[Spelling Progress Bulletin, Fall 1982 pp5-7]

SSS Conference 3: Spelling in other languages and international aspects of English Spelling continued.

"Teaching English in Francophone Africa,"

by Henry Niedzielski, Ph.D.*

*Univ, of Hawaii at Manoa, Honolulu, Hawaii.

Abstract.

As many people know, English has replaced French as the main language used for international communication. One example of this gradual loss of French to English can be witnessed in Burundi. The ambition of the Burundi government is to make the country trilingual in Kurundi, French and English.

Presently, the country is officially bilingual. The academic, social, and administrative elites communicate in French; most other Burundi speak almost exclusively in Kirundi. However, much of the country's international trade must be conducted in English because Burundi is a landlocked country and its nearest ports of access are in Tanzania or Kenya, both English speaking countries.

Consequently, more and more people are studying English, officially or privately.

Obviously, not all Francophone African countries feel the same necessity to learn English. Many, however, find it useful in order to read scientific, commercial or technical publications. Most also find it nearly indispensable for traveling and for attending international conferences. These latter purposes require the teaching of spoken English. In this respect, a question comes to mind. Is it preferable to teach English directly in the learner's native language or through French, the general and official language of instruction? I have set up an experiment to discover the answer(s) to this question. Five areas of language learning and proficiency are being analyzed. A preliminary report on this experimentation is given in this paper. The findings could be generalized to most Francophone (or Lusophone) African countries.

(Comment by VY: This paper, on learning English in a polyglot African country, illustrates forcefully that English is an international language, spelling and all, and reminds us that spelling must be considered in the context of other aspects of the English language too.)

Corpus.

1. Introduction.


As many people know, English has replaced French as the main language used for international communication. This process of substitution, noted already in the first half of this century, has recently gained momentum in some "French speaking" African countries. One example of this gradual loss of French to English can be witnessed in Burundi. It was made clear on Apr. 22, 1981 when the Minister of National Education proclaimed, at the seminary for secondary school teachers of English in Burundi, that the ambition of the Burundi government is to make the country trilingual in Kirundi, French, and English.

However, the teaching of English in Burundi suffers from many problems, some of them inherited from the colonial days. Introduced in the late fifties by the Belgians, it was taught more or less like Greek or Latin. Slowly, after obtaining their independence, Barundi have come to realize its' importance as the language of communication both regionally and internationally. Presently, in the secondary schools it is given the same number of instruction hours as French, and at the post-secondary level, it is a compulsory subject in all institutions and in all departments. It is also taught on the radio, where it was first introduced with French explanations and now presented in Kiswahili. We thus observe the first problem. At least 95% of the Burundi population speaks Kirundi as the native language, yet English is taught through another language, mostly French.

This observation led me to conduct an experiment to check the question whether Barundi would learn English more efficiently if it were presented with occasional explanations in Kirundi rather than in French.

2. The Problem.


Interferences caused by the linguistic background of the Foreign Language learner may vary with the skill being acquired. Therefore, five components were determined for this experiment: lexicon, structure, conversation, pronunciation, and spelling. Actually, this division was dictated partly by the fact that investigation was carried out by eleven students in psycholinguistics and that there were ten classes available in Bujumbura for the experiment. It was also influenced by the opinions collected in short essays written by 23 advanced students in the department of English on the topic of "Present linguistic and other interferences in the learning of English in Burundi." The following statements summarize these students' thinking.

2.1 Lexicon.


Since some words are better known in French than in Kirundi and many are graphically and/or semantically similar, the class felt that vocabulary might be taught better through French rather than through Kirundi.

However, in some situations and because of cultural contexts, some lexical items are easier to understand through Kirundi than through French (e.g. small pox, petite vérole, and akarandi). Another minor danger in teaching vocabulary through French is caused by the abundance of deceptive cognates (e.g. parents, parents', relatives).

2.2 Structure.


English grammar might be better approached from French because both languages share many structures and classify their parts of speech in a similar manner. Kirundi, for instance, has no articles and no real relative pronouns.

On the other hand, the English tense system could possibly be taught more efficiently through a direct contrast with the two main groups of Kirundi auxiliaries of aspect (Mategeko 1971: 123).

2.3 Conversation.


Here again opinions were divided. Because culture shapes the way of thinking (Whorf's hypothesis), and since there are many more similar cultural contexts in French than in English, English conversational skills should be built on situations with which the students are already familiar in French.

