Opening perspectives from an integrated analysis: language attitudes, place of birth and self-identification

ABSTRACT There is a theoretical and empirical tradition demonstrating the influence of the place of birth and self-identification in the shaping of language attitudes. But very few works analyse their joint effects. The main aim of this study is to analyse both the disaggregated and joint impact of these variables on the shaping of attitudes towards Catalan and Spanish languages in Catalonia. The sample consists of 309 adolescents (aged 14–16 years) coming from Ecuador, the rest of Latin America and Morocco. Both variables have an influence on attitudes towards Catalan separately, but when analysed jointly in the same explanatory model, self-identification outweighs place of birth. This is more evident among young people who develop the most and the least favourable attitudes. Furthermore, attitudes towards Spanish are only influenced by the place of birth, and by none of the variables in the case of those with the most and least favourable attitudes. These results open new lines for research and reflection on the process of construction of language attitudes and their role in a successful process of integration.

This had an effect on educational, reception and integration policies. Within the area of education, recent reviews of the Plan for Language and Social Cohesion (Pla per a la Llengua i Cohesió Social), better known as Pla LiC (Generalitat de Catalunya 2007, had the main aim of 'promoting and consolidating social cohesion, intercultural education and the Catalan language within a multilingual framework' (Generalitat de Catalunya 2009, 14). In order to achieve this end, a new model of language immersion is promoted in which Reception Classrooms and Environment Educational Plans are reinforced in addition to pedagogical activities adjusted for the classroom.
From a more general perspective, integration policies stress the identification with Catalonia as one of their principles. Within this identitarian project, Catalan should be the backbone of a common and multilingual project of coexistence, including the knowledge of the languages present in Catalonia (Catalan / Spanish / immigrants' L1) as an element of progress and social mobility (Generalitat de Catalunya 2009. But as Lewis (1981) states, the success in the design and implementation of any policy directly or indirectly related to languages is tied to the knowledge of the population's language attitudes. Thus, the effectiveness of a policy would be related to its effect on the attitudes of the actors involved, its capacity to transform the attitudes of those who do not share them and the analysis of the causes behind these attitudes.
Therefore, the analysis of language attitudes transcends the cognitive, affective and behavioural elements that compose them, and connects directly with elements of selfidentification (Garret 2010; Hogg and Smith 2007; name deleted to maintain the integrity of the review process 2010; Tajfel and Turner 1986;Woolard 1991). For 'autochthonous' people, but especially for the descendants of immigrants 1 , such self-identification refers to fundamental aspects such as how they perceive and feel their sense of belonging or not in society, the development or not of bonds of affection with the place, and their attitudes (Haines 2007; name deleted to maintain the integrity of the review process 2010; name deleted to maintain the integrity of the review process 2014; Portes Vickstrom and Aparicio 2012).
Different approaches analysing the role of language in the construction of identity sustain that language attitudes are not a mere marker of identity (Castelan and Giles 1997), but they also become the attribute par excellence of identity itself (Edwards 2009;Fishman 1977; name deleted to maintain the integrity of the review process 2008). Consequently, they reflect belonging, and are mutually tied to ideologies and language preferences (Alarcón and Parella 2013;Trenchs and Newman 2009;Woolard and Schieffelin, 1994).
In this sense, some works that do not specifically focus on how self-identification influences language attitudes, do focus implicitly on closely related processes such as ideologies or language preferences. At this point it is impossible not to mention Woolard's classic works (1989) and Woolard and Gang's (1990). They were both developed during the 1980s, coinciding with the access to education of the children of immigrants from different parts of Spain received in Catalonia during the 1970s. The first of these works, carried out in 1980, when the process of language normalisation was in an embryonic state, concluded that both the group with a Catalan background and the group with a Spanish background showed a clear preference for their own language in the affective dimension. At the same time, rejection was observed towards Spanish speakers whose accent was noticeable when speaking Catalan. The study was repeated seven years later and, although the preference towards one's own language persisted, the perception of the Catalan spoken by Spanish speakers had improved. Newman, Trenchs and Ng (2008) and Trenchs and Newman (2009) adapted and completed a similar analysis focusing, in that case, on the 'autochthonous' and Latin American groups that had reached Catalonia during the last immigration wave. Their main conclusions suggested that there were differences between the two groups with regards to their language ideologies that would directly affect attitudes towards languages, on the basis of a preference towards multilingualism (Cosmopolitanism) or monolingualism (Parochialism). Moreover, the Latin American collective does not reproduce the exact same pattern followed by Spanish speakers in the 1980s. Although they value the peninsular Spanish as a language of international prestige used in Catalonia and in the rest of Spain, Catalan is also valued as a language of integration and social progress. This is added to the symbolism conferred to their non-peninsular Spanish, tightly connected to their identity (Trenchs and Newman 2009 Name deleted to maintain the integrity of the review process (2010) complement this study with an analysis of the discourses that sustain the language attitudes of migrated people.
Their analysis shows these discourses are determined by the aforementioned variables, but point out the condition that the young immigrants identifying themselves more with Catalonia and feeling valued and integrated at school and the society, exhibit better attitudes towards languages, higher competence levels and more balanced patterns of language use. One of the reasons for the development of this self-identification and recognition is the possibility to receive classes of the language and culture of their country of origin.
Finally, in Madariaga, Huguet and Lapresta (2013), 1,206 'autochthonous' and immigrant students in their 2nd and 4th years of secondary education in Catalonia completed a questionnaire on language attitudes. Although attitudes are mostly positive towards Catalan and Spanish, it can be concluded again that, in the case of immigrant students, these vary depending on being Spanish-speaking or not, the family language, the length of stay and the age on arrival. Thus, Spanish speakers, who mostly use Spanish as their L1, show more positive attitudes towards this language than towards Catalan when compared to other groups.
In the same way, those born in Catalonia show better attitudes towards Catalan.
As a general trend, all these analyses tend to reveal the significance of origin and the importance of aspects which are interconnected with self-identification. This heterogeneity is also evident in the self-identification with the receiving society. In an analysis about predictors of self-identification with Spain among second-generation teenagers in Madrid and Barcelona, Portes, Celaya, Vickstrom and Aparicio (2012) show that young people coming from Latin America present a differentiated behaviour: being Ecuadorian is a predictor for stronger self-identification with Spain, while being Argentinean, Chilean or Venezuelan is not (again, with very similar scores among them).
With this background and with the aim of advancing the knowledge on the construction of language attitudes of young people descended from immigrants with a greater presence in Catalonia (from Maghreb and Latin America), we need studies with different approaches that can shed new light on how these variables are interlocked.
The impact of self-identification and the place of birth on the language attitudes of young immigrants, whether in isolation or combined has been rarely analysed. That is, given the proven importance attributed to language/s in Catalonia in educational, identitarian and welcoming policies, the main objective of this paper is to analyse the impact and interrelation of (i) the place of birth and (ii) self-identification with Catalonia and Spain in language attitudes towards Catalan and Spanish.
We focus on the most important areas of origin and, specifically, on the most prevalent countries of origin (Morocco / Ecuador / Other Latin American countries). Also, in order to make a more in depth analysis of their interrelation, a global analysis is performed, as well as an analysis of the subjects showing the most and least favourable attitudes towards the languages in the study.

