Results,Interpretation & Implications

Before the results regarding the acceptance or rejection of major and minor hypothesis are presented and discussed, it is quite necessary to detail the points related to validity and reliability of the pre-test. This is of particular significance, since the same test was also used as the post-test.

At the outset, it is conceded that the preparation of the pre-test was not as systematically pursued as it is conventionally done in constructing achievement tests. In other words, since the major aim of the project was to develop the Bridge Course and as it was a very urgent need, a full-fledged project of standardizing a test inclusive of the five language skills was not feasible. Apart from this, it was thought that in a way the project, being an experimental one, had built-in elements of sensing and demonstrating face validity. Because of this shortcoming, utmost care was taken in preparing the pre-test. The procedure for selecting passages for the Bridge Course and the pre-test was very well thought of in order to ensure a high content validity. The following steps, namely, the appointment of a top level committee of creative writers, language professors, linguists, textbook writers, and educationists, scanning and screening of materials from a variety of sources such as textbooks, encyclopedias, popular magazines and then making a wide selection of passages related to academic and non-academic topics of general interest to P.U.C. entrants, the final selection and grading of the passages following systematic and scientific procedures using common criteria, analyzing the content and the language of the selected passages, and finally framing various types of questions with respect to the five language skill objectives, in no small way contributed to the content validity of the test. Thus, the pre-test did contain sufficient and substantial language materials for testing what it purported to measure. It must be pointed out here that as a pioneer research project in language education, this attempt represents a noteworthy step, for, ordinarily a research worker would not have been able to muster the support and help from so many experts.

One of the techniques adopted by psychometricians to establish the validity of a test under study is to correlate it with a number of available tests which claim to measure the same characteristic such as Intelligence, Aptitude or Language Skills. If correlational indices are high and of course significant (usually above .70), the test constructor feels that he has satisfied one of the technical objections regarding validity. This is known as concurrent validity, as against the concept of divergent validity that can be taken as one, which, by the virtue of low correlational indices shows that the tests, though believed to be measuring the same characteristics, are in fact independent, having only a small amount of identical elements. In simple words, they do not measure the same characteristic or even though they are meant to measure a particular global concept, the components which constitute the concept are divergent and hence independent.

It is felt that the application of this concept in explaining the indices of correlations presented in Table 4 will be quite appropriate.

CORRELATION COEFFICIENTS SHOWING RELATIONSHIPS AMONG THE FIVE LANGUAGE SKILLS AND TEST-RETEST RELIABILITY BETWEEN THE PRE- AND POST-TEST SCORES OF THE EXPERIMENTAL AND THE CONTROL GROUPS FOR EACH SKILL

	LC	LNC	RC	GC	EP
LC	--	.506	.73	.319	.242
LNC	--	--	.516	.328	.271
RC	--	--	--	.472	.196
GC	--	--	--	--	.206
EP	--	--	--	--	--

All r values are significant beyond .001 level. N=730

Group

LNC

TLS

Experimental
Control

.470
.471

.680
.653

.439
.554

.438
.177*

.324**
.348

.789
.713

Attention needs to be drawn to the fact that uptil now language ability, like academic achievement, has been considered as a single, global characteristic. Rarely, scientific attempts have been made to analyse these characteristics and investigate their true nature indeed a very important question from the point of view of validity. The very fact that the team of linguists and psychologists felt the need for dividing the so called language comprehension into 5 independent objectives for development and testing is a sufficient proof for making an assumption that significant differences exist among them and hence the abilities required for mastering them may be different and independent. That such an assumption was implied in the objectives and one of the hypothesis of the project needs no comphasis. To reinforce the above argument, the investigators would like to quote Thorndike and Hagen (1970) on this subject. They state: "Validity is always specific to a particular curriclum (under line is ours) or a particular job… The test must always be evaluated against the objectives of a specific program of instruction… Validity must always be in relation to a situation as similar as possible to the one in which the measure is to be used."

Inter-correlations among the five skills. The indices of correlations among the five skills presented in Table 4 lend substantial support to the assumption of independence of the five skills discussed above. The range of indices is .196 to .516, not very high indeed. They all are significant at .001 level. These results derive double support from the finding of significant differences among the sum of ranks of the rive skills obtained through the Friedman's ANOVA by Ranks (difficulty) and from the McQuity's test of showing cumulative hierarchy among the same (see the discussion on both later in the chapter). The r coefficients indicate that at best just moderate correlations exist between LC-LNC, LNC-RC and RC-GC. The other correlations are quite low, i.e., below .37, thereby indicating more divergence and independence than identity of elements among them. The indices show extremely meaningful trends. The correlation of LC to LNC is the second largest (.506), and that is what it ought to be. While it shows that LC and LNC certainly have some identical elements, it also shows that LNC contains elements which are different from those in LC. It is extremely interesting to note that the highest correlation exists between LNC-RC (.516), again indicating similarity as well as divergence of elements between them. It is obvious that LNC has something common with LC and with RC. Is it possible that the authors who have, at the risk of being excessively repetitious, made reference to audile and visual symbols might have identified the right processes underlying language learning? The data surely substantiate this assertion. LC and GC should have a lower correlation than that between RC and GC for, common visual symbols between the latter set of skills will exceed that of the former. The correlation .472 being greater than .319 lends clear support to it. In the same vein, both the sets RC-LNC and RC-GC, again having common visual symbols, show a higher correlation than the one obtained for RC-LC (.516 .478 > .373). In fact, minute examination of the correlations in Table 4 reveals too obvious a trend such as mentioned by cross-examining the indices. The first row shows that the correlations of LC to LNC, RC, GC and EP go on decreasing, as common audile elements also go on decreasing (.506, .373, .319 and .242) LNC, truly a combination of audile and visual symbols, keeps up its more or less delicate balance between the two as its correlations to LC, RC, GC and EP shows the following value i.e., .506, .516, .472 and .196. Similarly, RC also follows the rule; its correlation to LC is low (.373); while the magnitude of those to LNC and GC (.516 and .472) has sufficiently increased, that to EP has been reduced to .196. Of course, this drop needs explanation. This warrants examination of EP's correlations to all the other four skills. They are very low, indicating thereby extremely limited similarities to the others. The question is, why should the correlation between EP and RC be lower than that between EP and LC, when it essentially involves manipulation of visual symbols? The only answer to this may be that the abilities or mental processes needed for manipulating two languages together may, although inclusive of all other skills, be still different, and perhaps visual symbols of the mother-tongue might be interfering because of identical elements (in two language passages) demanding two different response patterns from the learner. Finally, GC's correlations to the other skills more or less show the same trend, the largest being to RC, the lower to LNC, still the lower to LC and the lowest to EP.

