A.K. SRIVASTAVA AND RASHIDA KHATOON
INTRODUCTIONThere is a very strong section of opinion favouring mother tongue (MT) as the medium of instruction (MI) for every child. Gandhiji strongly supporting this view, thus wrote "... mothertongue is as natural for the development of man's mind as mother's milk is for the development of the infant's body" (Sakrikar, 1979; p.11). This view is based on sound pedagogic principles. At the time of entering school, a child's competence to understand and manipulate his MT is well developed. The cognition of a child bears the stamp of the particular linguistic environment and culture to which he has been exposed. One of the most important principles of educational psychology is first to ascertain what the child already knows, and then to start accordingly (Ausubel, 1968). The child knows well his MT. Therefore, if the MT becomes the MI also, education becomes more meaningful where the cognitive faculties are ready to receive and act, transfer and integrate.
Supporting the plea for making MT of a child his MI also, the UNESCO report (1953) writes, "Psychologically, it is the system of meaningful signs that in his mind works automatically for expression and understanding. Sociologically, it is the meaning of identification among the members of the community to which he belongs. Educationally he learns more quickly through it than through an unfamiliar linguistic medium". (p.11). The principle of teaching through MT was strongly recommended by the Education Commission (1964-66). The three language formula recommended by the Commission, emphasizes this principle and it also forms part of our National Policy on education (Srivastava et al.,1978). However, for inclusion in the school curriculum only a broader definition of MT, which includes only the regional languages recognized in the VIII Schedule of constitution, has been taken. Therefore, if MT is defined as the language spoken from the cradle, there is bound to be difference in the MT and the regional language which is the MI.
Several studies have empirically supported the beneficial effect of teaching through MT on child's performance and mental ability (Anand, 1971; Kamakshi, 1966, Prator, 1950; Skutnabb, 1976; Yela 1975). Anand (1971) in his study of Kannada and English medium students found the former significantly superior in achievement and verbal intelligence and the latter in non-verbal intelligence. Skutnabb-Kangas (1976) on the basis of her study of 700 Finnish immigrant children claims that unless a child has him MT as the instructional medium he cannot learn a second language properly. In fact forcing second language in places where the MT has low status will result in double semilingualism in which neither of the two languages will be proficiently acquired. Yela (1975) found support for the hypothesis that since children taught through their MT used simple and direct strategies, as opposed to the complex and indirect ones of the bilingual children, the former would performs significantly better than the latter on verbal problems involving reflective thoughts, grammar and syntax. All these studies demonstrate the beneficial role of teaching through MT one a child's all-round development.
Even though MT is selected as the instructional medium, the child of today cannot afford to remain unilingual and he has to learn more than one language. There is an emphasis on bilingual education and several studies have been conducted to assess the impact of bilingualism on child's cognition, achievement and personality (Cummins, 1976; Srivastava, 1976). The programme of bilingual education has generally followed two approaches. The first one has taken the form of total and partial immersion programmes. In total immersion, from the very beginning the MI for teaching of all the subjects is the second language. In partial immersion, the MI is equally divided between the first and the second languages. In the second approach the MI is the child's first language which is followed later by introducing the teaching of second and third languages. In India while the English medium school follows the first approach the first language medium schools follow the second one.
Supporting the necessity for bilingual education, Ben-Zeev (1976) suggests that bilingualism involves an intense degree of cognitive conflict, which facilitates language learning. In his study he found the bilinguals significantly superior in performance, analytical grasp and system understanding. Quite a number of recent studies have empirically supported the view that bilingualism accelerates cognitive growth, including both convergent and divergent thinking and achievement (See Srivastava 1976, review). These studies however, compared the groups of bilinguals and monolinguals. They have not compared the two situations, where both groups are bilinguals but for one of them the MT has been the MI also and for the other, the two have been different. The total and partial immersion programmes (Barik and Swain, 1975, 1976, a, b and Lambert and Tucker, 1972) and several other studies (Dave and Dave, 1971; Macnamara, 1966; and Reddy, 1977) have studied this aspect of the problem. They have reported either beneficial effect or no detrimental effect on the cognition and achievement of the child if he is educated through a medium other than his own.
