Structural Bilingualism and Marketing Strategy : The Case of Franco-Arabic Bilingual Segmentation

Marketing research is generally interested in the linguistic aspect of communication within the framework of the choice between the standardization and the adaptation of an international and/or intercultural communications strategy (Clark 1990). However, with the development of foreign languages through the immigrant communities in the rich countries on the one hand, and the economic growth of the structurally bilingual developing countries on the other hand, the markets’ multilingual problem became increasingly important and the strategic choices became rather complex (Usunier and Shaner 2002).


Introduction
Marketing research is generally interested in the linguistic aspect of communication within the framework of the choice between the standardization and the adaptation of an international and/or intercultural communications strategy (Clark 1990).However, with the development of foreign languages through the immigrant communities in the rich countries on the one hand, and the economic growth of the structurally bilingual developing countries on the other hand, the markets' multilingual problem became increasingly important and the strategic choices became rather complex (Usunier and Shaner 2002).
In terms of marketing communication, a bilingual or a multilingual market is thus a market in which the consumers regularly use two languages, or more, at the same time.However, all the used languages do not have the same status and thus do not have the same effects on the consumer's behavior.First we have the official languages, often related to the cultural background and the regional dialects; they are often related to the traditional uses.Then, we find the foreign languages, related either to immigration factors or to _________________________________________________________________________________ ______________ Slim Khalbous and Sawssan Radhouan (2013), Journal of Marketing Research and Case Studies , DOI: 10.5171/2013.364200.historical factors; and finally the mixed languages or code-switching, stemming from a complex mixture between the original languages and the borrowed languages (loan word) (Paulin 1994).However, this linguistic plurality, become a worldwide phenomenon, which generates more researches on the immigrant bilingual communities, than on the structurally bilingual emerging markets (Ueltchy and Krampf 1997, Khalbous 2003).The question here is whether multinational enterprises have enough information about the difference of the potential impact on their marketing strategy in terms of the used language today?Do they know enough the differences between the bilingual groups forming the emerging markets?The numerous case studies of managerial failures, allow saying that the answer is rather negative.This research exactly suggests approaching the problem of structural bilingualism of the emerging countries, where consumers' bilingualism may vary in quantity and in quality; and where the decision concerning the language to use in marketing communications strategy often remains difficult to take, more specifically the case of Franco-Arabic bilingualism in Maghreb countries which will be the subject of this research.The key issue being the following: can we identify a bilingual segmentation to better understand and to better target Franco-Arabic consumers?
The objective of this research is consequently double; it's at once methodological and managerial.Methodological, because the research presents a rather rich approach (qualitative, and quantitative) in order to construct a measurement scale of Franco-Arabic bilingualism (attitudinal and behavioural) which allows to identify the main segments of this bilingualism.From a managerial point of view, the research tries to present an operational bilingual segmentation tool for the companies.This research is presented in two sections.The first section is abstract; it approaches more particularly the question of structural bilingualism as opposed to community bilingualism, especially through Franco-Arabic bilingualism.As for the second section, it deals with the methodology and the results, presents the approach stages and especially the identified Franco-Arabic bilingual segmentation.

Structural Bilingualism: Linguistic and Cultural Aspect
Linguistic acculturation and the concept of structural bilingualism: what definitions to retain and which are the questions which remain unresolved?

