training of teachers in bilingual isiXhosa education and the increased production of isiXhosa literature, including teaching and learning materials, as well as the further funding and celebration of the literary isiXhosa arts.
b) The educationally-sound use of CSM practices should be more strongly included within teacher training curricula. These practices must be considered powerful teaching tools that should be valued and used wisely and only after appropriate training. As defining features of disadvantaged SA classrooms, CSM practices cannot be ignored. The practices are both necessary and beneficial for learning in such contexts, but problematic and damaging to educational outcomes if not used correctly.
c) The extent to which responsibility and accountability for LoLT implementation exists within schools is to strengthen if learner performance improvement is to result from LoLT policy implementation.
d) In order to honour the SA Constitution (RSA, 1996a) as well as adhere to expressions of it in legislation pertaining to the right of parents to choose the LoLT that their children are educated in, SGBs must be more closely monitored for their compliance with the SASA (RSA, 1996b).
e) Initial as well as in-service teacher training must improve both in terms of capacity and quality. Furthermore, greater emphasis should be placed on language education practices within teacher training curricula.
f) Teaching and learning material resource inadequacies within schools should be addressed.
While the LoLT policy problem in SA education is the result of a conundrum of various factors from different realms each with its own effect on learning, both positive and negative, at constant and dynamic play, interacting across time and space, understanding the LoLT policy problem as it exists in SA disadvantaged classrooms for teachers and learners needs not be so complex. It is simply a problem of being able to communicate. In order to teach. In order to learn.