Others felt that materials should be created and adapted from Kirundi by building on speech acts and speech modes equivalent to those familiar to Barundi in Burundi speech situations.

2.4 Pronunciation.


Presently, the teaching of pronunciation is introduced through the IPA. Students suggested that the English phonological system should be introduced gradually through contrasts with the Kirundi phonological system which is much less complicated than the French. Phonetic notation could be introduced much later for reinforcement.

2.5 Composition.


Most Barundi students perform on a much lower level in writing than in speaking. It might be due to the facts that eloquence is highly valued in Burundi culture and that, until recently, written literature did not exist. Therefore, developing and organizing techniques already acquired in French could be utilized.

2.6. General problems.


A certain percentage of students speak Kiswahili instead of Kirundi. There is also a shortage of local teachers, and European teachers - mostly Belgians - do not know Kirundi nor Kiswahili. Ideally, the English teacher would need a working knowledge of English, French, Kirundi, and Kiswahili. It will still take a few years before a sufficient number of so qualified Barundi is ready to teach.

Finally, there are no texts in Kirundi because the language of instruction has been French, even to the extent that university courses in Kirundi literature or culture are taught in French.

3. Previous Research.


A thorough analysis of various "memoires" (equivalent to U.S. M.A. theses) did not reveal any formal interest in the teaching of English directly from Kirundi. The general assumption has been that English could, should, and would be taught through French.

Interviews with various methodologists and program writers at the BEPES (Bureau d'Etudes Pedagogiques de 1'Enseigment Secondaire=Planning and Programming Office for Secondary Schools) have confirmed this lack of interest in the teaching of English directly from Kirundi. Nobody had even thought of doing it. This may be explained by the fact that until recently both English and Kirundi were studied as an additional language by French majors. It seems therefore traditionally logical that the teaching of English should be based on a good knowledge of French.

The only notable exception to this state of affairs has been a M.A. thesis written in 1971 by a Burundi on a Fulbright grant at the UCLA. Zacharie Mategeko, the present chairman of the English department at the Univ. of Burundi, wrote in his introduction to A Contrastive Analysis of Parts of the English and Kirundi Tense Systems that Kirundi can be efficiently used in the teaching of the English tense system and of the English language in general.

Ten years later this statement had not yet been tested. This experiment is the first attempt at doing it.

4. The Method.
4.1 The Subjects.


Altogether 375 students distributed in ten different groups were taught in May 1981 some rudiments of English sounds, vocabulary, spelling, structure, or conversational skills. They were all first year students in Bujumbura secondary schools, who had studied French for about four years and would start studying English the next year. Table 1 shows the number of people in each group. There were only 43 females, 42 in the two conversational groups and one in the French pronunciation group.

The age of the students ranged from 12 to 21 years, the average being 16 for the boys and 15 for the girls. Between 20 and 40% of them came from Bujumbura and some of them may know Kiswahili better than Kirundi. The remainder come from up country and are native speakers of Kirundi. Parallel Kirundi and French groups were chosen in the same schools in order to ensure that the socio-cultural background would be identical and that the only decisive variable would be the use of French or Kirundi in the experiment.

4.2 The Instrument.


Ten short teaching programs were written, discussed and dry run in my psycho linguistics class. They lasted 45 minutes except for pronunciation and conversation classes which took 55 minutes.

All writing groups were tested within the same class period. The members of the conversation and pronunciation groups, which required oral testing were examined individually during the next class period under such conditions that no communication took place among the testees. Each one was questioned by the same experimenting university student with another one evaluating the answers.

5.0 The Results.


A synopsis of results is found in Table 1. Higher performances were achieved in all groups taught directly from Kirundi, except in pronunciation where those who had been taught through French obtained better results.

The total number of errors and the average number of errors per student in each group are shown to give a quantitative indication of overall performance of the various groups and to provide a comparison between the Kirundi and the French groups for each skill under study. The average score in percentile is not computed on the basis of the average number of errors subtracted from the perfect score. Rather it is derived from a formula multiplying a perfect score by the number of students in a group, then subtracting from this total perfect score the number of errors in that group, and finally dividing this remainder by the number of individuals in the group. It seems that this approach helps to spread the numerical results.

Conscious that for each group of individuals, statistics may be misleading because of a few unusually high or unusually low performers, all computations were redone after eliminating about 10% of the top students and 10% of the bottom students in each group. The results appear in the lower half of Table 1.

