Thesis (Ph.D. (Educational Psychology)) --University of Limpopo, 2012 / This research project was an endeavor to investigate the effects of the languages of instruction (English and Shona), to teach Tsonga (Shangani) speaking children in Chiredzi district of Zimbabwe. Because of the nature of the study, a mixed method design was used where both qualitative and quantitative methods were adopted to study the performance of the Tsonga (Shangani) minority language speaking learners in five purposively sampled schools. 222 learners participated in the study. The main aim of the study was to investigate the effect of the language instruction in teaching Tsonga (Shangani) speaking Grade Seven children in Environmental Science. In fact, the researcher was interested in finding out whether teaching learners in a foreign language was a bridge or barrier to learning. In this case, the research did not only look at the effect of using English for instructional purposes, but also investigated how other major or dominant indigenous languages which are used for instructional purposes affect the performance of minority language speaking children in Chiredzi district of Zimbabwe. Data for this study were collected using lesson observation, document analysis, the questionnaire and a knowledge test. In this case, fifteen lessons were observed. Fifteen Tsonga (Shangani) speaking Grade Seven learners per school were purposively selected and taught in Tsonga (Shangani) only and the other fifteen Shona speaking Grade Seven children per school were also purposively selected and taught the same topic in Shona, and a third group of fifteen Grade Seven learners per school, were randomly selected and taught in English only. A knowledge test was given to each group thereafter. Children from each language condition were allowed to answer questions in their home languages, except for the third group which was taught in English. This group answered the questions in English with the restricted use of Shona. Each of the test results from the knowledge tests were analysed using a One Way Anova of Variance (ANOVA) and conclusions drawn. The results from other data collection instruments were analysed using qualitative methods like narrative discussions of data. A sample of five learners per school had their exercise books analysed. Data were presented in tables. The results from the knowledge tests given showed a significant difference in the mean marks obtained from the three groups (the Shangani, Shona and English group). The result showed that language has a significant influence on the performance of learners since the p – value was 0.000. This implies that the performance of learners between the three groups is significantly different. On the basis of these observations, the Null hypothesis was rejected. The same picture was also shown in document analysis and in the questionnaires. Consequently, conclusions were drawn and recommendations made.