In the immersion studies, the English speaking children were taught from the very beginning through the medium of French and their performance in the second language French, native language English, arithmetic and mental ability were compared with those English children who were taught through the medium of their MT besides also learning French as a second language as part of the curriculum. No detrimental effect was reported on the immersion children who generally did not differ significantly from the comparison group on measures like arithmetic and mental ability. In their reanalysis of data, Barik and Swain (1976c) reported significantly higher IQ for the immersion children as against those from the comparison group. The immersion children were better in the second language French and also in some of the skills of first language English. Lambert and Tucker (1972) concluded that in the MT-MI different group there was transfer of skills across the language and the knowledge of contrastive linguistics helped the children in vocabulary building and in comprehension of complex linguistic functions. They also developed linguistic detective capacity like being attentive, patient and developing a concern for words and their meanings. Lambert and Tucker (1972) call this process as 'two-way bilingual relay of concepts and linguistic principles'. It results in reciprocal gain of both the first and the second language. Macnamara (1966) found that English speaking children who had schooling through the medium of Irish did very well in comparison to those taught only through their MT, be it either English or the Irish groups. In fact these children had higher nonverbal IQ than the other comparison groups. Reddy (1977) compared the college students studying through the English and the regional media. His finding was that when the influence of intelligence on achievement was controlled the English medium students achieved bettering several subjects of science and humanities. On the basis of the positive trend in their findings, Lambert and Tucker (1972) laid down the principle that 'in any community where there is a serious widespread desire or need for a bilingual or multilingual citizenry, priority for early schooling should be given to the language or languages least likely to be developed otherwise, that is, the languages most likely to be neglected.' (p.216).
On the other hand quite a numbers of studies have also reported the adverse effect of bilingualism or MT-MI difference on both divergent and convergent mental ability and the achievement of the pupils (Anand 1971; Srivastava 1976). Cowan and Sarmad (1976) in their longitudinal study in Iran, compared the reading performances in Persian and English of several types of schools. These were total immersion English medium with 11/2 hours devoted daily to Persian, partial immersion (MI 50 per cent English and 50 per cent Persian) and two types of monolignual Persian schools, where MI was Persian but English was taught only for 2 or 3 hours per week. In one type of monolinguals school children from upper SES studied and in the other type lower SES children studied. The last two types of schools served as controls. The rationale of this study according to Crown and Sarmad (1976) was that "the child should first learn to make the symbol-sound-meaning associations, which constitute the earliest step in reading, in the language he is most proficient in, his own. Once a firm basis for reading has been established in the first language, transfer of reading skills in the second language may occur' (p. 355). On the basis of this study the authors concluded that, "the overall trend evident groups in their language-as sufficiently distinct from the results of other studies… particularly the St. Lambert experiment…" (p. 372).
Discussions so far will demonstrate the highly controversial nature of findings showing that the issues involved are yet to be settled. McLaughlin (1977) after reviewing the studies in this area arrives at the conclusion that it remains to be proved yet that interference is an inevitable consequence of bilingualism or that its effect on achievement and cognition is either positive or negative. It was decided to conduct a fresh study on the subject in order to find out the effect of different between MT-MI on pupils' intelligence, achievement and creative abilities. Now onwards MT-MI different group will be called as Different, and the MT-MI same group, as same. A comprehensive study on the problem was done by Anand (1971). However, the phenomenon of balanced bilingualism was not taken care of and thus each selected group might have developed different levels of proficiency in the two languages- English and Kannada. Studies have reported significant effect of this phenomenon on a child's allround performance (Cumins, 1977; Srivastava 1976). Moreover, the criterion of achievement in Anand's study was achievement test scores only in three subjects, general science, mathematics and social studies. Achievement tests which measure a pupil's performance at a particular point only do not take into account his performance spread in course of a number of years in his life. For all these reactions it was decided to conduct a fresh study.
Design and procedure - It was decided to compare two groups of students - one whose MT-MI was same and the other with the two different. The former category of students were selected from Kannada medium schools and the latter form English medium, and their achievement, intelligence and creative abilities were compared. The average of the total marks obtained I all the academic subjects of the previous six consecutive examinations served as the criterion for academic achievement. Six examination marks were taken in order to minimize the influence of chance factor on the criterion variable and also to have measure of a student's performance spread in course of several years. In order to select balanced bilinguals the average of the last two previous examinations in first and second languages were found out for every student. The test showed that no significant difference was found for the two groups in their performance in the two languages except for girls who differed significantly (p>.01). The tests were administered at a time on a test was administered in a day to each such group.
Sample: the samples consisted of students from VIII standard only. This was done in order to study those pupils who were for years exposed to either of the two instructional media-Kannada and English, and in whom its effect was expected to have consolidated in either direction. The sample included both boys and girls. Care was taken to select the schools from different localities so that, as far as practicable, boys and girls from both lower and middle classes were represented. Only those schools which maintained their records for the previous examinations properly were taken into consideration. From such schools all the boys and girls were initially selected for study. Some of the schools where facilities for data collection were found to be inadequate were dropped. Those schools which showed reluctance in offering these facilities were also dropped. Finally Kannada medium boys and girls were selected from five schools - Ramakrishna, Bhagini Seva Samaj, Vanita Sadan, Hardwicke and J.S.S. English medium boys and girls were selected from 5 schools - Demonstration, Royal English, Ramakrishna, J.S.S. and Nirmala Convent. In these schools all those students were selected for study who had their previous six consecutive examination marks available. In all, the present study was conducted on 65 girls and 50 boys from English medium schools and 41 girls and 43 boys from Kannada medium schools. The total sample was 84 for the same and 115 for the different group.