The Definition of the Concept of Bilingualism
The individuals' bilingualism was approached according to various points of view according to the linguist, the psychologist, the psycholinguist, the sociologist, the sociolinguist, and more recently the marketer disciplines, etc.The definitions are numerous and different.
They emphasize either the criterion of linguistic skill, or the criterion of the attitudes and languages use according to the communication context (Hormann 1972, Paulin 1994, Demorgon and Lipiansky 1999).With respect to the objective of this research, which is the identification of several types of consumers' bilingualism, the approach of Demorgon and Lipiansky (1999) seems to be the most adequate.Indeed, it integrates the dimension of imbalance in the process of bilingualism; whereby the term bilingual may take various senses; a common sense and a scientific sense.A bilingual, according to the common sense, is the person who perfectly masters two linguistic systems at the same level as a native speaker of each of them.While according to the scientific sense, which is more pragmatic, a bilingual is anyone who uses, in more or less important doses, elements of two different linguistic codes.This dimension is very useful in marketing because the companies which operate in a bilingual target must be aware that the 2 nd language (L2) of the market is at once an instrument of further communication compared to the original language (L1), but it can also sometimes be, a more or less conscious obstacle to the understanding (Luna and Peracchio, 2005).The researches have showed that the language of the message influences the advertising efficiency through the cognitive processing procedure and this through influencing the consumer's attention and his understanding of the message (Wyer 2002).
Similarly, the language affects the system of thought as it influences the state of mind and the feelings of the individual.When a person faces an advertisement in his original language (L1) or in his second language (L2), he will automatically activate the system of thought relative to one of these two languages (Noriega and Blair 2008).Puntoni and al (2009) have showed that an advertising slogan realized in the original language (L1) of the target is perceived as more emotional than the other which is realized in the second language.And this effect is not only due to the activation of specific stereotypes in the language or to the lack of understanding but it also depends on the frequency of use of the language.Also, the interaction between the two languages (L1) and (L2) may draw the attention to the target of the company and increase its memorization (Ahn andLaferle 2008, Krishna andAhluwalia 2008) as it may influence the positioning of the company's product (Krishna and Ahluwalia 2008).Caroll and Luna (2011) have revealed that the interaction between the content of the advertisement and the language affects the attitude towards the advertisement in a bilingual context and the latter may favorably influence the implication towards the advertisement (Bishop and Peterson 2011).
The example of the Hispanic bilinguals' typology in the United States developed by Marcos (1988) confirms this notion of dominance of a language over the other.The author has identified two types of bilinguals: "Subordinate Bilingual", who can be differently successful in each language and often experiences certain difficulty with the second language; and the "Competent Bilingual" who shows certain lucidity and speech fluency in both languages, however they are less numerous.

From Community Bilingualism to Structural Bilingualism
In the literature on marketing, most of the researches which have been realized essentially apply to the case of minority bilingualism that is to say the case of 'niche market' of an immigrant community in a country which is not necessarily bilingual (Hispanics in the USA, Maghrebians in France and Turkish immigrants in Germany…).
Several examples can be quoted, Cofresi and Gorman (2004) showed that bicultural Hispanic immigrants maintain close relations with their original culture throughout their life, and acquire an adapted behaviour to their two cultures by developing two separated value systems and attitudes.Guttfreund (1990), Toffoli and Laroche (2002) and Cofresi and Gorman (2004) have tried to study the effects of bilingualism, by showing that the language, and in particular L2, creates a specific access and an impact on the intelligence, the perception, the attitudes, the beliefs, the memory as well as the personality characteristics.
Generally the studies on acculturation and linguistic acculturation in particular, often develop the idea that bilingualism inevitably leads to a confusion of linguistic and psychological models necessary for the development and the stability of the individual.This instability including other reasons is due to the fact that the environment in which the individual lives is not bilingual.Thereby, there is a phase shift between the individual and the society.
Then, what would be this effect if the society is in general, historically and culturally bilingual?It is the case of all the formerly colonized countries: for example, the English language in India or Egypt; or the French language in the Maghreb countries or Africa.

