Journal of Psycholinguistic Research, Vol. 12, No. 2, 1983
Bilingualism and the Perception of Professional Concepts Janice Monti-Belkaoui I and Ahmed Belkaoui 2
Accepted August 23, 1982
A selected set of professional concepts was subjected to analysis through two separate multidimensional scaling techniques, the INDSCAL and TORSCA models, to evaluate the intergroup perceptual differences of four experimental groups, made up of unilingual French, unilingual English and bilingual students. The linguistic relativism thesis provided the research hypotheses on the relationship between language access and usage and concept perception. The multidimensional scaling techniques were applied to the matrix of subjects' similarity judgments on pairs of concepts, thus enabling the identification of three dimensions. The dimensions were labelled as conjunctive, relational and disjunctive, and were assumed to be related to the criteria employed by the subjects in their similarity rankings, An analysis of variance of the individual saliences on each dimension provided evidence of linguistic relativism for both the relational and disjunctive dimensions, these findings support the contention that unilingual speakers of separate languages differ from each other and from bilingual speakers in their perception of professional concepts.
INTRODUCTION Because language mediates our world view, it plays a central role in the development of cognition and perception. Individuals, as they learn a language, acquire not only a store of lexical and grammatical characteristics but also a linguistic mode of cognition and perception:
I Departrnent of Sociology, Rosary College, 7900 W. Division, River Forest, Illinois 60305. 2Department of Accounting, University of Illinois at Chicago, Box 4348, Chicago, Illinois 60680. 111
0090-6905/83]0300-011150.300/0 9
Plenum Publishing Corporation
112
Monti-Belkaoui
Language is not a substitute for direct experience; it is itself a mode of experiencing and an activity of apprehending and transforming direct experience as well as symbolically mediated experience (Schrnidt, 1973; p. 119).
Speakers of different languages therefore acquire and maintain separate world views. An extension of this line of reasoning suggests that bilingual individuals possess a different world view from unilinguals (Lambert & Tucker, 1973). It is not surprising then to note that professional organizations and corporations who operate internationally are paying increasing attention to language and linguistic differences. One outcome of this awareness is the growing demand for the international standardization of professional concepts and techniques. This paper investigates the hypothesis that unilingual speakers of two different languages and bilingual speakers of the same languages hold different perceptions of professional concepts. The study is also an attempt to validate one of the levels of the "Sapir-Whorf hypothesis" of linguistic relativism.
S Y S T E M A T I Z A T I O N OF THE " S A P I R - W H O R F HYPOTHESIS"
Anthropologists have always emphasized the role of language in their studies of culture. Sapir's investigations of the linguistic symbolism of a given culture view language both as instrument and as communication of thought (Sapir in Mandelbaum, 1956). A given language predisposes its users to a distinct belief. The idea that language is an active determinant of thought forms the basis of the principle of linguistic relativism. According to the Whorfian version of the principle, ways of speaking are reflections of the metaphysics of a culture (Whorf in Caroll, 1956). These metaphysics constitute the unstated premises which shape the perception and thought of those who participate in that culture and predispose them to a given method of perception (Belkaoui, 1978). The work of Fishman (1960) is an attempt to systematize this set of assumptions, known as the "Sapir-Whorf hypothesis." Fishman's fourfold analytical scheme (Table I) distinguishes between two levels of language, lexical and grammatical, and two types of behavior, linguistic and nonlinguistic: (i) The lexical level refers to all words which compose a language. Languages differ in the number of terms they possess to describe phenomena. (ii) The grammatical level refers to the manner in which the structural units of a language are organized.
Perception of Professional Concepts
113
Table I. Fishman's Schematic Systematization of the Sapir-Whorf Hypothesisa i
Data of (cognitive) behavior Date of language characteristics
Language data ("cultural themes")
Nolinguistic data
Lexica/or "semantic" characteristics
Level 1
Level 2
Grammatical characteristics
Level 3
Level 4 ii
m
asee Fishman (1960).