In conclusion, it needs re-emphasis that the data demonstrate more divergent than concurrent, more independent than dependent nature of skill objectives reported in this book. Since the investigators started with clear cut objectives using a specific curriculum schema, it seems justifiable to make a claim for divergent validity for the evaluation test. Doubtless, these findings are quite revealing and hence extremely useful in elaborating and expanding the concept of language attainment for further research.

Test-Retest Reliability

The administration of the same test as the post-test to both the Experimental and the Control groups provided the investigators an opportunity for demonstrating its test-retest reliability. The correlation coefficients showing the test-retest reliability of the entire test as well as the sub-tests are presented in Table 4. The overall values showing association are very high indeed. (.789 and .713 for the Experimental and Control groups respectively) although, the coefficients for different sub-tests vary from a very low value of .139 to a considerably high value of .680. The correlations of the Experimental Group are quite satisfactory, although the r value of .324 for EP may not be so considered. The indices for the Control Group for GC and EP are also quite low. It will be safe to conclude that while the total test possesses high consistency and accuracy, the sub-tests can be claimed to have them only in moderate.

The experimental part of the project was planed solely with a view to evaluating the efficacy of the prepared materials and the procedures of teaching therein suggested in promoting language skills which seemed to have not been fully developed due to neglect in schools. The data presented in Table 5 lend clear support to the major hypothesis formulated for testing. The F values with respect to the total language skill (42.90) and with respect to the other skills, except LC, i.e., LNC, RC, GC and EP (34.15, 36.9, 14.20 and 19.99) were significant beyond .005 level. The F value for LC was significant at only .05 level. These F values were obtained through the covariance statistical designs where the effect of the pre-test scores was partialled out and consequently only the significant differences between the gains of the Experimental and the Control Groups were tested. The data strongly support the first part of the hypothesis that the college entrants trained through the Bridge Course in Kannada would be better in their language skills than those who were not. It is quite pertinent to point out here that since both the groups were more or less the same on the pre-test (see Table 5), the statistically significant gains by the Experimental Group pitted against those of the Control reflected very favourably on the effectiveness of the training given through the newly developed Bridge Course. Let it be stressed again that this strongly supports the claim of validity made earlier.

ANALYSIS OF COVARIANCE FOR THE PRE AND POST TEST SCORES OF THE
EXPERIMENTAL AND THE CONTROL GROUPS ON LANGUAGE SKILLS
(df = 1/166)

Skill

Test

Scores

Percentages

Values

Level of signifi- cance

Means of

Exp. Group

Means of the Control Group

Means of

Exp. Group

Means of the Control Group

Gains of

Exp. Group

Gains of the Control Group

TLS

Pre

Post

111.83

155.32

109.58

131.73

39.14

53.13

38.24

44.41

13.99

6.17

42.90

P < .005

Pre

Post

17.08

23.04

16.44

20.79

34.16

46.08

32.88

41.58

11.92

8.70

6.60

P<.005

LNC

Pre

Post

41.04

57.71

38.31

47.94

41.38

57.71

38.31

47.94

16.63

9.63

34.15

P<.005

Pre

Post

36.65

53.05

38.62

45.69

52.36

75.79

55.17

65.27

23.43

10.10

36.90

P <.005

Pre

Post

10.73

12.04

9.75

9.52

42.92

48.16

39.00

38.08

5.24

-.92

19.99

P <.005

Pre

Post

6.29

9.48

6.46

7.79

25.16

37.92

24.85

29.16

12.76

3.32

14.20

P<.005

CHI SQUARE VALUES AND FREQUENCY OF PASSED AND
FAILED STUDENTS IN THE EXPERIMENTAL
AND CONTROL GROUPS

Group	Final P.U.C. Examination		Chi Square
Group	Passed	Failed	Chi Square
Experimental	30	46	5.44
Control	17	60	P < .02