Instrument: Raven's (1960) Progressive Matrices Test was administered to measure the non-verbal intelligence. Non-verbal test was selected to rule out the influence of language factors on intelligence. Raven's test as a measure of non-verbal IQ has been used by several studies including those on immersion programmes. Both verbal and non-verbal tests of creativity prepared by Mehdi (1973a, b) were used. The dimension measured by the verbal test these were elaboration and originality. Since the tests were in Hindi, they were translated into English. (Translation given in the test manual) and Kannada for administering them to samples from each medium. The Kannada translation was done with the help of Kannada experts in the Central Institute of Indian Languages. In order to see if the translations were comprehensible to the students from VIII standard, the two tests were initially administered on a group of randomly selected students from each medium. Since they were found to be comprehensible, no modification was made in them.
Analysis of Data: Data were analysed separately for boys and girls and then for the combined sample from each group - Same and Different. Mean and SN for each group were calculated and t-test was used to find out the significance of the difference between the means of the two groups on each variable studied. Homogeneity of variance was tested (Meddis, 1975) for satisfying the assumptions associated in order to control the influence of intelligence on achievement and creativity. Since Ramakrishna and J.S.S. were the only two schools, which were common for both Same and Different groups, their data were separately analysed to check against the school conditions influencing the performance. These schools were for boys only. However, in this paper the analysed data for only the Ramakrishna School has been reported. Sample from J.S.S. School was very small and therefore, they have been reported. Sample from J.S.S. School was very small and therefore, they have not been reported, by the samples from the two schools were not combined for statistical treatment, in order to control the heterogeneity introduced their combination.
RESULTS
Assumptions for homogeneity of variance was satisfied in the case of all the variances except a few. However, for cases where they were not satisfied, data were not transformed. This was done on the strength of the evidence that the parametric tests like t-test remain robust even when the assumptions about homogeneity of variance have not been satisfied (Boneau, 1960; Srivastava, 1967).TABLE 1
Mean, SD, t and Significance levels of Inter-group Difference on Progressive Matrices Test
Groups
Mean (SD)
t
p
girls
Different (N=65)
Same (N=41)34.52 (9.67)
23.44 (9.70)5.57
< .01
boys
Different (N=50)
Same (N=43)39.54 (12.66)
28.44 (13.59)4.02
< .01
combined
Different (N=115)
Same (N=84)36.65 (11.33)
26.00 (12.11)6.26
< .01
As will appear from the above table, the boys, girls and the combined sample of the groups - Same and Different-differ significantly in non-verbal intelligence. In all the analysis the Different group has scored significantly higher than the same group.
TABLE 2
Mean, SD, t and Significance levels of Inter-group Differences on the Achievement measure
Groups
Mean (SD)
t
p
girls
Different (N=65)
Same (N=41)55.92 (14.44)
50.90 (15.041.61
NS
boys
Different (N=50)
Same (N=43)64.36 (14.02)
50.33 (14.65)4.58
< .01
combined
Different (N=115)
Same (N=84)59.33 (15.76)
50.61 (14.85)4.11
< .01
The above table will show that the achievement of boys and the combined sample from the Different group is significantly superior to their counterparts from the Same group. However, though the Girls from Different group have scored higher, the mean difference has failed to reach the test of statistical significance.TABLE 3
Analysis of Covariance of Scores on PMT and AchievementA perusal of the above table will show that when the influence of intelligence is controlled that achievement differences between the girls, boys and combined samples of the two groups becomes insignificant. This shows that the difference in medium of instruction does not being significant difference in achievement, which is only found when intelligence is not controlled.
Groups
F
P
girls
0.18
NS
boys
3.76
NS
combined
1.03
NS
TABLE 4
Mean, SD, t and Significance levels of Inter-group Differences in First Language Achievement
Groups
Mean (SD)
t
p
girls
Different (N=65)
Same (N=41)69.18 (13.30)
55.90 (11.33)5.25
< .01
boys
Different (N=50)
Same (N=43)71.40 (15.60)
59.14 (15.12)3.94
< .01
combined
Different (N=115)
Same (N=84)70.14 (14.37)
57.56 (13.50)6.33
< .01
It is evident from the above table that the achievement in the first language of girls, boys and combined sample from the Different groups is significantly superior to those from the Same group. Even though for the Different group the medium of instruction is different from the first language its performance is better than those who have been taught from the very medium of the first language.
TABLE 5
Analysis of Covariance of Scores on PMT and First Language AchievementThe insignificant F values in the above table show that when the effect of intelligence is controlled, the significant differences, as reported in Table 4, in the first language achievement of boys and girls from the Same and Different groups, disappear. However, even after controlling intelligence, a significant difference remains between the combined samples from the two groups. TABLE 6
Groups
F
P
girls
0.99
NS
boys
2.79
NS
combined
11.17
< .01
Mean, SD, t and Significance levels of Inter-group Differences in Second Language Achievement
As will appear from the above table, the second language achievement of the girls, boys and their combined samples from the Different group is significantly superior to their counterparts from the Same group.