(iii) Linguistic behavior refers to choices among words. (iv) Nonlinguistic behavior refers to choices among objects. This last distinction will be clarified in the following explanation of the cells in Table I. (a) Cell 1 posits a relationship between the lexical properties of a language and the speaker's linguistic behavior. Linguistic behavior, the choice of words for describing a particular phenomenon, differs from one language to another. (b) Cell 2 posits a relationship between the lexical properties of a language and the nonlinguistic behavior of the users of that language. This refers to the idea that speakers of a language that makes certain lexical distinctions will be able to perform tasks better and more rapidly than speakers of languages that do not make such distinctions (Lenneberg, 1973; Brown & Lenneberg, 1954; Lantz, 1953). (c) Cell 3 posits a relationship between grammatical characteristics and linguistic behavior. This refers to the idea that speakers of a language that uses specific grammatical rules acquire a distinctive world view from speakers of other languages that do not employ such pales (Hoijer, 1951; Ervin-Tripp, 1969). (d) Cell 4 posits a relationship between grammatical characteristics and non-linguistic behavior. This refers to the idea that speakers of a language that has certain grammatical characteristics will perform nonlinguistic tasks differently from speakers of languages that do not have these characteristics (Carol & Casagrande, 1958; Belkaoui, 1980). This paper reports a study of the perception of professional concepts by unilingual and bilingual speakers as measured by a task of linguistic behavior. The research is designed to answer the following question: Do access and use of two distinct language systems result in differences in the
114
Monti-Belkaoui
perception of professional concepts? Accordingly, the specific research hypotheses derive from Cell 1 and deal with the relationship between lexical properties of a language and the linguistic behavior of its speakers.
RESEARCH HYPOTHESIS
The research design is based on a perceptual task assigned to three groups of university students, two of which were composed of unilingual and one of bilingual speakers. The groups were selected and matched for their exposure to and understanding of a set of professional techniques and concepts. The major research question is to determine whether language access and usage have any appreciable effect on perception. The first of the hypotheses represents a direct test of the "SapirWhorf" proposition as outlined by Cell 1. It tests whether linguistic behavior, in the form of the choice of words for a particular phenomenon, will differ from one language to another. Specifically, in this case, the hypothesis tests for differences in the perception of professional concepts by unilingual subjects from two different language groups. This is stated in null form as follows: HI: The perception of professional concepts by unilingual speakers from two different language groups, as measured by the individual weights assigned to the dimensions of the cormnon perceptual space, will not differ.
The second hypothesis is an extension of the basic relationship expressed by Cell 1. This formulation investigates differences in perception of professional concepts between bilingual and unilingual speakers. It questions whether the two language systems available to bilingual speakers provide cognitive enrichment or linguistic and perceptual confusion. In null form: H2: The perception of professional concepts by bilingual and unilingual speakers, as measured by the individual weights assigned to the dimensions of the common perceptual space, will not differ.
The third hypothesis also deals with bilinguals. It tests whether switching from one language to another leads to better perception. Although not strictly derived from Cell 1, this hypothesis is supported by the literature on bilingualism. Language switching has been found to be related to higher levels of creativity (Peal & Lambert, 1962; Landry, 1974), cognitive flexibility (Balkan, 1970), concept formation (Liedke & Nelson, 1968), verbal intelligence (Lambert & Tucker, 1973; Kittell, 1963), and psycholinguistic abilities (Casserly & Edwards, 1973). The null form of this hypothesis is:
Perception of Professional Concepts
115
H3: The perception of professional concepts by bilinguals using each of two languages, as measured by the individual weights assigned to the dimensions of the common perceptual space, will not differ from one language to the other.