The second part of the hypothesis, that these entrants would also show better academic achievement than their counterpart Controls, was tested by the Chi Square. The data presented in Table 6 show that the Chi Square value of 5.44 is at .02 level of significance, thereby lending support to the above hypothesis. The investigators would like to interpret this part of the hypothesis a little cautiously, as they are aware of two major weaknesses; first, the sampling bias which was frankly pointed outing the description of the selection of the subjects for the Experimental Group and second, the absence of pre-measurement of academic achievement of both the groups. Both the points are quite relevant and need some discussion before the interpretation of results is attempted. Doubtless, the bias was present in the selection of the subjects for the Experimental Group. With respect to the second point, it should be readily accepted that a similar kind of the subjects for the Experimental Group. With respect to the second point, it should be readily accepted that a similar kind of bias regarding academic achievement might have been also present in the same group, i.e., it may be argued that to start with the Experimental Group had an edge in academic achievement over the Control Group. Since the data pertaining to their previous academic achievement is unfortunately not available, one guess is as god as the other. Just the same, the investigators would like to present their points, of view and offer some clarification and explanation about this.

One point needs re-emphasis. The Control subjects were free from any such bias as their selection was done by following a systematic random procedure. The question that needs to be explored and satisfactorily answered is whether these randomly selected Controls can be considered equal or better or poorer in language skills and academic achievement than those in the Experimental Group. The means of the Experimental and the Control Groups on the pre-test for al the five skills presented in Table 5 very clearly show that the differences between them were negligible, suggesting thereby that both the groups were equal in their language skills. What about their academic achievement? Would it be prudent to assume, particularly in the light of the data of no difference in language skills just discussed, that the students who were quite good in their academic achievement tended to volunteer for the course? A commonsense guess which the investigators would like to accept and which was supported by the general observations of the lecturers conducting the course was that students opting for the Kannada medium were academically inferior to those opting for the English medium and, consequently, those volunteering for the course were the weak ones strongly felt the need for such training to better their academic achievement in the public examination. Having rigorously argued the point, the investigators still would not like to accept this assumption and proceed. Instead, they would prefer to accept the assumptions that both the groups were equal in their academic achievement and had equal initial potential to pass the examination. If this is accepted, which seems reasonable, the conclusion from the reported result (Table 5) is straightforward. That is, the improvement in their language skills seemed to have helped the Experimental Group to better their academic performance.

Support to Theoretical Schema
These results extend support to the basic assumptions implied in the theoretical schema propounded in Chapter 1. Since this is the first systematic empirical testing of the schema, it is appropriate that a full discussion on it is presented. The statistically significant gains in language skills attained by the Experimental Group strongly indicate that instruction pursued with clear cut, observable and measurable skill objectives (Input) and appropriate and suitable teaching methods (Process) not only improves learning but also does it with far more economy and efficiency. The reader could not have missed the point that the Bridge Course in fact filled the gap, within a short period of a month or two, which was the result of not one, two or three years but of more than ten years or so. As a matter of fact, it is proper to remind the reader that the basic reasons for undertaking this developmental-cum-experimental project were (1) unsatisfactory state of teaching of the mother-tongue, (2) lack of proficiency on the part of the college entrants in language skills needed for pursuing further studies, and (3) interference of this deficiency in increasing their content competence in the chosen field. This experimental evidence raises a genuine interest which can be stated in hypothetical terms as follows: What will be the gains if from the primary stage itself the teaching of language is pursued with reference to clear cut and specific goals stated in terms of language skills and by utilizing a variety of methods specifically designed to promote continuous and frequent interactions between the teacher and the taught? Although no systematic evaluation of these interactional processes was attempted, the investigators have a feeling that in no small measure did they play their part in raising the language achievement levels of the Experimental Group. Equally important part was played by the new techniques of evaluation employed in the Bridge Course. Two major points need attention in this regard: one, every question was objective-based and two, evaluation was continuous; so continuous that, as was pointed out earlier, it was difficult to distinguish where instruction stopped and evaluation of the imparted instruction started. Both these points are of extreme relevance and importance, especially when the teaching of the mother-tongue to a great majority of teachers in India, irrespective of the levels at which they are teaching, is nothing more than reciting poems, reading prose passages and then explaining intricate points of beauty of all such language gems to students who have not been fortunate enough to comprehend what is spoken to them, to express their answers in simple and meaningful language, to read a passage with comprehension and or communicate their thoughts in a simple written form. Even if it sounds as cliché or too simple to be told, the language teacher needs to be told, perhaps again and again, that he must plan his teaching with specific skill objectives in view and that he must not merely teach literary forms but through them develop efficient communication skills which in turn will develop capacities in children to reach higher goals of appreciation and enjoyment of literary forms of high aesthetic value. The point of continuous evaluation with respect to specific learning outcomes is of even a great importance to language education than it is to Science, Arts or Technology. Comprehension of a certain idea or concept related to a specific content ca wait for a delayed evaluation, but it just cannot afford to wait in language education, say with respect to LC. A person will be completely un-functional if he has to be given instructions over and over again for executing a piece of work. Immediate, on-the-spot and objective-based evaluation should be part and parcel of every language lesson given by the teacher. Delayed evaluation of a poems, such as what the central idea of a poem is, particularly of the one taught from a textbook all aspects of which have been thoroughly and eloquently thrashed out by the teacher, is absolutely meaningless and has no place in the modern language teaching. Thus, the findings in general suggest that if instruction and evaluation are planned as suggested in the schema, they may help promote better learning than when they are not.

