Development of Balanced Reading in Sepedi and English of Grade 7 Limpopo Readers Pheladi Florina Fakude

Reading in more than one language remains a debatable issue in literacy pedagogy. The present study aims to illuminates this debate by measuring, and comparing learners’ writing, resolving anaphors, and making inference on text using both Sepedi and English among grade 7 readers in three randomly selected schools in rural Limpopo. Two schools speaking the same dialect of Semmamabolo (language closer to standard variety of Sepedi) were selected randomly from Mankweng circuit. One school was conveniently selected from Tzaneen to serve as a pseudo group to measure dialectal influence for reading performance of the learners for both Semmamabolo and Selobdu dialects. The dialectal schools were compared to measure if there were differences in performance and Selobedu School served as a psedo-school, to determine the dialectical difference. The learners were tested on reading comprehension with the use of making inference test and anaphoric resolution test. A onetime series design was used to study and compare the two languages and dialect reading scores within and between subjects. The sample consisted of n=150 participants for the tests in both English and Sepedi. Analysis of Variance (ANOVA) results reveal low reading levels in both the participants’ home language and English with scores of (-+ 40%) below 75% and no differential dialect performances within Sepedi. The results indicate that learners are not about the same level in their reading comprehension skills in both the languages. These results are interpreted within various theoretical frameworks on the relationship between L1 and L2 reading development.


Introduction
Reading comprehension levels have been reported very low by previous researchers (Fakude, 2014;Makalela, 2010;2012;Makalela and Fakude, 2014;PIRL, 2006;Pretorius & Lephalala, 2011;Phokungwana, 2012;Pretorius and Mampuru, 2007,).Of particular significance has been the phenomenon of the transfer of reading skills from one language to another.Very little, however, is known about the reading development among primary school learners in many African countries.Within the field of literacy the assumption about the significant transfer of reading skills from one language to another language remains a debatable issue in literacy pedagogy apart from what is said by Cummins (2000).Reading in more than one language remains a significant phenomenon for transferring reading skills across L1 to L2. Reading comprehension of both L1 and L2 has become the most important issue throughout the country and beyond.Studies conducted in Africa that focused on three local languages and English in Eritrea brought fourth suggestions that there is higher reading comprehension variance in home language than in English.L1 reading comprehension results showed significant results while the script-based difference of L2 language proficiency, L2 reading comprehension, and L1 wordreading results were not significant ( Paran & Williams, 2007).From this study the results revealed that learners were not able to read and more variance was observed in home language than second language.The results depart from the South African perspective by Pretorius & Currin, (2010) who examined the reading levels of Grade 7 both in L1 and L2.
Their study looked at correlations and assessment of English and L1 (home language).The results of the study showed more variance between the two languages with Northern Sotho comprehension performing below 40% and English 47.8%.More variance was observed in L2 than L1 in the study conducted by Pretorius & Mampuru, (2007).
From a theoretical proposition of Linguistic Interdependence Hypothesis, it is expected that skills acquired in one language are transferable to another language and the preferred order is L1 skills transferring to L2.The results of empirical data, however differ remarkably among African researchers and less so in the developed countries.Very little, however, is known about reading development among primary school learners in many African countries.Therefore; the present study aimed to illuminate this debate by measuring and comparing reading strategies used in Sepedi and English among grade 7 readers in three randomly selected schools in rural Limpopo.primary languages.Pretorius identifies anaphoric resolution "referring back phenomena" as a tool to assist learners or students with comprehension when reading study materials.According to Pretorius reading is not an integral part of many people's daily life.Reading is therefore associated with a formal study setup and it especially happens in L2, the language of schooling.In this study the research focuses on anaphoric resolution as a point of departure.The study investigated anaphoric resolution by first year English as second language students during the reading of expository texts.The study examined more closely the anaphoric inferencing abilities of English as a second language (ESL) reading of expository texts.This was set up to determine whether there was a relationship between anaphoric resolution and academic performance.The subjects were 68 first year students with a mean of 18.9 years who wrote 30 paragraphs that had a total of 38 different anaphoric ties.The study suggests a strong relationship between anaphoric resolution and academic performance, albeit a few small groups of students showed mastery performance in anaphoric inferencing with distinctions.
On contrary, a study conducted by Cramer in (1971) states that the incident of reading failure is higher among children whose dialect is significantly divergent from Standard English.In this case reading failure among nonstandard speakers may be generated by the mismatch between the dialect of the learner and the dialect of materials reading instruction.If this is the case then there is an instructional problem rather than language problem.In his study he suggested three distinct teaching alternatives that could assist learners in developing their reading.The first aspect he suggested is (1) Writing instructional materials in the child's language or the dialect of the individual learner.This is not practical in Northern Sotho since there are over 27 dialects that are known to date (Mojela, 2008).Cramer also questions the matter of practicality.The second aspect is (2) Teach the child to speak the standard dialect.After he has learned to speak standard dialect, reading instruction can be started with standard materials.The third and final aspect is (3) Using materials written in standard dialect, but allow the child to translate the standard language into his dialect without penalizing him by considering differences in pronunciation and syntax as errors.All these three aspects cannot hold for South African learners especially Northern Sotho speakers with over 27 dialects.Another challenge is that some of these dialects are not alphabetized and have no lexicography.What can work for the learners is for the educators to re-design reading instruction to fit the kind of learners that they receive at schools.
With regard to teaching, Georgiana (2012) agrees that teaching English cross-lingual as a learning aid was disapproved in favour of intra-lingual approach.She identified problems that come with this teaching method of using Mother Tongue in an English language environment.Mother tongue is a relevant tool to be used but more English should be used in the classroom.From the above study it can be suggested that in an environment where learners' dialect is divergent from standard language instruction learners are faced with a situation where they use their home dialect to cope with the learning environment.She further states that using L1 is not a problem but the problem is when and how you use it in the classroom and English must be used throughout and L1 must be used when necessary.In this study learners come to the classroom with their own home dialect and they find themselves in a position where they have to learn a standard language.For them, it's like learning a second language; therefore English becomes their third language.According to the results of this study learners who spoke a language that is divergent from the standard variety performed higher than those that had an advantage of learning in a language that is of their mother tongue.These learners are in a position juggle between their dialect and the standard variety.