METHODOLOGY
Subjects Experimental subjects were recruited from two sections of an accounting theory course given at the University of Ottawa, Ontario, Canada, in the spring semester of 1981. All subjects were fourth-year accounting majors who had fulfilled the same course and degree requirements in preparation for the bachelor degree as well as for the professional degree of C.A., the Chartered Accountant certification. One section of the course was taught in English and the other in French by the same instructor, a bilingual speaker and one of the paper's coauthors. English- and French-language editions of the same textbook, written by the instructor, were used in each respective language section (Belkaoui, 1981 a,b). Testing and evaluation performed at the beginning of the semester revealed that students in both sections were equally unfamiliar with certain advanced accounting concepts and techniques. (The professional training of the students prior to the accounting theory course had been restricted to lower-level courses in conventional bookkeeping). The research design insured that each student was introduced to the same set of concepts, in his or her own language, in the same time period by the same instructor and textbook. Other things being equal, the only difference in the acquisition of the professional concepts was the language of instruction and learning. Because of this, any differences in perception of the concepts may be attributed in great part to the impact of the language used. Before the conduct of the experiment, some of the students enrolled in the English section were found to be bilingual speakers. Specifically, these students were bilingual Francophones who had elected to take the course in English (primarily because of considerations related to future job-market opportunities in the predominantly English province of Ontario). Because of the presence of these bilingual students, four experimental groups were designated: (i) Twenty-seven Francophone students enrolled in the French section who performed the experimental task in French formed the unilingual French group (UF).
116
Monti-Belkaoui
(ii) Twenty-two Anglophone students enrolled in the English section who performed the experimental task in English formed the unilingual English group (UE). (iii) Seventeen bilingual students enrolled in the English section who performed the experimental task in English formed the bilingual English group (BE). (iv) A week later, the same seventeen bilingual students in the English section performed the experimental task in French, thus forming the bilingual French group (BF). Research Design Subjects in all four experimental groups were given the same questionnaire in an English- or French-language version. The questionnaire required subjects to assign similarity judgments to paired sets of twelve concepts. In multidimensional scaling techniques, such similarity judgments are interpreted as "psychological distances," representing a "mental map" through which respondents view pairs of concepts that are " n e a r " each other as similar and pairs of concepts that are "far apart" as dissimilar. If numerical measures are provided for the similarity judgments, multidimensional scaling techniques may be used to construct a "physical" multidimensional map whose interpoint distances closely relate to the input data. Two separate multidimensional scaling techniques, the TORSCA and INDSCAL models, were applied to individual similarity judgments to estimate the dimensions of the common perceptual space and each respondent's saliences. Regression was then used to measure the relationship between saliences and selected background variables which included the subjects' age, the number of accounting courses taken, and the grade received in the accounting theory course. One of the multidimensional scaling techniques used in the study is the TORSCA nonmetric scaling routine (Young, 1968). Given n(n-1) 2 similarity/dissimilarity measures, the TORSCA program first yields a set of orthogonal coordinates for the final configuration and then estimates the dimensionality of the data. The other algorithm used in this study is the INDSCAL Model (Caroll & Chang, 1970). In contrast to the TORSCA solution, the stimulus configuration obtained from the INDSCAL algorithm is uniquely oriented. The INDSCAL model assumes that all individuals share a common perceptual space, but assigns differential weights or saliences to the different dimensions of the group stimulus
Perception of Professional Concepts
117
space. These individual saliences provide an ideal operational measure for the evaluation of possible inter- and intra-group perceptual differences.
Professional Concepts and Experimental Decisions The twelve concepts used in the experiment were chosen to reflect two categories of professional concepts of relevance to accounting theory construction. The terms "going concern," "entity," "stable monetary unit," and "periodicity," represent underlying assumptions of accounting theory while the terms "cost," "revenue," "matching," "objectivity," "consistency," "full disclosure," "materiality," and "conservatism" represent generally accepted accounting principles within the profession (Belkaoui, 1980). At the first class meeting of each language session, the students were required to assign familiarity ratings to the twelve concepts as well as to other unrelated issues and hypotheses in the field of accounting. The seven-point familiarity rating scale ranged from "not familiar" to "extremely familiar." The results revealed that the students had an overwhelming unfamiliarity with the twelve concepts at the beginning of the semester. Two weeks before the end of the semester, the UF experimental group was asked to complete a French language version of the questionnaire containing the experimental task while the UE and BE groups were asked to complete the English version of the questionnaire. One week later, the bilingual subjects, as the BF group, were again asked to complete the questionnaire, this time in a French-language version. The content of both versions of the questionnaire was identical. Subjects were asked to: (a) provide linguistic 3 and background information, (b) assign to each of the paired concepts an integer rating, on a seven-point scale ranging from "very dissimilar" to " v e r y similar," and (c) list the criteria they used for assigning similarity ratings.