Hierarchy in Language Skills
In the opinion of the investigators, the support to a minor hypothesis, which is rather a by-product of the project, is perhaps more valuable than the major finding, as it offers a substantial contribution to the theoretical schema. The presentation of the data and the discussion following thereafter have specific reference to an elaborate theoretical discussion on the Bloom model and to hierarchies in language skills and content objectives offered in Chapter 1. Attention the Bloom model of educational objectives with an ad hoc assumption of cumulative hierarchy has been available for the last two decades, the attempts of its empirical validation are not older than five or six years (Pachauri, 1972; Dave, 1972; Dave and Anand, 1972); and (2) This paper having critically examined the conceptual and the theoretical rationable of the Bloom model has made an initial attempt to separate the language and the content aspects of the same, and then to propose first, an existence of a parallel cumulative (or independent ) hierarchy(ies) in language objectives, second, an empirical validation of it (Table 7a and 7b) and third, an explanation of inter-relationship between both (Figure 3 and 4).

This is the proper place to present the data regarding the acceptance of the hypothesis that language skills are hierarchically related. There are two criteria of the concept of hierarchy which needed empirical substantiation, one, difficulty and the other, complexity. Two appropriate statistical tests which could independently show significant differences and yet become complimentary and reinforcing to the concept of hierarchy investigated were necessary. In order to investigate the criterion of difficulty, i.e., E>GC>RC>LNC>LC, the choice fell on the Friedman's test of Analysis of Variance by Ranks, for it not only provided information regarding the significant differences existing among the five skills but also the rank order in which they fell (Walker and Lev, 1965). Although the separate frequency distributions of scores with respect to the four skills showed (except EP) approximations to normality, it was considered fallacious to assume homogeneity of variance among scores, since the inter-correlations (Table 4) demonstrated more divergence and independence than concurrence among them.

For testing complexity, i.e., EP is more complex than GC, and GC is more complex than RC and so on, there was no choice but to select the McQuitty's Hierarchical Syndrome Analysis, as it being the only statistical test available for such investigations.

It is necessary to mention that hierarchy implies cumulativeness which is so well stated by McQuitty (1966). It is stated: "The theory says that every individual represents a succession of types, first an individual type, then types analogues to a species, a genus, a family, etc. As more and more individual types are classified together to represent higher and higher orders of hierarchical type, the successive categories become better representatives of pure types, which exist only in theory".

That Bloom and his associates also started with such an assumption as evidenced by the title "Taxonomy of Educational Objectives", that they gave to their book. Hierarchy in ordinary use also implies the same meaning, e.g., in academic terminology, the post of Professor is higher than that of Reader and thus inclusive of the same; the post of Reader is higher than that of Lecturer and thus inclusive of that. The following graphical presentation clarifies this concept and shows what kind of results are obtained through the use of McQuitty's test.

LC : LC
LNC : LC + LNC
RC : LC + LNC + RC
GC : LC + LNC + RC + GC
EP : LC + LNC + RC + GC + EP

Therefore, the hypothesis regarding hierarchy was split into the two following major assumptions, and were tested by the two statistical tests discussed above.

(1) EP > GC > RC > LNC > LC
(2) The order of hierarchy in language skills is as
LC-LNC-RC-GC-EP

ANALYSIS OF VARIANCE BY RANKS (FRIEDMAN'S TEST)
FOR LANGUAGE SKILLS SHJOWING HIERARCHY
WITH RESPECT TO DIFFICULTY CRITERION

Language Skills

LNC

Chi

Square

ç R

X¨ Ranks

Rank Order

N=730

2486

3.41

1994

2.73

944

1.29

2185

2.99

3341

4.58

1645.6
df=4

P < .001

941.34
df: n₁ = 3
n₂= 2187
P < .001

The results with significant values of Chi Square (1645.6) and F (941.34) at .001 level presented in Table 7a were obtained through the Friedman's test of Analysis of Variance by Ranks. The sums of ranks and the mean-ranks with respect to the five skills (Table 7a) obtained with the help of the Friedman's test strongly substantiated the assumption regarding the existence of a hierarchical structure in skill objectives of language. While doing so, they rejected the rank order which was derived from the conventional assumption regarding the order of development of language skills, i.e., EP>GC>RC>LNC>LC. The data supported a different hierarchical order as follows: EP>LC>GC>LNC>RC. The two sets of five graphs drawn to examine the frequency distributions and cumulative frequencies of the five skills clearly demonstrated a systematic decrease in the mean performance scores of the same large group on the five skills along with the movement of their frequency distributions towards the lower end of the continuum of attainment in the following descending order, 1. LC 2. LNC 3. GC 4. LC and 5. EP.

Having obtained a new hierarchical order for the five language skills with respect to difficulty criterion, their inter-correlations were subjected to the McQuitty's hierarchical syndrome analysis test. The data and matrices (1-4) presented in Table 7b again lend substantial support to the two assumptions of hierarchy, although again the rank order in which they fall is different from the one originally predicated, i.e., LC-LNC-RC-GC-EP and even slightly different from the one found through the Friedman's ANOVA test by ranks.