Aims
The study was guided by the following research questions; (1) is the observed relationship between L1 and L2 for each school with regard to anaphoric resolution test scores significant?, Alternative Hypothesis-The observed relationship between L1 and L2 for each school with regard to anaphoric resolution test scores will be statistically significant.(2What are the Grade 7 readers' inferencing skills proficiency?(3) Is the standard variety form influenced and affected by spoken language in School B "Selobedu" dialect?, Alternative Hypothesis-The written language form will be influenced and affected by spoken language in School B "Selobedu" dialect, (4) Are there correlations between variance of inference skills in L1 and L2 and are they statistically significant?

Research Design
A descriptive study was utilized to define mean and standard deviation for the study.A correlation design was also employed to determine significance between Sepedi scores in School B and School C for the dialect factor.In addition, ISSN 2039-9340 (print) Mediterranean Journal of Social Sciences MCSER Publishing, Rome-Italy Vol 5 No 23 November 2014 954 three schools were compared to each other on two variables, namely, Sepedi and English by using a paired-sample t test.Lastly, the Sepedi scores were compared within School B and C to examine if dialect spoken in School B affects learners performance during inference test.

Participants
The participants were Grade 7 second language English speakers, with Sepedi, the most commonly spoken official language and a predominant first language in Limpopo.According to the information obtained from the teachers, all learners who participated in the study had Sepedi as their home language.Grade 7 was chosen as it represents the final stage of primary school and this stage will give a great overview of reading development levels before learners move to high school to ensure or to examine if they will cope well with the subject content presented at high school or not.At this stage the learners would have had at least 3 year learning their subjects through the medium of English.
The participants were selected irrespective of their academic performance.Information in terms of learners' impairment was not obtained from the educators and they also did not bring that to attention.This kind of procedure was followed in order to contain a various kinds of learners that are in our mainstream school.Participants were selected from 3 schools in two geographic contexts of Limpopo province [Polokwane (two schools) and Tzaneen (one school)] representing three educational contexts found in South Africa.That said, the selected schools were what Nkabinde (1997) categorised as school directly under state control.The sample consisted of 150 learners whose mean age was 12.7 years at the time of investigation.

Description of learning contexts
The school varied in terms of socioeconomic status and included a representative sample of the population of Grade 7 learners.A total of 150 (n=50) learners (mean age of 12.7 years) were selected from two schools in Polokwane and one in Tzaneen.These learners came from a variety of residential areas, including; Makanye village, Mamotintane, Iraq and Mankweng (Unit A-F) for Polokwane and New-Rita village, Rita village, Ramalema village, Myakayaka, Pulaneng, Lenyenye, Mohlatlareng, and Sunnyside.Sepedi was the only language recognised as the most predominantly spoken at the areas of investigation.The teachers were predominantly bilingual or multilingual with Sepedi as their first language.The language of instruction in all the schools was Sepedi from Grade R to Grade 3 and English from Grade 4. All the schools were generally not well-resourced with materials with respect to either English or Sepedi materials, expect for School A with exposure to English materials.