Groups
Mean (SD)
t
p
girls
Different (N=65)
Same (N=41)75.83 (11.39)
51.07 (14.77)9.60
< .01
boys
Different (N=50)
Same (N=73)72.26 (16.76)
54.67 (17.59)5.08
< .01
combined
Different (N=115)
Same (N=84)74.28 (14.08)
52.92 (16.35)9.89
< .01
TABLE 7
Analysis of Covariance of Scores on PMT and Second Language AchievementThe significant F values in the above table show that the significant difference in the second language performance of the two groups. Same and Different - is existent even after the influence of intelligence has been controlled.
Groups
F
P
girls
16.02
< .01
boys
8.60
< .01
combined
33.69
< .01
TABLE 8
Mean, SD, t and Significance levels of Inter-group Differences on Fluency Dimension of the Verbal Test of CreativityA perusal of the above table will show that in all the three comparisons made between the Same and Different groups, the latter has been found to be significantly superior. It appears that those taught through the medium of a language different from their first language have the tendency to come up with more ideas on a problem as against those taught through the medium of their first language.
Groups
Mean (SD)
t
p
girls
Different (N=65)
Same (N=41)29.09 (10.35)
17.67 (7.52)5.39
< .01
boys
Different (N=50)
Same (N=43)48.47 (25.02)
25.95 (16.75)4.35
< .01
combined
Different (N=115)
Same (N=84)35.98 (19.44)
22.03 (13.83)5.74
< .01
TABLE 9
Analysis of Covariance of Scores on PMT and Scores on Fluency DimensionIt appears from the above table that when intelligence is controlled, only the boys from the two groups - Same and Different-differ significantly in fluency of ideas. The girls and the combined sample of boys and girls from the two groups do not differ significantly.
Groups
F
P
girls
4.93
NS
boys
8.01
< .01j
combined
5.57
NS
TABLE 10
Mean, SD, t and Significance levels of Inter-group Differences on the Flexibility Dimension of the Verbal Test of Creativity
It is evident from the above table that on the flexibility dimension, the two groups-Same and Different-differ significantly between themselves in all the comparisons made between them. The different groups are throughout superior to the Same groups and therefore, in face of any problem in a changed situation, they have the tendency to be more flexible in their approach.
Groups
Mean (SD)
t
p
girls
Different(N=65)
Same (N=41)19.33 (6.30)
11.28 (4.57)6.24
< .01
boys
Different (N=50)
Same (N=43)26.19 (8.48)
15.33 (9.11)4.94
< .01
combined
Different (N=115)
Same (N=84)21.77 (7.87)
13.41 (7.60)7.60
< .01
TABLE 11
Analysis of Covariance of Scores on PMT and Flexibility Dimension
A perusal of the above table will show that even after the influence of intelligence has been controlled, the difference reported on the flexibility dimension, between boys and girls and the combined sample from the Same and Different groups remain significant. The Different groups are throughout superior to their counterparts from the Same groups.
Groups
F
P
girls
33.23
< .01
boys
7.83
< .01
combined
14.72
< .01
TABLE 12
Mean, SD, t and Significance levels of Inter-group Differences on the Originality Dimension of the Verbal Test of Creativity
A glance at the above table will show that on the originality dimension of creativity test, except for girls whose mean score is higher though not statistically significant, those from Different group score significantly higher than their counterparts from the same group.
Groups
Mean (SD)
t
p
girls
Different (N=65)
Same (N=41)4.17 (5.20)
2.67 (2.94)1.49
NS
boys
Different (N=50)
Same (N=43)12.22 (19.11)
4.20 (6.48)3.89
< .01
combined
Different (N=115)
Same (N=84)7.03 (7.66)
3.47 (5.18)3.56
< .01
TABLE 13
Analysis of Covariance of Scores on PMT and Originality Dimension
Groups
F
P
girls
0.04
NS
boys
6.22
NS
combined
1.06
NS
The above table shows that when the influence of intelligence is controlled, no significant difference is found in originality of ideas between the girls, boys and their combined sample from the two groups - Same and Different.
TABLE 14
Mean, SD, t and Significance levels of Inter-group Differences on Elaboration Dimension of the Non-Verbal Test of Creativity
Groups
Mean (SD)
t
p
girls
Different (N=65)
Same (N=41)25.03 (6.37)
18.87 (5.75)5.08
< .01
boys
Different (N=50)
Same (N=84)23.40 (6.12)
21.20 (5.35)1.69
NS
combined
Different (N=115)
Same (N=84)24.87 (6.40)
20.18 (5.65)5.39
< .01
It appears from the above table that boys from the Same and Different groups do not differ significantly. However, in the other two comparisons made, that is for the girls and the combined sample of boys and girls from the two groups, a significant difference has been reported. It appears that in the capacity to elaborate and build on the basic ideas, the girls from the Different groups are significantly superior to those from the Same groups.