CORRELATION INDICES (rs) ANALYSED THROUGH THE McQUITTY'S TEST FOR
SHOWING HIERARCHY AMONG THE FIVE LANGUAGE SKILL OBJECTIVES

Matrix 1						Matrix 2
	RC	LNC	GC	LC	EP		RC-LNC		GC	LC	EP
RC	--	.516 1	.472 3	.373	.196	RC-LNC	--		.400 2	.400 1	.234
LNC	.516 1	--	.328	.506 2	.271	GC	.400		--	.319	.206
GC	.472	.328	--	.319	.206	LC	.440 1		.319	--	.242 3
LC	.373	.506	.319	--	.242 4	EP	.234		.206	.242	--
EP	.196	.271	.206	.242	--

Matrix 3						Matrix 4
											Final Hierarchy
		RC-LNC-LC		GC	EP			RC-LNC-LC-GC		EP	RC-LNC-LC
RC-LNC-LC		--		.360 1	.238 2	RC-LNC-LC-GC		--		.222 1	GC-EP
GC		.360 1		--	.206	EP		.222 1
EP		.238		.206	--

It is very important to note that these data and matrices reveal some more relevant and subtle information, as they were analysed through a test developed with a specific Aim of investigating the phenomenon of hierarchical syndromes in a set of obtained data. While being complimentary and reinforcing to the results reported by the previous test, these data and matrices demonstrate how the hierarchy of the five skills is built up step. Matrices 1and2 show that RC-LNC-(LNC-RC) forms the first reciprocal pair of a Typal Representative, thereby showing a mutual association. Matrices 2 and 3 further show that RC-LNC-LC forms a higher Typal Representative, thereby representing A hierarchical order of LC, LNC and RC as ranks 1,2 and 3 respectively. Similarly, Matrices 3 and 4 show higher Typal Representatives as follows: RC-LNC-LC-GC and RC-LNC-LC-GC-EP. The rank order has slightly changed, as LC and GC have acquired the third and fourth ranks. The previous rank order instead showed GC as the third and LC as the fourth rank. Except this reversal, the other ranks have retained the same positions. For the purpose of interpretation, the investigators have preferred the rank order obtained through the Friedman's test. There are two specific reasons for making this choice. One, it provides the level of significance for the differences among the sums of ranks of the five skills and two, the frequency distribution of LC clearly demonstrates its higher order than that of GC, it being farther on the lower continuum of the scale.

It is pertinent to discuss in what way these results are complimentary and reinforcing and yet different from the above ones. They demonstrate that one, the five skills are different and two, they form a definite hierarchical structure. At the same time, the correlations have shown some concurrence among them also. The contribution of this test is to show divergence and concurrence at one and the same time. The data analysed through this test show that RC is an individual type that comprises a set of factors unique to itself. LNC is another individual type with another set of factors. When RC-LNC forms a higher Typical Representative, it means that LNC has all the factors of LC plus some factors unique to itself. Similarly, when a Typal Representative RC-LNC-GC is formed, it means that GC constitutes a set of factors which are inclusive of both the sets of factors comprising RC and LNC plus its own unique set. In the same vein, EP is composed of all the factors constituting RC-LNC-GC-LC Typal Representative plus its own unique set. Thus, each typal representative contains some similar factors to the ones which form the lower hierarchical type (concurrence) as well as a different set of factors unique to itself (divergence).

What are these factors? Are they audile and visual symbols? If they are assumed as such, then how can LC, supposedly being composed of audile symbols only, be the end highest in the hierarchical order? More critically, how can audile and visual symbols be integrated in a single hierarchical structure? The investigators would like to make an intelligent guess. As may be recalled, whether audile or visual, the pre-test with the help of which these data were collected and which subsequently were analysed through a variety of statistical tests comprised a series of questions to which answers were sought from the pupils. So, they should be considered as stimuli that were presented in the forms of questions. Each question-irrespective of the usage of audile or visual symbols-might have aroused certain mental processes in order to elicit a correct response. Perhaps, these factors can be explained in terms of mental processes aroused when a particular skill is used at the time of instruction and evaluation. If this guess or hunch is correct, it would be necessary to analyse these skills into specific objectives or EBOs. Such an attempt is made in the modified version of curriculum schema. It is surprising to note that such obvious trends with respect to content objectives have yet to be reported by any researcher in the West. One of the investigators along with another colleague (Dave and Anand, 1972) has recently published a paper which reports the first set of evidence of the empirical validation of hierarchy of educational objectives in India. These data hint that, RC is the least complex and least difficult of all and comparatively is less complex and less difficult than LNC and, while LNC is less complex and less difficult than GC and LC, EP is the most difficult and complex of all the five skills as it demands proficiency in two languages. Put more simply, it is easier, to develop reading comprehension (RC) than to develop listening and not taking competence (LNC); is easier to develop than to develop the skills required for composition (GC); and while it is quite difficult to develop listening comprehension (LC), the skill of comprehending a passage of another language and then translating it into one's own mother tongue (EP) is the hardest of all the skills studied here.

Theoretical Significance : One may wounder how seemingly simply findings like these can have theoretical significance of any consequence of serious nature. Well, many a time, simple phenomena in science turn out to be the most meaningful, if pursued seriously and systematically. For example, should it have taken so long for psychology and education to realize that learning - in language or content-may be hierarchical in nature and not normative? Do not teachers come across situations every day which convincingly show that when a simple question aimed at testing rote reproduction of given information elicits almost a hundred per cent response, a complex question aimed at testing, say, ability to interpret or to formulate a hypothesis, hardly elicits a response worth the name from the same group of students? Why, then, have the psychometricians constructed tests of all kinds choosing items of so-called average difficulty and discrimination values? Why can't there be tests which are based on assumptions that are different from those implied in the normal probability distribution? As a matter of fact, it is quite logical to argue that the characteristics such as language skills should follow hierarchical development and tests of measurement should reflect the reality which has been now supported by empirical evidence.