ANALYSIS AND RESULTS
Preliminary Findings The input to TORSCA was a single rank-order similarity matrix computed by averaging the cell ranks obtained from all subjects. The 3Linguistic informationwas required to insure against the presence of any undetected bilingual speakers in the two unilingualgroups.
118
Monti-Beikaoui
resulting average stress indices were .325, .625, and .062 for dimensions 2, 3, and 4, respectively. Based on these results, "goodness of fit" was provided by three dimensions (stress < .05). The input to INDSCAL was a 12 x 12 matrix of similarity judgments for all subjects. "Variance accounted for," a measure of "goodness of fit" in the INDSCAL model was 57%, 65%, and 63%. Goodness of fit improved for three dimensions but tapered off for the fourth. Dimension 4 explained no more variance than could be accounted for in random data. On the basis of both the " s t r e s s " and "variance accounted for" measures, the three dimensional solution has been employed. The graphical portrayal of this solution appears in Fig. 1.
Identifying the Dimensions In the last part of the experimental task, subjects were asked to list, in order of their importance, the criteria used in assigning similarity judgments. Mirroring the results of another experiment where the same treatment was used, (see Belkaoui, 1980) the similarity judgments appear to be assigned on the basis of the presence or absence of perceptual qualities common to each pair of concepts. This process resulted in the three-fold classification of concepts as either conjunctive, relational, or disjunctive (see McDavid & Harari, 1974; pp. 78-79; Hunt & Hovland, 1960; Belkaoui, 1980). Conjunctive concepts are defined as those which have in common one or more perceptual characteristics while disjunctive concepts differ in one or more characteristics. Relational concepts refer to pairs of terms which are linked by a fixed relationship. This classification scheme provides labels for the three dimensions obtained in the I N D S C A L solution (see Table II and Fig. 1). The stimulus configuration depicted in Fig. 1 places the concepts "objectivity," "materiality," and "conservatism" with nearly equal magnitude on the same side of dimension 1. The pairs "period" and "conservatism," "consistency" and "disclosure," "entity" and "stable monetary unit," and "consistency" and "going concern" are also of fairly equal magnitude and placed on opposite sides of dimension 2. " E n t i t y , " "going concern," "stable monetary unit," "period," and "materiality" are found on the positive side of dimension 3 while the remaining concepts are on its negative side.
Perception of Professional Concepts
119
11
y II
I [I I11 I
Conjunctive Dimension Relational Dimension Disjunctive Dimension Entity Assumption
2 Going Concern Assumptien 3 Stable Monetary Unit Assumption 4 Period Assumption 5 Cost PrincipTe 6 Revenue Principle 7 Matching Principle 8 0 b j e c t i v i L y Principle 9 Consistency Principle 10 Full Disclosure Principle 1 1 M a t e r i a i i t y Principle 12 Conservatism Principle
Fig. 1. Stimulus configuration.
120
Monti-Belkaoui
Table II. Saliences for Professional Concepts on a Three Dimension Solution Illl
Saliences Professional concepts 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.
