[The following is an excerpt from Helen Abadzi's book, Efficient Learning for the Poor: Insights from the Frontiers of Neuroscience, which I recommend for those involved in MLE.]
For many children, education in another language is more difficult than expected. The deficits in native language development common among the poor may inhibit the rapid acquisition of a second language. Mother tongue instruction is a prerequisite if Education for All is to be achieved, particularly when the official language has complex spelling rules. The official language should be taught to children as early as possible. However, it should become the platform for learning new information only after children know it sufficiently well to process complex sentences and vocabulary. A gradually decreasing percentage of mother-tongue instruction seems to be an effective way to introduce an official language.
Visitors to French-medium primary schools in sub-Saharan Africa are surprised to find out that children may understand little of what they are told and merely repeat verbatim what they hear. Sixth graders in rural areas may read haltingly or in monotone and be unable to answer comprehension questions on simple passages. Why are schooling outcomes so poor?
Many countries have multiple languages and a need to teach in a common language. In countries like Romania or Indonesia, children speaking minority languages must learn the official language of instruction. In many others–including most countries in Africa and the South Pacific–the lingua franca is foreign to everyone (for example, English, French, or Portuguese). The countries with multiple languages have various language-instruction policies. In some countries, students may study in their mother tongues in lower primary grades and then switch to the lingua franca. In others, logistical and political complexities result in the use of the lingua franca for all grades. The latter approach is preferred in much of Africa and impacts some of the world’s poorest countries.
There are definite advantages to learning a second language at an early age. The earlier grammar is learned, the easier and faster it is mastered. The time to learn a language most efficiently is from age 3 to about 8; then the ability falls off dramatically and steadily until adolescence and then again until adulthood. After age 8, children are no longer in their prime language-learning years, but as their working memory and reasoning ability increases, they can put explicit grammar rules to use. Non-native students in primary and secondary years become fluent in about a year and eventually competently master a language. With help, they may be able to catch up with students studying in their native language, although the latter double their vocabulary every two years between grades 1 and 5, and by grade 5 know 40,000 words. However, the literacy rate among speakers of minority languages worldwide is low, reported at 20-30 percent. If children are poor and also study in an official language they do not know well, how serious are the problems that arise?
Proficiency in a first language predicts success in studying a second. Better-off children can benefit from bilingualism because a second language added to a rich knowledge of a first language results in complex knowledge networks (additive bilingualism). The reason is not well understood; the result seems to be a limited knowledge of both languages and a vocabulary too limited to make sense of classroom material. It then becomes difficult to build knowledge networks on various topics and attach new information. For example, a child who knows the meaning of “justice” or “honesty” in one language can acquire the terms in another but faces a harder task if she has to acquire both the label and the concept in her second language. In particular, poorly fed students who often have a more limited working memory may need more time to acquire vocabulary.
Beginning learners and readers must quickly reach a threshold of language knowledge where they start to learn subject matter from context. Research on students with English as a second language shows that in English, with just 1,000 words, one covers 72 percent of the text. But to successfully guess the meaning of unknown words, at least 95 percent coverage is needed. If books are written relatively simply (such as novels for teenagers), 3,000 words provide a 95 percent coverage in English. Thus a benchmark of roughly 2,000-3,000 words is needed for children to understand the content of lessons sufficiently. To get to this level and keep increasing vocabulary as native speakers do, children should learn 1,000-2,500 words a year. But hearing a language is not sufficient to learn it. The speech must be directed at students, and in large classes the opportunity may be minimal.
When students know only limited grammar and vocabulary, it is not even possible to create a basis to which to peg complex concepts, and the classes may get stuck in a simple level of discourse. For example, grade 4 students in Guinea still learn the names of clothes in French, while grade 5 students are asked to name what is in a picture (water, birds, pebbles). Their vocabulary does not approach the 40,000 words that the typical native speaker of French knows by grade 5. Unfortunately, such problems may not be obvious to a casual observer. Children’s ability to repeat sequences verbatim can fool adults, making them believe that children actually know what they say.