The Case of Franco-Arabic Structural Bilingualism in Maghreb
The linguistic landscape of the Maghrebian market can be considered bilingual and even multilingual in so far as there are several languages of different statuses which coexist: dialect or spoken Arabic (L1: mother tongue originating in an oral tradition), literary Arabic (L1 bis: standard and written official language) and the French language (2 nd common language).Today, this form of Franco-Arabic communication makes an integral part of the daily conversations of the Maghrebian societies.Bentahila (1983) has described bilingual Moroccan as people who are incapable to correctly speak two languages and incapable to valorise them.He has discovered in his study of bilingualism (Moroccan dialect versus the French language), that the French language reflects a status of wealth, a social importance and a higher educational level.On the other hand, the Moroccan dialect reflects religious dimensions.In the same vein, Riahi (1970) and Steven (1983) have noted that the French language in the Tunisian context has a connotation of modernity, whereas the literary Arabic is more associated with traditionalism and conservatism.
However, more recent studies have shown that attitudes and behaviours towards bilingualism evolve to reach a sense of identity withdrawal and thus refusal of the French language.Lawson and Sachdev (2000) have studied the attitudinal and the behavioural dimensions of code-switching (CS) in Tunisia.The results have revealed that attitudes towards CS seem to be relatively negative compared with the results of studies obtained in other contexts.These discoveries also suggest that these negative attitudes of Tunisians towards CS have not yet been reflected by real languages use.Indeed, the negative and ambiguous attitudes towards bilingualism (Tunisian dialect versus the French language) are not reflected by the linguistic behaviour of the majority of the Tunisians.They assert a sense of identity which results from a more important use of CS in private, in family (intra-group); rather than in professional situations (out-door).This difference in the use of bilingualism from a group to the other which has been confirmed by the studies of Khalbous (1999 and2003) who has identified different bilingual groups according to two (2) discriminating axes: professional and personal use of bilingualism.These studies have shown that the impact of advertisement was very different in terms of targeting these bilingual groups.This literature review, confirms that from the point of view of marketing research, particularly in structurally bilingual markets, it is important to identify the groups of consumers according to their attitudes and their behaviours towards bilingualism; which will allow researchers and practitioners in marketing to better know the behaviour of their consumers and especially to know how to better target them by linguistically and culturally accurate advertising campaigns.

Application in the Case of Franco-Arabic Bilingualism
The history and the geographical location of Tunisia, between Africa, the Arab world and Europe, explains that Tunisia belongs to the countries which have inherited an ethno-linguistic diversity.Tunisia has always been described as a structurally bilingual country and even multilingual, which makes of it a good field of application in order to measure the effects of structural bilingualism in marketing and particularly Franco-Arabic bilingualism.

The Followed Methodology
The research has been conducted over several phases according to Churchill's paradigm (1979)

Advantages and disadvantages of bilingualism
Attitude towards bilingualism among adults Attitude towards bilingualism among children

Situation of use of French or literary Arabic
Attitude towards the use of the local language (Tunisian Arabic) Attitude towards the use of the official language (Literary Arabic) Attitude towards the use of the foreign language (the French language)

Bilingualism and consumer behaviour
Behavioural use of bilingualism Variables to be explained

Bilingual's Measuring Instrument
The questionnaire has been at first drafted in the French language and then it has been translated into Arabic.This translation has been realized by three perfectly bilingual readers according to several methods (Marchetti and Usunier, 1990;Khalbous, 1999)._________________________________________________________________________________ In order to test the reliability of the measures several consecutive principal component analyses (PCA) have been conducted and followed by an analysis by Cronbach's Alpha.All the retained items at this stage had a Cronbach's alpha superior or equal to 0,6.The factor analysis has ended and the reliable items have been reserved.We have initiated a confirmatory analysis of the retained factor structure.

______________
A confirmatory factor analysis (CFA) has been conducted through structuralequation method under AMOS 4 in order to validate the final form of the questionnaire.
The parameters of the model; the reliability, the convergent and discriminant validity of these factors confirm the dimensions of the construct.Table 3 summarizes the set of the adjustment indices and the measurement scale of the convergent and discriminant validity.(1971) for the analysis of the internal coherence of the scale.There is no exact rule of evaluation, if it is not the application of the same criteria as for Cronbach's alpha.
In the case of this scale, we have obtained for each construct coefficients superior to 0,8.This has demonstrated the stability of the scale.Thereby, the test of internal coherence according to Joreskog (1971) has revealed that the asked questions have shared enough common notions; so that the same instrument can produce similar results when it is administered to various populations.
In order to calculate the nomological validity, we have used the measure of Valletta-Florence (1988) which represents the percentage of the contained information in the explained dimensions by the construct.The nomological validity confirms a significant explanatory power of i on the j.This has thereby allowed globally validating the model as well as the dimensions which have been retained in the previous factor analyses.

The figure below presents the scale attitudes towards reliable and valid
Franco-arabic bilingualism composed of 16 items.

Cluster Analysis
The implementation of typology includes a classification in dynamic contrast followed by a discriminant analysis.The first method has allowed us to determine 3 groups of homogeneous iterative bilinguals (G1=60 individual; G2=80 individual and G3=39 individual).And the second method has allowed us to confirm the belonging to a segment according to the items of the questionnaire.Naturally, certain number of statistical processing have been realized at first by using the "step by step" calculation method, then by verifying both the normal distribution of data and the equality condition of variance-covariance matrices by the box test (see table 4).