Having discussed the evidence of the hierarchical skill levels in language, a more rigorous claim of validity and efficacy for the Bridge Course is warranted. The Experimental Group had shown substantial gains not only in RC but also in all the other skills (see Table 5), indicating thereby its capacity to develop higher order skills. This doubtless gives more credence to the effectiveness of the Bridge Course.

Now the time is most opportune to discuss the finer points of theoretical significance, since it is empirically demonstrated that the Input (Instruction), the Process (Modes, Media and Methods) and the Output (Evaluation) are an integrated whole, that one part does not exist without the other and, that there seems to take place a dynamic interaction among the three when any kind of learning is attempted. Furthermore, the above explanation suggests that the more a teacher consciously tries to increase this dynamic interaction among the three, the better seem to be the chances for promoting learning, particularly higher levels of learning.

In the light of the above findings that there exists two separate and independent hierarchies of objectives for language and content, that the former precedes (or must precede) the latter and that the gains in language attainment may help increase achievement in content, it will be quite profitable to re-examine the propounded schema from all possible different angles. It must be readily pointed out at this juncture that the subtle contribution to the theory lies in the evidence that these independent hierarchies seem to be interlinked. It is very imperative to recall the relevant argument from the Introduction, that

"…While a direct link between KNOWLEDGE and
CONTENT is possible, the links of other levels with
Content can be attained only through the medium of
Language."

This medium of language seems to hold the key to the understanding of learning. Let this be expanded as follows, gradually by modifying the original curriculum schema. Now the Input consists of two types of objectives and behavioural outcomes.

The Input (Instructional Objectives)
|
________________________________ |________________________________

Language	Content	Levels
	Knowledge	1
RC	Comprehension	2
LNC	Application	3
GC	Analysis	4
LC	Synthesis	5
EP	Evaluation	6

Suppose a teacher is teaching a major concept such as conduction in physics. Ordinarily in an Indian classroom most of the teaching is done through Oral Expression (OE) and learning takes place through LC. Although it is difficult to conceive a situation where instruction is imparted through only one skill, for a while let it be assumed that in a controlled experiment*

learning takes place either through RC, LNC, LC or EP. Assuming that the whole group of students was more or less equal in all these skills, the data reported indicate that the probability of producing greater learning will decrease from RC to EP in that order, they being hierarchical. That apart, the probability of developing higher levels of learning will also be decreasing. In other words, the chances of producing higher levels of learning are quite high through RC, whereas the same through EP are comparatively very low. This very strongly indicates that learning through LC is also quite limited in quality and quantity.

But how valid is the assumption that a normal random sample of children at any stage is equal in language skills? Obviously, they are not and they can never be. Then, it is certain that a particular skill through which learning is attempted will from one kind of LAYER which will interfere in the process of learning. At once, it is realized that the medium of language, loosely labeled as a communicative skill, surrounds INSTRUCTION with not one or two but with a number of layers which may interfere in the process of learning. Let the above schema be now further modified with respect to LC only.

It can be well imagined how many LAYERS will be interfering when an ordinary teacher makes use of several language skills at a time in teaching a certain portion of subject matter.

The reader needs to be reminded that the argument mentioned a while ago is more appropriate to the Output than to the Input as the original Bloom schema was essentially an evaluative one. The inter-links are more easily perceivable in the Output than in the Input. Suppose the pupil has developed a certain expected level of learning. What about LAYERS in evaluation? As was the case in instruction, layers of language again interfere in measuring and evaluating the Real Learning Outcome (RLO). Assume that the pupils have retained learnings equally well. The schematic presentation below depicts what will happen and how the layers well. The schematic presentation below depicts what will happen and how the layers of language will interfere in evaluation

It is clear from the schematic presentation above that the two layers again interfere in measuring the real outcome. That apart, it is fair to assume that as the complexity of language increases, the probability of correct responses decreases. This phenomenon has been investigated by Dave and Anand (1972) and they have reported that the pupils studying through the English medium seemed to have experienced interference in solving items of Understanding

(comprehension) and Application levels, when their counterparts studying through the medium of the mother tongue had done significantly better on the same in the tests in mathematics, general science and social studies. This again shows how learning through the mother tongue, a situation similar to EP learning, is more difficult than learning through the mother-tongue, and how high levels of learning are obstructed when the ability to use language is severely limited. It shows how the layers of language skills interfere in content attainment. It must also be pointed out that evaluation through RC, LNC, LC and EP again will have a decreasing probability of bringing out correct responses in terms of quantity and quality, as it follows the principle of hierarchy so evident with respect to Instruction. This further suggests that if instruction is imparted through RC when learning points are always present and which may be referred back and forth any time during the process of learning and if evaluation is also done in such a way when the elements necessary for arriving at the correct response are present right there for reference, the development and measurement of learning outcomes are more facilitated.

The foregone discussion and the schematic presentations of certain points of theoretical significance clearly indicate the need for modifying the learning schema presented earlier. It is very clear that teaching-learning process is bound by LAYERS of language skills which require permeation for achieving Expected Behavioural Outcomes (EBO). As a matter of fact, human knowledge related to any subject involves manipulations of either audile or visual verbal symbols. Similarly, it is only through these verbal symbols can one evaluate how much knowledge is acquired by an individual. Thus, it is fallacious to think of a curriculum schema exclusive of the medium of language. Hence the Modified Version of the curriculum Schema (MVCS). This new schema (Figure 5) expands the Integrated Model of Learning proposed by Dave (1972) incorporating the essential components of language.