The educational contexts are:
School A: a peri-rural school (Mankweng Unit A) where both L1 and L2 is used and well-supported at school with English given more attention because of resources availability.
School C: a rural school (Makanye Village) where both L1 and L2 is also used and well-supported, although the learners are more exposed to Sepedi than English with lack of resources in both languages.The teachers at this school are either bilinguals or multilinguals because of the nature of the environment, while School B: a rural school (New Rita Village), this school is the same context with School C, except that at this school learners are exposed to a third language that was not tested.Learners at this school spoke a dialect called "Seroka-Sekgaga" which is predominantly spoken at New Rita Village and it is more influenced with "Selobedu".According to Mojela (2008), this dialect contributes 2% to the standard variety of Sepedi.

Procedure
The learners were assessed using a self-developed one-time series designed test to measure reading comprehension with the use of a text that demanded their knowledge of anaphoric resolution and inference skills.To ensure that learners were given a fair chance in the anaphoric resolution testing, the test was divided into two sections; fill in and multiple choice questions and both sections had 10 questions.The text was extracted from the learner's textbook prescribed for Grade 7 English as Second Language (ESL).The inference test had learners read a passage of 400 words and write a paragraph on what they could have done if they were in the shoes of one of the characters.

Test administration
The tests were administered by L1 Sepedi and L2 English speaker.The assessment took a maximum of 45 minutes to complete: Anaphoric resolution test took 30 minutes while the inference test took 15 minutes.At each school, the selected Grade 7 learners were assessed as a group of 50 in their classroom.Each learners was given a question paper and answer sheet combined in which to write responses.In order to ensure that that the learners understood the test well, and to clarify any words that may have been incorrectly read, the tester went through the question paper with the learners.Section one of the test was the inference test and section two was the anaphoric resolution test.Learners were given enough time to check their papers if they had left blank spaces before they submit.At all the school, learners wrote the L1 first and then L2, with an interval of an hour.

Results and Discussion
To determine if there was any observed relationship between L1 and L2 for each school with regard to anaphoric resolution test scores, the raw data is presented in order to check the cause and effect of individual results.To further examine if there was any balanced reading by looking at the scores.The raw data represent 15 learners from each school.Raw data shows that majority of the learners performed better in Sepedi than English.As evident from the above spread sheet of raw data on SPSS, the scores on both English and Sepedi tests reflect a wide range of scores.Learners in all three schools therefore differed greatly in the scores they obtained, with some scoring 2/20 whilst others obtained 17/20 on the same test.The only small exceptions to this were relatively small ranges seen in School B where one learner obtained 0 score for Sepedi and 5 scores for English, while another learner scored 2 scores for Sepedi and 0 score for English.All schools performed very low, because none of the schools performed above 50%.These results suggest that all schools may therefore be at a disadvantage academically because of their low scores.Although, the Sepedi scores on anaphoric resolution are higher than English scores, there is still a need for the mother tongue instruction to be developed ISSN 2039-9340 (print) Mediterranean Journal of Social Sciences MCSER Publishing, Rome-Italy Vol 5 No 23 November 2014 956 to a level that is sufficient for schooling, considering the above scores.This was further investigated by using a correlation test to examine the relationship and to further check if different marks presented on the table are statistically significant.

Relationship between L1 and L2 for the three schools
(1) is the observed relationship between L1 and L2 for each school with regard to anaphoric resolution test scores significant?, Alternative Hypothesis-The observed relationship between L1 and L2 for each school with regard to anaphoric resolution test scores will be statistically significant.The above table represent School A with both English and Sepedi results during the anaphoric testing in order to analyse if there is a statistically significant correlations.The hypothesis that stated that the observed relationship between L1 and L2 for each school with regard to anaphoric resolution test scores will be statistically significant is supported by the above table .The results show that all schools reveal statistical significant at 0.01**, with correlation at 1.The results show a relationship between L1 and L2 with a significant value of 0.001.