Table 4: Test of the Null Hypothesis of Covariance Matrix
Variable to be explained

Box's M test F approached Signification
Groups of bilingualism 517,118 1,630 0,000 Finally an examination of the proper values has been conducted, as well as the measurement of the discriminant power of the explanatory variables among the different segments of Wilks coefficient (see Table 5).
1. Bilingualism among children develop their learning capacity 2. Bilingualism among children broadens their information horizon 3. Bilingualism among children help them in their studies 4. Bilingualism decreases our authenticity 5. Bilingualism among children makes them psychologically disturbed between two languages 6. Bilingualism distorts our language 7. Bilingualism is a sign of lack of self confidence 8.The use of a foreign language allows better expressing ideas 9.The use of a foreign language is a way to follow global development 10.The use of a foreign language is mandatory for scientific fields and research 11.The use of a foreign language is mandatory for technical and technological fields 12.The use of the spoken language is a way to facilitate communication between people 13.Your conversations and discussions with your close friends 14.Write your various reports and business letters 15.Watch TV 16.Your conversations and family discussions

Attitudes towards bilingualism in general
(7 items)

Attitudes towards the use of the local language
(1 item)

Behavioural use of bilingualism (4 items)
From item 1 to item 12 = Likert scale « Agree / Disagree » 5 points.At this stage of the analysis we are going to interpret the segments and to describe the different groups.

Discriminant Function Analysis
The results of this type of analysis have provided us with two discriminant functions.They have a good capacity to well place the consumers of 96,1 %.Similarly, they have an important explanatory power: F1=67,24 % and F2=62,41 % of the total variability (Table in appendix 1).The first function is mainly defined by items ( 5), ( 6) and ( 7); it is the axis which mainly represents the negative or disturbing side of bilingualism (identity and psychological arguments).The second function is mainly defined by items ( 1), ( 3), ( 10) and ( 11); it is the axis which mainly represents the positive or the necessary side of bilingualism (pedagogical and technological arguments).

Identification of Franco-Arabic Bilingual Segments
The analysis of the items' averages which discriminate most shows that the three groups distinguish themselves more clearly through cognitive and emotional attitudes then through behavioural attitudes.The below table summarizes bilingualism average per group.The graphic synthesis of the results (see Figure 3) shows that the most correlated proposals to axis 1 allow sharply distinguishing G2 from G1 and G3.Whereas the most correlated items to axis 2 allow sharply distinguishing G1 from G2 and G3.
Group (1) rather has well-balanced and functional attitudes.It is neither particularly hostile, nor particularly favourable to the foreign language (L2).
The consumers belonging to this group admit the fact that bilingualism may have negative effects on the identity, but without excess and without ideology.They think that the contribution of the foreign language is particularly important for the technological domains.Similarly the behaviour of the consumer is in accordance with his attitudes, because he exclusively uses the Arabic language (local dialect: L1) in his personal conversations for example.On the other hand, for media consumption, the use of the Arabic (L1) and the French (L2) languages is practically equivalent.The consumers of this segment will be considered as: "Pragmatic bilinguals".
Group (3) has double negative attitudes towards bilingualism.It tends to maximize the harmful effects of the foreign languages and to minimize their possible advantages.This group is made of culturally hostile consumers to any opening, but since the study concerns a structurally bilingual market, the study of the behaviour shows that this group uses the French language (L2) as well, even though in sharply lower doses than the other groups.The consumers of this segment will be considered as: "Forced bilinguals" or "Frustrated Monolinguals".
The extremely negative attitudes of this group (group 2) are rather astonishing.Indeed, this group presents at the same time the most elevated use rate of the French language (L2) in personal conversations, in media consumption, and in professional use.This is why the consumers of this segment will be considered as: "Ideal bilinguals".
These results are fairly consistent with the unique typology of the Franco-Arabic bilingualism than those which are presented in the study of Khalbous (1999).The author has equally distinguished three groups with respect to the bilingual behaviour: the "Literals" which present a rather technical use of the foreign language rather close to the "Pragmatic bilinguals", the "Monolinguals" those who tend to reject the foreign language, rather close to the "Frustrated Monolinguals", and the "Bilinguals" those who really master and strongly use the French language, are rather close to the "ideal bilinguals".