Language Symbols : It is many times heard that a language of an individual must be cultivated. What is meant by this? The investigators would like to deal with this question at length, lest the whole argument may look too simple to have any theoretical relevance. Usual reactions to a presentation like this are "old wine in a new bottle!" or "we have been telling the same thing for years!" and so on.
The investigators hold a strong view that any learning situation should be aimed at developing a certain mental ability or a set of mental abilities which can be measured and evaluated in terms of specific behavioural outcomes. Cultivation of language most of the times means a quick comprehension of a communication and a facile or an eloquent expression. Although these language outcomes are very important and are evidence of good language development, in themselves they cannot be taken as evidence of good mental development. Many a time it is experienced that eloquence suffers from superfluity and the words used are just hollow words. In other words, eloquence itself does not guarantee a higher decree of depth of knowledge in a subject, although eloquence combined with depth in a subject matter will always be superior to the former. The authors would like to suggest a shift in the focus of developmental language goals. Let not cultivating language skills, i.e., LC, RC, OE, WE and combination of these, o developing competence in physics, chemistry, history, etc., be the goal of education. Instead, let that goal be of developing basic mental abilities through language and content. Then, language development should also mean developing hierarchical objectives and EBOs, comprehension being just one of them. The findings of the developmental-cum-experimental project has demonstrated this, although in a very limited sense. A variety of passages belonging to different subject mater were used in the Bridge Course. The questions on linguistics were minimal; but as was explained in methodology, the questions on content required ability to reproduce as well as analyse, synthesize and evaluate. It is readily conceded that there is no data to support the above claim, even partially, as no classification in terms of EBOs was done and tested by each question; still the fact remains that many questions required abilities to answer far beyond those required in mere comprehension of a passage, either listened or read. The Modified Version of Curriculum Schema lends an excellent scope for embarking upon research for the empirical validation of many underlying assumptions such as this. The authors would like to interpret the data by saying that since the attainment in content depends upon the development of certain mental capacities, it is quite probable that the training through the Bridge Course might have helped improve these abilities which in turn might have facilitated the attainment in both language and content. True, this is a little stretched interpretation; but the investigators would like to proceed on this assumption rather than on any other for the sake of maintaining consistency in the logic of developing the MVCS.

This interpretative argument needs to be pursued. It boils down to asking a question, whether it is possible to develop a certain set of EBOs varying from a simple verbatim reproduction of a fact or an idea or a concept to a production of a new or unique idea of his own by an individual learner even if the language material used is only a simple story Herein a strict technical sense a question of a particular subject or content does not arise at all. The authors positively believe that it is quite possible to TEACH a simple story to young children with reference to a hierarchical set of specific objectives and EBOs; comprehension being only one of them. This is the shift in focus on developmental goals which the authors would like to emphasise. As can be seen, the emphasis on the development of certain skills is secondary as it is assumed that the process of developing a certain EBO cannot but assure the development of means (skills) through which they have to be tested. For example, a child who is required to develop a unique production of a new idea on his own will have to be apt in comprehending and using symbols-audile or visual-to produce one. What is most pertinent is that this type of teaching provides a unique opportunity which helps a child to acquire mental abilities that can help him to reach the depth of content through quick analysis of communication provided in the new forms of language.

In order to be more concrete and precise, an attempt is made to draw a set of EBOs under four different objectives, i.e., KNOWLEDGE, UNDERSTANDING, APPLICATION and CREATIVITY, the continuum being production and familiarity. It is necessary to draw attention of the reader to the fact that while the authors have retained the first three categories of the Bloom system of objectives in tact, they have combined the other three, i.e., Analysis, Synthesis and Evaluation (Judging) under a new category labeled as Creativity. According to them, these three can be equated to the subcategories called outcomes under the first three (see Figure 5).

10. What disturbed him so much? (Recall and Recognise)
11. What do you think of his attempts to (Judge)
get the caps back?

12. If you consider his actions as correct, (Establish Relationship)
give some reasons.

13. If you consider his actions as incorrect, (Discriminate)identify his biggest mistake.

19. Which action of the monkeys might have produced the new thought in his mind? (Establish Relationship)

21. Why did not the pedlar throw his cap on the ground as soon as the new thought came to (Formulate and Establish
his mind? Hypothesis)

22. What is the connection between the mistakes committed by the pedlar and the throwing
away of the caps on the ground by the monkeys? (See Relationship)

24. There is a particular relationship between his anxiety when he lost the caps and his happiness
when he got them back from the monkeys. Identify that relationship. (Establishing Relationhip)

25. What kind of precautions would the pedlar take if he happened to rest under the same tree some
time in future? (Infer)

26. Suppose you were in the place of the pedlar,how would you have got the caps back from
the monkeys? Think it out your way, other (Analyze and Synthesize) than the one given in the story.

A few important points need to be added. This passage has been given as an illustration, and a caution has to be given that, as it is, in English it will certainly appear to be difficult for Indian children. But its representation in the mother-tongue even for children of age 5 or 6 will not be too difficult. As a matter of fact, the investigators have tried to translate this popular vernacular story into English with some modifications.

It is interesting to note how a variety of ways can be used for teaching this story and evaluating the EBOs.