Instructional languages with orthographic complexities spell trouble. Even if children do not know a language well, they could learn fluent reading when the spelling is simple (as with Spanish, Romanian, Bahasa Malaysia, Armenian, or even Hindi). But official languages have often been written for centuries, resulting in irregular spelling. Children learning to read through English must memorize words spelled unpredictably, but they may not yet know their pronunciation, meaning, or related context. In French they must learn how to deal with dashes, accents, and endings as they were pronounced about 600 years ago. As a result of these issues, most of the school time in sub-Saharan Africa is taken up with language and spelling instruction, such as putting the right accents on French. Ironically, the neglected African languages tend to be phonetically written and very suitable for reading instruction.
Phonological awareness and decoding skills transfer from one language and script to another. However, visual features such as letter shapes do not transfer. For example, students who learn to read fluent Arabic in Koranic schools may be able to link sounds with letters but still need a commensurate amount of reading practice in Latin script. However, if they read an African language in the Latin script, transfer to English or French will be relatively simple, despite spelling complexities and differences in some letter values (for example, from Malagasy to French). Countries that have a choice (such as writing a language using Arabic or Latin characters) might choose the script of widest use and apply it to multiple languages.
Children in foreign-language instruction need more time but get much less. In principle, language problems can be overcome through increased hours of effective instruction, private tuition, and a wealth of audiovisual materials. Instead, textbooks are scarce, teachers are poorly trained (and may not know the lingua franca well), absenteeism is high, and class time is used poorly. The poor cannot afford private tuition and the national curricula may have insufficient space for both a new language and basic skills. To make things worse, hours in the crowded schools of some countries (like Niger, Mali, and Burkina Faso) have been reduced by about 40 percent to enable more students to attend (a policy known in French as “à double flux” or “double vacation”). Ironically this policy serves to screen students and identify the 2-5 percent who are gifted enough to learn with limited exposure.
All these issues create a bar that is too high for all but the most privileged to overcome. With school days lasting at best 2.5-4 hours a day, sufficient knowledge takes years to acquire. Unless students are very bright or well to do, their test scores are very low and many simply drop out or graduate from grade 6 functionally illiterate (as is the case in Niger and Guinea).
The Benefits of Bilingual Education
Teaching basic skills to poor children through language immersion may be detrimental, but bilingual education is a much more effective option. U.S. students receiving instruction in a native language and English at different times of the day were found to make the most dramatic gains in reading performance compared to their English-only peers. This research is pertinent to multilingual low-income countries.
Students require at least 5-7 years to approach grade-level norms on school tests that measure cognitive-academic language development in English (Figure 3). Students who arrived in the United States between ages 8 and 11, and who had received at least 2-5 years of schooling in their native language and home country, were the best achievers and took only 5-7 years to catch up in English. Those who arrived before age 8 required 7-10 years or more to catch up. The children arriving during early childhood (before age 8) had the same background characteristics as the 8- to 11-year-old arrivals. The only difference between the two groups was that the younger children had received little or no formal schooling in their primary language; this factor appeared to be a significant predictor in these first studies. (In countries where there are almost no native speakers in the schools, these reference points may be lost.)
The number of years of instruction in the first language is the most important predictor of reading performance in a second language. It is not important what the first language is, but rather how much cognitive and academic development the student has experienced in it. The higher the students’ achievement in the primary language, the faster they will progress in the second language.
Deficiencies in the second language may not be apparent in lower grades, but they increase after grade 4-5, when the concepts become more challenging. For this reason, students in the United States who are taken out of bilingual education no longer catch up with English-speaking students. Merely immersing students in a second language or giving them separate intensive language instruction results in low reading scores many years later. Poor students who receive only immersion instruction are also more likely to drop out. Socioeconomic status becomes more important when the program is of low quality, and the poorer students do less well. (Relate to CLE program in the Philippines.)
To catch up with native English speakers, minority language students must receive effective bilingual instruction (that is, studying parts of subjects in both languages). Otherwise they continue to score lower on basic skills tests throughout their 12 years of school. The performance difference between students receiving bilingual education and those receiving only English or English-as-a-second language instruction may be dramatic.
Research suggests that a very effective model of language introduction is 10 percent of foreign language in grade 1, gradually increasing to 20 percent, 30 percent, 40 percent, and 50 percent by grade 5. Students who start school performing at grade level and receive such gradual bilingual education will catch up with English-language students by grade 5.
