Exploring linguistic thresholds and reading comprehension and skills-transfer in a grade 6, isiXhosa-English additive bilingual context

Reading is the key to knowledge and learning and by implication, life success. Most South African children „learn to read‟ in their home languages (HL), such as isiXhosa in the Eastern Cape, and then at the beginning of Grade 4 are expected to make two significant transitions: they must begin to „read to learn‟ and they must do so in an additional language (usually English). The research evidence is damning: Intermediate Phase children are failing to read and failing to learn. This study is concerned with two of the possible, and often conflicting, reasons for the reading problem: 1) that too little time is spent developing learners‟ English language proficiency and 2) that the development of learners‟ reading comprehension skills in the HL is neglected, preventing the transfer of skills to reading in English additional language (EAL). This thesis explores the relations between English Language Proficiency (ELP) and isiXhosa Reading Comprehension (XRC), and between ELP and English Reading Comprehension (ERC), in a unique, additive bilingual context in the rural Eastern Cape, where isiXhosa is maintained as part-LoLT (language of learning and teaching) to the end of Grade 6. The Linguistic Threshold and Linguistic Interdependence Hypotheses constitute the theoretical framework of the study. The design of the research is exploratory and descriptive. The Woodcock-Muñoz Language Survey was used to measure the language proficiency (English relative to isiXhosa) of the sixteen Grade 6 learners in the study, while two sample, expository passages from the Progress in International Reading Literacy Study (2006) were used to measure the reading comprehension abilities of learners, in both isiXhosa and English. A questionnaire provided additional information – about the learners‟ perceptions of reading– which assisted in the interpretation of the statistical data. „Mean scores‟ and „standard deviations‟ were used to describe the ELP (relative to the isiXhosa language proficiency) of the participants, while „frequency‟ was used to describe the reading comprehension scores. Correlational statistics were then employed to test the strength of the relationships between the variables, while regression analyses were used to predict the relative contribution of each of ELP and XRC to ERC. The study reveals that while the learners‟ isiXhosa language proficiency far exceeded their English language abilities, their reading comprehension scores in both languages were equally poor. ELP correlated significantly with ERC; and XRC and ERC were also covaried, thus corroborating the findings of international research: that in this particular context, second language (L2) reading is a consequence of both ELP and first language(L1) reading ability. The regression analyses showed that while the potential for reading comprehension transfer in the direction L1 to L2 existed, this possibility was short circuited, both by learners‟ poor ELP and their poor L1 reading skills