Comparison between Sepedi and English results in each group
Is the difference reading scores observed in each school significant for both languages in anaphoric resolution test?Alternative Hypothesis-The different reading scores observed in each school will be significant for both languages in anaphoric resolution test.
On the Sepedi test, School B learners obtained a high mean score.Their overall mean score on the Sepedi test was M=8.5.School A followed with a mean score of M=7.9 for Sepedi test.Lastly, School C achieved a mean of M=7.9 for their Sepedi scores.The English reading scores showed low means in all the school with School A obtaining a mean of M=5.7, School B also obtaining a mean of 6.4 and lastly School C with M=5.3 which is the lowest performing school on the Anaphoric resolution test in English.The overall results shows learners performed better in Sepedi than English.
There is correlation with different values presented above the table.They further show statistical significant figures.All schools have significant results when paired-sample t test was employed.There is a relationship between L1 and L2 with a significant value of 0.001.Paired-sample t test was utilized in order to analyse if the relationship between L1 and L2 in each school is statistically significant.All schools showed that the relationship between the languages is statistically significant P<0.05.For the paired sample test, the mean and standard deviation yields different results from the paired sample statistics.This analysis was intended to establish the relative reading proficiency in English and Sepedi in each of the schools.The hypothesis that stated that the different reading scores observed in each school will be significant for both languages in anaphoric resolution test is revealed to be true and therefore the results are in support.

Figure 2 on Inferential Proficiencies of School B and C for both Sepedi and English
The mean scores are about half of the required score (maximum=3).Learners performed well in Sepedi than in English.Sepedi reading and teaching instruction needs to be developed since there is still a number of learners who are unable to write in their mother tongue.

Dialect influence on learners' performance in School B for Sepedi scores
(2) Is the written language form influenced and affected by spoken language in School B "Selobedu" dialect?, Alternative Hypothesis-The written language form will be influenced and affected by spoken language in School B "Selobedu" dialect.
In Sepedi, the School B learners obtained high mean scores on the inference test.Their mean score on the Sepedi test was 1.58.The School C learners' mean score was 1.42, with a difference of 0.16.The standard deviation for both the schools is similar with the value of SD=1.26 and an overall percentage of 51 for School B and 47% for School C.
Learners at School A and C were fortunate enough to learn to read in their own dialect that contributes 30% to the standard variety of Sepedi.From the above results there is no evidence for dialectical influence that delays learners' progress.Learners seem to be able to juggle between their home dialect and the instructional language learning at school.The hypothesis that stated that the written language form will be influenced and affected by spoken language in School B "Selobedu" dialect is therefore rejected since the results of the inference test yielded different results.Table 2 shows correlation results for both School B and C on their differential dialect to examine if their differential performance is statistically significant.The table show a correlation at 1 and the relationship is not statistically significant with the value of 0.79.

Discussion
Bilingual reading comprehension is promoted in all South African schools.This article posed several questions; (1) Is the observed relationship between L1 and L2 for each school with regard to anaphoric resolution test significantly?(2) Is the ISSN 2039-9340 (print) Mediterranean Journal of Social Sciences MCSER Publishing, Rome-Italy Vol 5 No 23 November 2014 958 standard variety form influenced and affected by spoken language in School B "Selobedu" dialect?(3) Are there correlations between variance of inference skills in L1 and L2, and are they statistically significant?The results of this study suggest that learners are struggling to read in both their home language and in English.The reading scores are much worse in the language of teaching and learning.If these Grade 7 enter a high school without the proper literacy skills that will enable them to use reading as a spear for learning and success then it is a much bigger challenge for reading comprehension.
There is a need for a balanced reading pedagogy and more research to argument these findings since there is very little research done in reading development especially in African languages.Their low reading comprehension scores may suggest that the learners do not have a culture of reading in both their home language and English.Reading resources seems to be one of the identified factors that hinder the reading materials.However, this factor cannot be identified as a reading barrier to learners in this study since these schools did not experience lack of reading materials delivered at the schools.The only barrier that could contribute to the low reading levels is access of reading materials in these schools, reading instruction and material distribution to the learners.It is recommended that the second language readers should at least read at about 70% of the rate of first language users.Makalela (2010;2012) reported that primary school learners are reading very slowly and they are 4 years below their expected proficiency level.Matjila and Pretorius (2004) also reported similar case with Grade 8 learners performing below their maturation level.
The findings of this study indicate that Grade 7 learners are not much exposed to the reading instructions which deny them an opportunity and exposure to reading materials.Learners in School B might experience challenges in mastering the standard language variety which is not their mother-tongue.However, the findings of this study suggest that learners have reading problems irrespective of the language in which they read.The reading scores in home language suggests that though learners have poor reading skills but they are performing better in their home language than in English.These Grade 7 learners lack reading instruction skills which can be offered by their teachers, because given these skills these learners can enter high school with reading skills that will enhance them to be successful in the subject content of high school.
Correlation results reveal low reading levels in both the participants' home language and English with score of (-+50%) below 75% and no differential dialect performance between Sepedi inference skills for both School B and C. Mojela (2008:124) wrote in his article that Balobedu people are not happy with the decision taken by both the PanSALB and Department of Education.One of the Grade 12 learners said he could not write or speak Sepedi very well.He continued to say that he could not finish writing his exams because he was translating Selobedu into Sepedi, which took him long to do.He went further to say PanSALB and Department of Education are making things to be difficult for them at school.His views suggest that he feels that "Selobedu" should be recognised as an official language.In addition, Rain Queen Modjadji of Balobedu told the former president Nelson Mandela that her people are forced to learn Sepedi, which according to her is not their language.Mojela further stated that "the idea that a language for which there exists no written form, a language which has not yet been alphabetized, is for that reason that intrinsically inferior, not a real language but a mere dialect".
Learners at School A and C were fortunate enough to learn to read in their own dialect that contributes 30% to the standard variety of Sepedi (Mojela, 2008).From the above results there is no evidence for dialectical influence that delays learners' progress.Learners seem to be able to juggle between their home dialect and the instructional language learning at school.The hypothesis that states that the written standard language variety from will be influenced and affected by spoken language in School B "Selobedu" dialect is therefore rejected since the results of the inference test yielded opposing results.The study does not offer scientific evidence regarding the influence of dialect on the acquisition of initial reading ability.Learners at School B outperformed School A and C in inference scores for English test (A=0.66,B=1.02, and C=0.32).
One of the factors is that learners do not understand their reading materials because they do not understand the language of teaching and learning.If this is the case for learners learning English as a second language therefore learners who do not understand Sepedi reading materials might be experiencing the same problems when learning English.Anaphoric resolution is a reliable predictor for successful reading comprehension and if learners fail to resolve anaphors it leads results learners not comprehending the reading materials.Learners should be trained with the language that they understand in order to develop the CALP skills.