TABLE 15
Analysis of Covariance of Scores on PMT and Elaboration Dimension
Groups
F
P
girls
16.15
< .01
boys
0.36
NS
combined
12.09
< .01
The above table shows that even after controlling intelligence significant differences have been found between the girls and the combined samples from the Same and Different groups. However, no significant difference has been reported in the cases of the boys.
TABLE 16
Mean, SD, t and Significance levels of Inter-group Differences on Originality Dimension of Non-Verbal Creativity Test
Groups
Mean (SD)
t
p
girls
Different (N=65)
Same (N=41)4.62 (3.62)
3.39 (3.09)1.58
NS
boys
Different (N=50)
Same (N=43)6.70 (4.64)
6.03 (4.04)0.68
NS
combined
Different (N=115)
Same (N=84)5.49 (4.20)
4.87 (3.89)1.07
NS
A glance at the above table will show that for all the comparisons made between the two groups - Same and Different, no significant difference has been found on the originality dimension.
TABLE 17
Analysis of Covariance of Scores on PMT and Originality Dimension of Non-Verbal Creativity Test
The insignificant F values reported in the above table further confirm the fact that the boys, girls and their combined samples from the Same and Different groups do not differ significantly so far as uncommonness of the responses are concerned.
Groups
F
P
girls
2.09
NS
boys
0.02
NS
combined
0.02
NS
TABLE 18*
Mean, SD, t and Significant levels of Inter-group Differences on Several Variables for Boys from Ramakrishna School.
Sl. NO.
Variable
Groups
Mean (SD)
t
p
1
PMT
Different
Same42.21 (10.60)
47.40 (10.14)0.24
NS
2
Achievement
Different
Same64.23 (9.92)
64.20 (9.15)0.01
NS
3
First language
Different
Same74.97 (12.79)
74.27 (12.20)0.17
NS
4
Second language
Different
Same79.23 (11.74)
74.13 (12.56)1.32
NS
5
Fluency (Verbal)
Different
Same56.09 (24.78)
45.15 (15.15)1.41
NS
6
Flexibility (Verbal)
Different
Same28.83 (6.82)
25.85 (6.96)1.21
NS
7
Originality (Verbal)
Different
Same16.26 (9.05)
11.15 (7.12)1.71
NS
8
Elaboration (Non-Verbal)
Different
Same22.73 (7.36)
17.71 (4.35)2.37
0.05
9
Originality (Non-Verbal)
Different
Same7.93 (4.28)
6.43 (4.47)1.07
NS
The above table will show that no significant difference is found between the two groups on any of the variables studied, when both the samples are drawn form the same school. Keeping in view the significant differences reported earlier, it appears that the difference existing between two schools in the form of environmental and other relevant variables like teaching methods, materials and aids used for teaching and the general school climate may also influence the performance of the pupils.
Three clear-cut trends emerge from the data. The first one is the allround better performance of these whose MI is different form their MT in comparison to those for whom there is no difference between the two. They have scored higher not only in intelligence and general achievement but also in respect of the other variables studied except the originality dimensions of the nonverbal test of creativity where no significant difference has been found between the two groups (cf Tables 1,2,4,6,10,12,14 and 16). Secondly, when the influence of intelligence is controlled significant difference is found only for a few of the several comparisons made separately for boys, girls and their combined sample from the two groups (cf Tables 3,5,7,9,11,13, 15 and 17). Thirdly, when the two groups - MT-MI Different and Same - are selected from the same school and comparison is made between them on the variables studied, no significant difference in their performance is found on any of the several comparisons made except at 0.05 level for the elaboration dimension of nonverbal test of creativity (cf. Table 18).
The above trends lead to the following two implications which may form the hypotheses for future studies to arrive at some conclusion.
(a) The most one is that these findings lend support to a particular type of bilingual education where a language other than the MT of the group is medium of instruction. This type of bilingual education has at least no harmful effect on the growth of mental ability and creative behaviour and also on the allround academic achievement including in the first and second languages.
(b) The difference in the allround performance of the two groups - Same and Different - may also be due to the difference existing between them in intelligence and school related variables like teaching methods, materials and aids used for teaching and the general school climate. Therefore, all oft reported better performance of English medium schools over mothertongue medium schools may also be attributed to the fact that the former not only from the very beginning select bright children from upper SES, but also that the school conditions prevailing therein are more conducive to achieving behaviour.
The difference in the intelligence of the two groups might have been present from the very initial stage of their schooling and this, the students from MT - MI Different group have been able to maintain till tested in this study. We have no confirmatory evidence for this possibility. However, the MT-MI Different group has been selected from the English medium children from higher SES join such schools. They can both compete in the admission tests and also afford the high fees charged by these schools. On the contrary in many of the MT medium schools admissions are not so competitive. Generally lower SES groups and also those who could not be selected for the English medium, join such schools.