Complementary Descriptive Analysis of the Groups
Several complementary analyses of the three identified segments of bilingualism have been realized in order to verify the homogeneity degree and the representativeness of these segments.These analyses have thereby concerned the chosen language to answer the questionnaire, the city of residence and the socio-demographic dimensions.
Each respondent had the choice to answer the questionnaire either in French (L2) or in Arabic (L1).The free choice of the language had a double objective, first to eliminate language obstacle as a refusal motive for answering the questionnaire, and second to have an additional socio-cultural index which allows refining the results.The proportion of the consumers who have chosen to answer the questionnaire in Arabic is about 56,1 % versus 43,9 % in French.These percentages and the results of Pearson's Chi-squared test show once again that the daily expression of bilingualism as well as the chosen language to answer the questionnaire are not a significantly discriminant variables of bilingual segments.This result confirms that the use of one of the two languages in a structurally bilingual market does not foreshadow the attitudes of the consumers.
As for the other parameters such as the city of residence, the gender, the age, the educational level, the income, the profession and the civil status, the analysis have all shown insignificant discriminant results among the three bilingual segments.It would be clear from this that bilingualism may although represent a very strong and very significant segmentation criterion for marketers.

Conclusion
By referring to the realized on-site application within the framework of this research, the qualitative as well as the quantitative, it is possible to assert that structural bilingualism very closely affects consumers in their everyday life as well as their close environment: family, school, work, and media.Also the second-order confirmatory factor analysis has confirmed and validated four significant dimensions of structural bilingualism: the attitudes towards bilingualism in general, the attitudes towards the use of the foreign language (L2), the attitudes towards the use of the spoken language (L1) and the behavioural use of bilingualism.These different facets take into consideration, as it's shown in the below summary table, a large part of the diversity of the issues related to bilingualism and help to segment the consumers in three major bilingual  These consumers must be targeted beforehand and in priority with L2.Because in practice it is the target which uses and which most valorises the foreign language.But, the advertisers have to pay attention, because they are likely to encounter negative reactions in some cases, when their authenticity is threatened.But it is likely that this would be more connected to the contents rather than the language itself.They are consumers who must be targeted exclusively with L1.Because, even if they master L2 (often partially), they tend to systematically emphasize its negative aspects.They are likely to develop an avoidance strategy towards any marketing communication in L2.
The results have similarly showed that it was necessary to pay attention to the possible differences between attitudes and behaviours, because these delicate dimensions, like the identity, may even appear with a perfectly bilingual target.
Nevertheless, this survey deserves to be followed with a broader application.On the one hand, the size of the samples especially of the quantitative survey (70+150+180) can be increased in order to enhance the external validity, even if the size is technically acceptable for the method of structural equations used for scale validation.
On the other hand, a comparative intercultural application in other structurally bilingual markets would doubtless have a significant contribution to further refine the results.
Thereby, in terms of the investigation method, it would be interesting to use indirect methods, like projective techniques, in order to avoid possible bias of answers related to the subject's sensitivity, which emerged in diverse onsite stages.This identified Franco-Arabic bilingual segmentation in a structurally bilingual market opens interesting perspectives for marketing applications, particularly in terms of impact measure on numerous domains as diverse as advertising, point-ofsale display (POS), packaging, vendor negotiation, and commercial signs, etc.

SlimFigure 1 :
Figure 1: The Procedures of the Realized Transcultural Translation of the Questionnaire Sawssan Radhouan (2013), Journal of Marketing Research and Case Studies , DOI: 10.5171/2013.364200.
In such context of structural bilingualism what would be the attitudes and the behaviours towards it?Is it similar or not to community _________________________________________________________________________________ ?Actually, very few studies are available in marketing research.

Table 1 : Synthesis of Research Survey
are summarized in the following table: ______________ Slim Khalbous and Sawssan Radhouan (2013), Journal of Marketing Research and Case Studies , DOI: 10.5171/2013.364200.elements