LC. This can be taught to very young children who have not learned to read. First the complete story can be read to them. Then having read only one paragraph at a time, the questions pertaining to it may be put to them. The answer may be of supply or selection type, i.e., they may be asked to give an answers to an open-end question or a question and 4 answers to it may be read to them one by one and, subsequently, they may be asked to select the correct answer. It must be noted that the underlying process for the former test will be recall, whereas for the latter it will be recognition.

It will be interesting to undertake an action experiment with an individual or a group of individuals using different techniques of instruction and evaluation in order to see hierarchical nature of language skills.

First, as is described above, the child may be taught the passage through listening and at least 4 types of tests administered in which different layers will be interfering at the time of evaluation. The following schematic presentation depicts the situation clearly.

Second, the passage may be taught through listening and note taking, i.e., the students are instructed to take their own notes, when the passage is narrated to them. The same 4 tests mentioned in the above schema may be given to them. The difference between these two presentations is that in this situation the students are allowed to use their own notes in answering the questions.

The Third way of teaching this story is to ask the students to read the given story within a specific time. The written tests (3 and 4) may then be given to them. One can vary the situation by taking away the passage after a certain number of readings or letting them have the passage for ready reference to answer the questions.

Four, this story in English may be presented to the students in the all the three ways mentioned above, i.e., through listening, listening and note taking and reading. Afterwards, they may be administered the same tests in the mother-tongue. This will test their ability to epitomize a passage learned through the other tongue (EP). It is expected that scores on tests will show trends of hierarchy in language skills, and that an increase in hierarchical levels will be accompanied by a decrease in mean achievement by the group, in spite of the learning that might have taken place in the previous presentations. Furthermore, the recognition test will show better achievement than the recall one.

Since this is a pioneer that effort, no explanation can be perfect and fully convincing. All that one can say is that it is open for empirical validation and consequently it should be put to rigorous experimentation. Just the same, the authors would like to assert again that the teaching of different forms of literature such as essay, story, poem, play, etc., provide excellent opportunities to develop basic mental abilities varying from a simple verbatim reproduction to a unique unfamiliar production, and a teacher should be fully aware of the potential of such language materials. Furthermore, what needs to be stated in unequal terms and also without much reservation is that no EBOs (except recognition and recall) related to content are possible without the development of EBOs pertaining to language, and that the latter in forms of permeable layers precede the former.

Although many factors and conditions have been considered as filters, layers and blocks interfering in the process of learning, in the ultimate analysis language seems to be the only true layer of interference. What then is the part played by materials, methods and mode-media in learning? This is an important question to be answered before the MVCS is presented. They all are in true sense AIDS. They may help promote better learning. They do so by thinning the layers of language and thus make the permeation of layers easier. Apart from that, their role is not of much significance.

The schema (MVCS) presented in Figure 5, can now be easily explained, as most of the points and the assumptions implied in them have already been discussed. The Input contains two sets of Objectives and Expected Behavioural Outcomes, one under language and the other under content. Those related to language have been listed first to indicate that they precede those related to content. A few interactions between them. The arrows between language EBOs and language skills represent teaching-learning of any literary form. Sandwiched are 4 major skills, Reading, Writing, Listening and Speaking. The other skills studied in this paper have also been listed for reference, as they have been empirically supported. The rank order of language skills supported by the evidence in the project has been followed. Words 'comprehension' and 'expression' have been dropped as they from one of the categories of major objectives. Besides, they are redundant anyway. Further, the reason for putting skill objectives after language EBOs is to indicate their secondary importance to the latter and also to draw attention to the fact that the development of language skills depends upon the development of EBOs. That is to say, the greater the attempts for reaching a higher level EBO, the better are the chances for sharpening these skills. In simple words, an attempt to develop creativity in children cannot be attained unless and until they attain certain levels of listening, reading, speaking and writing. At the same time, not aiming consciously at the development of creativity may deprive children of a higher development in language skills. In addition to representing one, two, three or multi-way interactions, arrows also indicate the presence of layers of language skills.

Under the Process are listed the major components along with a few schematic diagrams to represent interactions and their directions among them. Since it is through teaching-learning process a certain EBO or a set of them is developed, the LAYERS of language skills and the problems of their permeation are also drawn I the schema. The following are the implied assumption:

1. No learning Experience, no Learning Outcome, i.e., without experience, no change in behaviour.

4. Interactions among Materials, Methods, Mode-Media aid in thinning the layers of language skills that surround stimuli, thereby intensifying learning experiences of an individual.

The output consists of only Real Learning Outcomes (RLO) and related language skills, again presented in a hierarchical order. Purposely the language EBOs and objectives of language and content have been omitted. In Evaluation, one is mainly concerned with those outcomes which are developed and also amenable to measurement. Since they can never be equal to EBOs, they are labeled as RLOs, i.e., what is really measured in quantitative terms. The process of measurement is also obstructed due to the layers of language skills. If a teacher decides to measure RLOs through listening, then his ability to present (write or speak) and present a test item in verbal symbols comprises one layer. Similarly, the student's ability to comprehend the spoken item (listen) and given an answer to it orally (speak) comprises two other layers. The same problems of permeations of 3 layers are present in a written test too.

Thus, this schema encompasses all the variables related to the Input, the Process and the Output and the interactions among them present in an ordinary classroom. Although some empirical evidence is gradually tricking in (Dave and Anand, 1973), it is open to experimentation and subsequent modification. The authors modestly hope that this will give an impetus to research in language and content curricula.