The second reason for the high intelligence score or MT- MI Different group of students may also be due to the effect of medium of instruction itself which is in a language different form their mothertongue. The different medium which was introduced from the very beginning of the schooling might possible have made the students cognitively more flexible.
A few other Indian studies have also reported higher intelligence scores for the English medium students (Anand, 1971, Dave and Dave, 1971). A large number of immersion studies, already discussed earlier, have reported no detrimental effect on the cognitive faculties of the child if he is educated from the very beginning through a medium other than his MT. In fact Barik and Swain (1976c) found that for the first three years there might not be significant difference in the IQ of the two groups - immersion and regular. But over the five year period, the immersion group, particularly the high achieves among it, had higher IQ than the regular group.
Ben-Zeev (1976) on the basis of this studies supports the hypothesis that language acquisition is enhanced by cognitive conflict since language is learnt through adjustment to this conflict. Bilingualism induced an intense degree of cognitive conflict and that is why it is beneficial for language learning. In another study Ben-Zeev (1977) found support for the hypothesis that a bilingual child was forced to develop coping strategies, conducive to cognitive development, to meet the mutual interference created between the two languages. Though both the English and Kannada medium students representing the two groups compared are bilinguals, the former have been taught from the very beginning through a medium other than their MT. In the latter case bilingualism has meant just the introduction of the second language at a later stage as a small part of the curriculum. Therefore, the chances of arousal of cognitive conflict and mutual interference between the two languages are much more in the case of those whose medium of instruction is different from the mothertongue from the very beginning of their schooling.
One very crucial factor in the bilingual development of a child is the age when second language learning is introduced. Quite a number of studies have reported that for several reasons early introduction of the second language is more beneficial for the allround development of the child (Srivastava, 1976). Engle (1975) reviewing the evidence on the topic suggests that since the age between six and eight is the age for early concrete operation, the introduction of a second language at this stage creates more difficulties for the child than its introduction before 6. Ianco-Qarral (1972) also suggests that simultaneous introduction of two languages in early childhood speeds up the separation of semantics and phonetics. In all the immersion studies bilingualism was introduced right from the very beginning. In this context if we examine the English and MT medium schools we find that in the former the students are simultaneously introduced to the first and second languages right form the very beginning of their schooling, at least before the age of 6. On the other hand, the MT medium schools normally do not simultaneously introduce the second language, but do it a later stage.
The above discussion attempts to explain why in comparison to MT-MI Same group, the MT-MI Different group of students have scored higher on the nonverbal test of intelligence. It also explains their significantly better performance in respect of other variables also.
The second trend shows the interaction of intelligence on the performance of the two sample groups. Without controlling intelligence boys from the English medium are found to be significantly superior to their counterparts in general achievement, first and second language performance, fluency, flexibility and originality dimensions of verbal creativity test. After controlling intelligence, significant difference is found only for second language, fluency and flexibility dimensions. The girls from English medium without controlling intelligence, are found superior in first and second language, fluency and flexibility and elaboration dimension of the nonverbal test of creativity. After the influence of intelligence is controlled significant difference exists only for second language, flexibility and elaboration. Similarly without controlling intelligence the combined group from English medium schools are found superior to their counterparts in general achievement, first and second languages fluency, flexibility and originality dimensions of verbal and elaborations of nonverbal test of creativity. After intelligence has been controlled, they are found superior only in first and second languages, flexibility and elaboration. To sum up, in all the 24 comparisons made between the two groups, significant difference was found in 18 cases. After controlling intelligence, it dropped down to 10 only.
The above trends show that the difference often reported in the achievements of English and MT medium schools reflects to a great extent the difference in the intelligence of the two groups and the school conditions. It has already been emphasized that the English medium education is the first choice for the bright children from the higher SES. They therefore, start with a double advantage. Moreover, the conditions prevailing in the two types of schools may also be responsible for the difference in the performance. Most of the English medium schools are not only autonomous but also charge high fees besides getting grants also. They have less constraints and have more resources to offer facilities for learning. The conditions in the mothertongue medium schools are very different. They have to meet the needs of the bulk of population with inadequate resources.
The third trend reveals the interaction of school related variable on the performance of the two groups. When they were selected from several schools, significant difference was found between them on almost all the variables. However, when both English and Kannada medium students were selected form the same school, no significant difference between them was found on any of the variables except for one on .05 level of confidence. These two groups come from Ramakrishna school which is a residential school with reputation for not only maintaining high standard of teaching but also for providing the pupils with excellent conditions conducive to their allround growth.
We shall not take up each variable, achievement, first and second language performance and creativity separately for discussion.