Entity Going concern Stable monetary unit Period concept Cost principle Revenue principle Matching principle Objectivity Consistency Full disclosure Materiality Conservatism
Conjunctive dimension (1) 0.68993 0.30491 -0.34511 0.01261 -0.29556 -0.27769 -0.17216 -0.10981 0.15559 0.84349 -0.18402 -0.08213
Disjunctive dimension (2) 0.10869 -0.15673 -0.11662 -0.43953 -0.08966 -0.23575 -0.35309 0.27462 -0.15810 0.17036 0.53271 0.40310
Relational dimension (3) 0.34030 0.45372 0.73046 0.18687 -0.01793 -0.00786 -0.08010 -0.33560 -0.44650 -0.50882 0.13003 0.04884 in
Intergroup Differences in Perception The I N D S C A L model produces a spatial representation of the similarity data obtained from subjects as well as a set of dimensional saliences. These saliences are unique to each individual and may be interpreted as indicators of the way in which the individual "distorts" the group stimulus space by " s t r e t c h i n g " (represented by the large weights) or " s h r i n k i n g " (the small weights) the distances between points along the weighted dimensions. The investigation of intergroup differences in p e r c e p t i o n was accomplished by a one-way analysis of variance of four e x p e r i m e n t a l groups. This determined the extent to which the groups' saliences differed on each of the three dimensions. The results of the analysis o f variance are presented in Table III. The general hypothesis of no intergroup differences was accepted for the conjunctive dimension but rejected for the disjunctive and relational dimensions. Each of the four e x p e r i m e n t a l groups assigned different weights to the disjunctive concepts, those which differ on the basis of one or more characteristics, and to the relational concepts, which are linked by fixed relationships. C o n s e n s u s among all the groups appears to exist only for the conjunctive dimension, for those concepts which share characteristics.
Perception of Professional Concepts
121
L) c~
> @
0
<
~r3 cq
o
u
~c
r--
o~ ce
II
II
122
Monti-Belkaoui
E
O~ U'-
g.
II
o~
~q
II ~L I I I t
II II
Perception of Professional Concepts
123
The differences along the disjunctive and relational dimensions were further explored in a t-test matrix for group means (Table IV). These findings provide a direct test of the three research hypotheses: (a) The significant differences in perception between the unilingual French group (UF) and the uniiingual English group (UE) along both the disjunctive and relational dimensions provide evidence of differential perception of professional concepts by unilingual speakers from two separate language systems and, thus, makes possible the rejection of HI. These findings also constitute a direct corroboration of the thesis of linguistic relativism by giving support to its main contention that language behavior, in this case, the perception of word choices for describing a particular phenomenon, i.e., professional concepts, differs from one language to another. (b) The significant differences between the bilingual English group (BE) and the unilingual French group (UF), along the disjunctive and relational dimensions, and between the bilingual French group (BF) and the unilingual English group (UE), along the relational dimension, provide support for differential perception of concepts between bilingual and unilingual speakers and for the rejection of H 2. These findings suggest that bilingual speakers who have access to two separate language systems perceive concepts differently than unilinguals. Of particular relevance to the linguistic relativism thesis are the findings that the bilingual speakers (who were all exposed to the professional concepts in the English language section of the course) displayed, in the BE experimental treatment, differences along two dimensions from the unilingual French group (who were trained separately in the French section of the course). However, in the BF experimental treatment, these same bilingual speakers differed only along one dimension from their English section classmates, the unilingual English group. (c) The significant differences in perception between the bilingual English (BE) and bilingual French groups (BF) provide evidence that the habit of switching from one language to another may result in differential perception of concepts which are linked by a fixed relationship. The perceptual differences are not significant, however, for the conjunctive and disjunctive dimensions. These findings provide some support for the contention that language switching may enhance understanding, thus enabling a partial acceptance of the null H3.
124
Monti-Beikaoui
lntragroup Differences in Perception The saliences for each of the three dimensions were regressed with the background variables, i.e., age, number of accounting courses taken, and the grade received in the accounting theory course, in order to determine if the differences would hold after allowing for variation in the maturity and preparation of the subjects (see Table V). These variables were shown to have no effect on the saliences. The reported differences in the perception of concepts appears to be directly related to language and independent of the effects of these background variables.
DISCUSSION This study provides some tentative support for the contention that the perception of professional concepts is shaped and mediated by the access to and acquisition of a language. Beyond the theoretical issues raised by the operationalization of the Sapir-Whorf hypothesis, these results also have some practical implications. One obvious implication is in the area of international economic relations and bilingual education. The increasing volume of international trade during recent years and the decline in the competitive position previously enjoyed by U.S. firms in the international market is seen in part as a function of the linguistic barrier faced by typically unilingual American corporate executives and professionals in dealing with overseas customers and colleagues. Language differences result not only in general communication problems across national boundaries but also in specific perceptual differences in understanding the same concepts used within disciplines, industries, and professions. As a contribution to the literature on the relationship between bilingualism and enhanced perception, our findings may have relevance to those who are concerned with training personnel for careers in inter-ethnic or international contexts. The evidence suggests that fluency in two languages aids in the uniform acquisition and comprehension of professional concepts. The need for (and efficacy of)~the bilingual professional is an area that requires further attention from social sciences, economic planners, and the standard-setters of professional organizations.