6. Conclusions and Implications.


First of all, it is remarkable that all Kirundi groups performed better than the French groups except the pronunciation group. The results obtained can be construed as an indication that Barundi teenagers can learn English more efficiently directly from their native, home language.

In fact, we have seen that a serious attempt was made to reduce all socio-cultural and physiological variables by choosing each set of parallel groups in the same school. In addition, psychological factors influencing the teaching and/or learning processes were also given full consideration. The student teacher in each group was convinced of the superiority of his/her approach and tried to share his/her enthusiasm with the subjects. The latter reacted to the experiment in an overwhelmingly positive manner.

In the pronunciation groups, the results obtained were somewhat puzzling. A more thorough analysis has revealed several reasons which lead us to conclude that this section of the experiment lacks validity. We have seen that only consonants had been taught and tested. Out of 7 these consonants, the most frequently reported as mispronounced were [r] 89 times in Kirundi, 30 in French; [θ] 130 times in Kirundi, 52 times in French; and [ð] 130 times in Kirundi, 64 times in French. These three sounds do not exist in Kirundi nor in French, and they account for over half of the errors in both Kirundi and French groups. On the other hand, those sounds which exist in Kirundi were reported as mispronounced more frequently in the French groups. This suggests that greater differences in phonetic features cause greater difficulties. But why should the Kirundi group have scored lower than the French group on difficult sounds? At least two possibilities may have existed. The subjects may have been more attentive in French because they were working in a foreign language, and the student teacher may have been stricter. To control the latter factor, the experiment could be repeated with a jury of at least three better trained examiners, the same for both groups. It is, therefore, recommended that a new experiment testing phonetic acquisition be conducted. It should also contain vowel sounds because the tables show much greater differences between the French and English vowel systems than between the Kirundi and the English vowel systems. The hypothesis should be that the Kirundi group would perform better, based on the observation, reported above, that the greater the difference between the source and the target phonology, the greater the difficulty.

In the spelling section, the number of words which could have been influenced by French was reduced to the minimum in order not to favor either the Kirundi or French group. In addition, the word school was chosen because its cognate in Kirundi is ishule. Professor was mispelled more frequently in Kirundi, often with only one s; exercise was once misspelled in French as exercice; coffee was spelled with one f three times in the Kirundi group. There are no double consonants in Kirundi. A new experiment could use French cognates (of which there are over eleven thousand). Preferably, they would be introduced without contrasting English with French in the Kirundi group. In the present experiment, no similarities or differences were pointed out.

It is probably too premature to draw any definite conclusion or recommendation for a preferable teaching approach. However, the results obtained encourage us to suggest that more thought should be given to the teaching of English directly from Kirundi. Since the subjects in this experiment are going to begin learning English next year, a longitudinal study could be conducted over the year with parallel Kirundi-French sections in the schools where this experiment took place.

Table 1.

Category LexiconStructure Conversation

Total number of students
Total number of errors
Errors per student
Total points possible
Average score in %
No. of students minus top 10%
and bottom 10%
Total errors of 80% of students
Errors per student
Average score in %
French
38.00
135.00
3.55
15.00
76.31
30.00

110.00
3.66
75.60
Kirundi
38.00
82.00
2.15
15.00
85.66
30.00

53.00
1.76
88.26
French
40.00
37.00
0.92
10.00
90.75
32.00

19.00
0.59
94.10
Kirundi
43.00
36.00
0.83
10.00
91.62
33.00

15.00
0.45
95.50
French
20.00
41.00
2.05
8.00
74.37
16.00

28.00
1.75
78.12
Kirundi
22.00
42.00
1.90
8.00
76.23
16.00

20.00
1.25
84.37
 
Category Pronunciation Spelling 

Total number of students
Total number of errors
Errors per student
Total points possible
Average score in %
No. of students minus top 10%
and bottom 10%
Total errors of 80% of students
Errors per student
Average score in %
French
44.00
546.00
12.40
49.00
74.67
34.00

331.00
9.73
80.14
Kirundi
41.00
514.00
12.53
49.00
74.40
33.00

417.00
12.63
74.22
French
44.00
89.00
2.02
15.00
86.51
34.00

60.00
1.76
88.26
Kirundi
43.00
77.00
1.79
15.00
88.06
33.00

51.00
1.54
89.43
 

Back to the top.