The insignificant difference found between the two groups in achievement after controlling intelligence is quite expected. This is because of the significant and positive correlation that exists between intelligence and achievement, the model value of being around 0.5 (Srivastava, 1967). The findings of the immersion studies lend support to these results. Anand (1971), however, found the Kannada medium students achieving significantly better than the English medium ones, whether or not intelligence was controlled. This striking difference may be due to the different criteria selected for achievement by each study. Anand used achievement tests in mathematics, general science and social studies. For the present study the criterion was average of previous six consecutive examination marks obtained in all the academic subjects. Achievement tests measure a pupil's achievement at a particular point of time in his life. On the other hand, the marks of school examinations spread over several years measure a pupil's achievement in a course of time reflecting ups and downs and the cumulative effect of several years of his school life on his achieving behaviour.
All the groups from the English medium schools have been found to be performing significantly better than those from the Kannada medium, both in first and second language is concerned the results are not at all surprising considering the emphasis given on English, which is also the MI, in English medium schools. Therefore, even after the influence of intelligence is controlled all the groups from these schools are able to maintain their superiority over their counterparts (cf Table 7). No significant difference however, is found in the case of first language achievement of boys and girls form the two types of schools after intelligence is controlled. But a significant difference is found between the two combined samples. The English medium students have performed significantly better even in first language performance. This result is surprising since in the Kannada which is both the MT and MI, for the students. On the whole, the English medium students perform better in both the first and second languages. It appears that in the case of English medium students there has been transfer of skills across the two languages which Lambert and Tucker (1972) have called the process of two-way bilingual relay. Vygotsky (1962) observed that '… a foreign language facilitates the higher form of native language'. (p. 110) However, the quality of this two-way bilingual relay process is comparatively poorer in the case of Kannada medium students. One explanation may be found in the difference in the conditions of language teaching prevailing in the two types of schools. This is borne out by the fact that no significant difference was found between the two groups both in fist and second language achievement when they were selected from the same school. At this stage a difference has to be made between additive and subtractive bilingualism. In the former L2 competence adds to L1 competence. In the latter it subtracts or it does not add. Possibly, what the Kannada medium students in this study are acquiring in the name of bilingual education is verging more towards subtractive bilingualism. In the case of English medium students of this study it may be approaching towards additive bilingualism.
In all the three dimensions of verbal creativity test, fluency, flexibility and originality, a significant difference has been found in all the intergroup comparisons made between the two types of groups, except in the case of girls who do not differ between themselves on the originality dimension (cf Table 8, 10 and 12). The MT-MI different students have been found to be better, though after intelligence has been controlled, significant difference was found only in the flexibility dimension. In the case of boys, however, the differences have been found significant also for fluency (cf Tables 9, 11, 13). The higher creativity score of MT-MI difference may be attributed to their higher intelligence score since a positive significant correlation has been reported between the two variables (Martin, et al. 1977). Khire (1976) found significant but low positive correlation between creativity and Progressive Matrices Test, the one administered in the present study. On the nonverbal test no significant difference was found between the two samples on originality dimension. (cf Tables 16 and 17). On the elaboration dimension girls and the combined samples from the English medium schools were found to be significantly better than their counterparts under both the conditions when intelligence was controlled and when it was not. Boys however, did not differ significantly. (cf Tables 14 and 15).
The most striking difference between the two samples is found on the flexibility dimension. In a few other inter-group comparisons made on other dimensions, significant difference has been found. Two reasons can be assigned for the better performance of the English medium students. First is the influence of reinforcement on the development of creativity (Locurto andWalsh, 1976; Milgram and Feingold, 1977). Possibly, either in school or in the home surrounding there are encouragements and conditions present for the growth of creativity. Secondly, Cummins (1977) has proposed that '… there may be a threshold level of bilingual competence which an individual must attain before his access to the two languages can begin to positively influence his cognitive functioning." (p.11) In his study Cummins found that the divergent thinking skills are related to the different levels of functional bilingualism attained by the balanced and non-balanced bilinguals. More balanced the bilingualism, the higher was the score on the test of creativity. In the present study, though attempts were made to control the level of proficiency of each group in first and second languages, the two groups differed in the level of functional bilingualism attained. This was evident from their respective performances in first and second language. These explain the significant difference between the two groups on some dimensions even after the influence of intelligence was controlled.
1. The authors thank Dr. D.P. Pattanayak, Director, CIIL for his valuable suggestions given from time to time. Throughout he remained a source of strength and inspiration.
2. Thanks are also due to all the Headmasters/Headmistresses of the schools selected for data collection, and also to the boys and girls from these schools who formed the samples for this study. Without their cooperation it would not have been possible to complete this work at all.
3. Dr. G.C. Ahuja and Dr. H.R. Dua gave their suggestions at the initial stage of the work. Sri. B. Mallikarjun helped in the translation of the Test in Kannada. Sri. R.K. Bhat helped in scoring some of the Kannada answer sheets. Sri. B.D. Jayaram and Miss Vani of Data Processing Unit helped in statistical treatment of the data. We are extremely thankful to all of them.
Anand, C.L. The study of the effect of socio economic environment and medium of instruction on the mental abilities and the academic achievement of children in Mysore State. Ph.D. dissertation, Mysore University, Sept. 1971.