Perception of Professional Concepts
125
~D [...
0 m
~J
0
0
126
Monti-Belkaoui
REFERENCES Balkan, L. Les effects du bilinguisme Francais-Anglais sur les aptitudes intellectuelles. Bruxelles: Aimav, 1970. Belkaoui, A. Linguistic relativity in accounting. Accounting, Organizations and Society, 1978, 3(2), 97-104. Belkaoui, A. The impact of socioeconomic accounting statements on the investment decision: An empirical study. Accounting, Organizations and Society, 5(3), 263-283. Belkaoui, A. Accounting theory. San Diego: Harcourt, Brace and Jovanovich, 1981(a). Belkaoui, A. Theorie comptable. Quebec City: Presses de l'Universite du Quebec, 1981(b). Brown, R. W., & Lenneberg, E. H. A study in language and cognition. Journal of Abnormal and Social Psychology, 1954, 454-462. Carol, J. B. & Casagrande, J. B. The functions of language classification in behavior. In E. E. Maccoby, T. M. Newcomb, & E. L. Hartley (Eds.), Readings in social psychology (3rd. ed.). New York: Holt, Rinehart and Winston, 1958. Caroll, J. D., & Chang, J. J. Analysis of individual differences in multidimensional scaling via N-way generalization of "Eckart-Young" Decomposition. Psychometrika, 1970, 35, 238-319. Casserly, M. C., & Edwards, A. P. Detrimental effects of grade one bilingualism programs: An exploratory study. Paper Presented to the Annual Conference of the Canadian Psychological Association, 1973. Erwin-Tripp, S. Sociolinguistics. In L. Berkowitz (Ed.), Advances in experimental social psychology. New York: Academic Press, 1969, pp. 91-165. Fishman, J. A. A systematization of the Whorfian Hypothesis. Behavioral Science, 1960, 323-335. Hoijer, H. Cultural implications of the Navaho linguistic categories. Language, 1951, 111120. Hunt, D. E., & Hovland, C. I. Order of consideration of different types of concepts. Journal of Experimental Psychology, 1960, 220-225. Kittell, J. E. Intelligence test performance of children from bilingual environments. Elementary School Journal, 1963, 64. Lambert, W. E. & Tucker, G. R. The benefits of bilingualism. Psychology Today, September 1973. Landry, R. G. A comparison of second language learners and monolinguals on diverging thinking tasks at the elementary school level. Modern Language Jorunal, 1974, 58. Lantz, D. L. Language and cognition revisited. Journal of Abnormal and Social Psychology, 1953, 454--462. Lenneberg, E. H. Cognitions in ethnolinguistics. Language, 1973, 463-471. Liedke, W. W. & Nelson, L. D. Concept formation and bilingualism. Alberta Journal of Educational Research, 1968, 4. McDavid, J. W. & Harari, H. Psychology and social bhavior. New York: Harper and Row, 1974. Peal, E. & Lambert, W. The relationship of bilingualism to intelligence. Psychological Monographs, 1962, 76. Sapir, E. Culture, language and personality: Selected essays. D. G. Mandelbaum (Ed.). Berkeley: University of California Press, 1956. Schmidt, W. H. O. Child development: The human, cultural and educational context. New York: Harper and Row, 1973.
Perception of Professional Concepts
127
Whorf, B. L. Language, thought and reality: Selected writings. J. B. Caroll (ed.), Cambridge, Massachusetts: MIT Press, 1956. Young, F. W. TORSCA-9: An IBM 360/75 FORTRAN IV program for nonmetric multidimensional scaling. Journal of Marketing Research, 1968, 319-320.