Accordingly, the pedagogie convergente of Mali gradually introduces French and reduces the local languages, but also introduces more “student-centered’ strategies. Despite a lack of political support, textbooks, and training, students have a much lower repetition rate (3.7 percent versus 18.1 percent of French-speaking classes). At the end of the first year, about 69 percent were promoted compared to about 52 percent of French-language schools. Classroom observations showed high rates of student engagement and writing skills in French, in contrast to French-only schools. Although the pedagogie convergente students are mainly rural and peri-urban, they scored slightly above French-only students in grade 6 standardized achievement tests conducted in French. Pedagogie convergente nominally costs 80 percent more to teach because of materials and teacher training, but it costs 27 percent less when repetition rates are considered. However, political support for this program has been limited.
How Many Years Are Needed to Teach Sufficient
Language in the Schools of the Poor?
Children must learn 2,000-3,000 words through interactive speech before they can understand the content of the lessons sufficiently. To keep up with the increased complexity of school, they learn about 1,000-2,500 words per year.
In second-language acquisition, a college student taking immersion class in a foreign language for three hours a week would require about four years or about 600 hours of interactive instruction and reading to attain relative mastery in one of the common languages. Given the limited instructional time and competence of many teachers, it is unknown how long it takes to provide sufficient levels of instruction.
Sources: Nation and Waring 1997, Putnam 1975.
The positive results of literacy in mother tongues are evident in a program piloted in 800 schools in Zambia. The children learned to read just in local languages in grade 1 while learning English orally at the same time. Efforts were made to use instructional time well and employ interactive, “student-centered” methods. English writing was introduced in grade 2. The results were astounding. In 1999, students read on average two grade levels below their grade level in English and three grades below their grade level in Zambian languages; but in 2002, students reading and writing scores were 575 percent above baseline for grade 2, 417 percent for grade 3, 300 percent for grade 4, and 165 percent for grade 5. Scores in Zambian languages ranged from 780 percent above baseline from grade 1 to 218 percent in grade 5. Subsequently, the program was introduced to all schools of the country.
Political ambivalence toward native languages. Governments seem unable to reconcile political issues with instructional outcomes, particularly since the children of the officials themselves usually have little difficulty with the foreign language in school. Parental preferences pose an unexpected obstacle in bilingual education. In the words of one Senegalese father during an IEG evaluation mission, “The child already knows Wolof; what is the use of studying in Wolof?” Parents with limited or no education may expect a higher status or better family income if their children are educated in the official language. Frequently, the strong objections to local-language instruction are due to concerns that the language may acquire political importance in areas which try to set tribal differences aside. Reactions against national languages in Africa have been widespread and have resulted in reversal of programs in Guinea and Madagascar. However, the success of programs in Zambia, Mali, and Burkina Faso are encouraging.
Immersion languages with spelling complexities such as English and French should be avoided for the poor at all costs. Children should first learn to read fluently in their mother tongue (assuming its script is used extensively in the country, or also used for the official language). Logistics to teach multiple languages may be indeed complex, but ultimately mother-tongue instruction in the early grades may prove cheaper. The cost-effectiveness of bilingual education needs to be calculated, taking into account the costs of grade repetition and social costs of illiteracy.
Teaching children the basics in a mother tongue includes the use of dialects, such as Maghreb Arabic and various Creole dialects. Phonetically based writing of dialects may greatly facilitate the acquisition of the mainstream language, particularly when its spelling is complex. Children also need to be taught the change patterns between the dialect and mainstream language and learn to use them consciously.
A communication strategy involving mass media is needed to explain to parents the rationale for mother-tongue instruction and underline its utility if children are to learn the official language sufficiently well to progress in their studies. To deal with the political concerns involved in mother-tong instruction, governments may explain that local languages have a preparatory role and may be phased out of curricula after grades 3 or 4.
In the most remote schools it may be ultimately impossible to teach sufficient language and make a transition to higher grades possible to many students. If that is the case, then fluent reading and math skills in a native language as well as basic concepts may be all that most may acquire. However, that may be sufficient for correspondence and personal needs.
A better understanding is needed about what is gained and what is lost when poorer children acquire a language for classroom use, given the instructional limitations of low-income countries. At what points and under what conditions does subtractive bilingualism occur? What ages and how much mother language knowledge create the most susceptibility? What is the number of hours poor children must be exposed interactively to English and French before they acquire the essential patterns of grammar and vocabulary? Is that number of hours realistic given the usual amounts of instructional time? Answers could result in improved sectoral strategies.