Conclusion and Recommendations
Higher reading scores determine the success of learners' reading skills.Reading comprehension levels showed a strong relationship during inference skill testing between School B and C. The study was a descriptive and sought to examine ISSN 2039-9340 (print) Mediterranean Journal of Social Sciences  (2014:27) the challenges experienced in reading comprehension will not be overcome if the L1 reading trajectory is not understood, followed and treated.The study further suggests a call for action on the part of educators to teach literacy pedagogy to primary school learners in their early schooling.Another study carried out by Makalela and Fakude (2014:76) elutes that that Oral Reading Fluency (ORF) showed a slow, inaccurate reading process that compares to a 'barking' phenomenon which they defined in the study as making incomprehensible noises towards a text without adequate intonation, rate and accuracy that match natural speech.The two above studies give further recommendations to this study.
Even though the home language reading scores were higher in Sepedi than English, the findings in this study indicate that the learners do not read well in either their home language and English, their home language is not developed to a level of CALP skills.Educators need to make sure opportunities are created in the classrooms for extensive reading of different kinds of texts in the learners' home language.The fact that learners leave primary school with poor reading abilities suggests that reading is not a priority in primary schools.Reading instruction should be developed in African languages in order to promote reading development.Different context effect that may have impacted on the results of the study should also be taken into consideration.More experimental studies are highly required and intervention programs for African languages to be implemented.Literacy pedagogy needs to be addressed and remedy the deficiency in primary schools.Continuous support for teacher training on the methodology of teaching, and the provision of learner-materials is needed at the school represented by the schools under investigation.

Acknowledgement
The data reported in this article form part of my MA dissertation (at University of Limpopo) which was funded by VLIR (Vlaamse Interuniversitaire Raad) and supervised by Prof. Leketi Makalela.I want to thank UPSET (Understanding and Processing Languages in Complex Setting) and the Faculty of Humanities (North-West University-VTC) for financial support.A word of thanks further goes to anonymous reviewers and editor of the journal for their insight that made this manuscript to form part of the contributions in this journal.

Figure 1 :
Figure 1: represents raw data for all three schools; School A, B and C

Figure 2 :
Figure 2: Represents reading levels of Grade 7 AR performance for School A, B and C

Figure 3 :
Figure 3: Line graph representing differential performance between School B and School C Line graph shows that there is no differential performance with 47% for School C and 51% for School B. Spoken dialect did not influence learners in School B; they even performed better than School C. Correlation design was used to investigate if similar results are statistically significant.The results show correlation which is not significant with the value of P>0.079.Table 2: Results on correlation between School A and C on Sepedi scores Correlation: School B and C Sepedi scores School: B School: C

Table 1 :
Representing relationship between Sepedi and English for the three schools Sepedi and English reading development of Grade 7. The low reading results indicate that there is a gap between what is happening in schools and what the curriculum seek to achieve.The present results calls for reading literacy research interventions in African languages.According to Fakude