Ausubel, D.P. Educational Psychology: a Cognitive view. HRW, 1968.
Barik, H.C. and Swain, M. Three-year evaluation of a large scale early grade French immersion programme: The Ottawa Study. Language Learning, 1975, 25, 1-30.
Barik, H.C. and Swain, M. English-French bilingual education in the early grades : The Elgin study through grade four. The Modern Language J., 1976, 60, 3-17. (a)
Barik, H.C. and Swain, M. A Canadian experiment in bilingual education: The Peel study. Foreign language Annals, 1976, 9, 465-479. (b).
Barik, H.C. and Swain, M. A longitudinal study of bilingual and cognitive development. International J. Psychology, 1976, 11, 251-263. (c)
Boneau, C.A. The effects of violations of assumptions underlying t test. Psychological Buletin, 1960, 57, 49-64.
Ben-Zeev, S. The influence of bilingualism on Cognitive Strategy and cognitive development. Child Development, 1977, 48, 1009-1018.
Cowan, R.J. and Sarmad, Z. Reading performance of bilingual Children according to Type of School and Home Language. Language Learning, 1976, 26, 353-376.
Cummins, J. The influence of bilingualism on cognitive Growth: A synthesis of research findings and explanatory hypothesis. Working papers in Bilingualism No. 9, 1976, 1-4.
Cummins, J. Cognitive factors associated with the attainment of intermediate levels of bilingual skills. Modern Language 1977, 61, 3-12.
Dave, P.N. and Dave, J.P. Socio-economic environment as related to the nonverbal intelligence of rank and failed students. Research Reports in Education and Psychology. Regional College of Education, Mysore, 1971.
Education and National Development. Report of the Education Commission (1964-66), NCERT, 1970.
Engle, P.L. Language medium in early school years for minority language groups. Review of Educational Research, 1975, 45, 283-325.
Ianco-Worrall. Bilingualism and cognitive development. Child Development, 1972, 73, 1390-1400.Kamakshi, A Study of the relative performance of students of English and Kannada medium classes in general mathematics of St X.J. of College of Education, (Karnatak University), Dharwar, 1966, 4, p. 66.
Khire, U. A report on exploration study of Advanced Progressive Matrices, Growth of reasoning capacity and some Correlates. Indian J. Applied Psychology, 1976, 13, 33-37.
Lambert, W.E. and Tucker, G.R. Bilingual education of Children: The St Lambert Experiment. Rowley, Newburg House, 1972.
Locurto, C.M. and Walsh, J.F. Reinforcement and self reinforcement: their effect on Originality. American J. Psychology, 1976, 89, 281-291.
Macnamara, J. Bilingualism and Primary Education. Edinburg. Edinburg Unviersity Press, 1976.
Martin, J.D. et al., Correlaiton of the object assembly and Block design test of Wechsler Adult Intelligence Scale and the Torrance tests of Creative thinking. Educational and Psychological Measurement, 1977, 37, 1095-1097.
McLaughlin, B. Second language learning in children. Psychological Bulletin, 84, 438- 459.
Meddis, Ray. Statistical handbook for nonstatisticians. McGraw Hill, 1975.
Mehdi, B. Verbal Test of Creative Thinking. National Psychological Corporation, Agra 1973 (a).
Milgram, R.M. and Feingold, S. Concrete and verbal reinforcement in creative thinking of disadvantage children. Perceptual Motor Skills, 1977, 45, 675-678.
Prator, C.H. Jr. Language teaching in the Philippines, 1950.
Raven, J.C. Guide to the Standard Progressive Matrices, Sets, A, B, C, D and E 1960, London, Lewis.
Reddey, A.V.R. Instructional Medium at University Level. The Hindu, July 6 (Wednesday), 1977.
Sakrikar, D. Mother tongue is best. The Illustrated Weekly of India, March 4, 1979.
Skutnabb-Kangas, T. Bilingualism, Semilingualism and School Achievement. Linguistische Berichte, 1976, 45, 55-64, in LLBA Abstracts, April 78, 12.
Srivastava, A.K. An Investigation into the Factors related to Educational Underachievement. Patna University Ph.D. thesis, 1967.
Srivastava, A.K. An examination of the Psychological viewpoints on the effects upon the learner of learning more than one language. (Cyclostyled) Paper presented on the occasion of Language Load Colloquium, Central Institute of Indian Languages, 1976, Mysore.
Srivastava, A.K. et al., The Language Load. CIIL Occasional Monographs Series, 1978. Mysore.
UNESCO. The use of Vernacular Languages in Education. Monograph on Fundamental Education, 1953.
Vygotsky, L.S. Thought and Language. Cambridge, Mass: MIT. 1962.
Yela, M. Verbal Comprehension and bilingualism. Revista de Psycologia General. Aplicada, 1975, 30, 1039-1046. In Psychological Abstracts (Abstract No. 5297), March 78, 59.