Objectives
The objectives of this paper are:

Participants
The total sample consists of 309 young people in 10 Secondary Schools in Catalonia.

Variables
The variables considered for our research are: *Attitudes towards Catalan: variable with a range between -10 and +10 where -10 is a totally negative attitude, 0 is totally neutral and +10 totally positive.
* Attitudes towards Spanish: variable with a range between -10 and +10 where -10 is a totally negative attitude, 0 is totally neutral and +10 totally positive.
*Place of birth: admits the categories Ecuador / Other Latin American countries 3 / Morocco.
*Self-identification with Catalonia: admits the categories NO / YES. 4 *Self-identification with Spain: admits the categories NO / YES.
The distribution of the sample according to the independent variables is shown in Table 1. The questionnaire is composed of twenty statements with dichotomous reply (YES/NO), ten for each language. Each of them is given a +1 / -1 score, depending on whether the answer is favourable or unfavourable towards that particular language. Thus, through score addition a scale ranging from -10 (completely unfavourable attitudes) to +10 (completely favourable attitudes) is obtained, 0 signalling neutral attitudes.

Treatment of results
With the aim of observing in a disaggregated manner the impact of the place of birth and the self-identification with Catalonia and Spain on language attitudes towards Catalan and Spanish, statistical descriptions (Mean and Standard Deviation) and mean comparison tests have been used for the data analysis. In this case, ANOVA and the Bonferroni test were used; the latter for more in-depth analysis when the first result is significant.
As a second stage, the General Linear Models (GLM) have been carried out to establish the joint effect, specify the interaction and the importance of the place of birth and the self-identification with Catalonia and Spain, in the explanation of the attitudes towards Catalan and Spanish variance.
These analyses have been used both with the total sample and with the subjects that show the most and the least positive attitudes.
Based on a previous data study regarding manifested language attitudes and their distribution, the young people in the fourth quartile, scoring 7 and more, were considered those with the best attitudes, while those scoring between 2 and -10 were considered the least favourable, corresponding with the first quartile.

Results
With the aim of obtaining the maximum clarity, results are presented in three sections. The first section contains an analysis of the total sample, the second section focuses on the young people who showed the most positive attitudes, and the third section focuses on the young people who showed the least positive attitudes.

Catalonia and Spain. Global Findings.
Taking the whole sample into consideration, the mean of attitudes towards Catalan among immigrant young people is favourable, with a mean score of 4.5469, and a standard deviation of 4.7845. In the case of Spanish the mean is higher: 7.009 (SD=3.0837), the difference between them being statistically significant (t(308)=-7.222; p<.000). 5 In the case of attitudes towards Catalan, both the place of birth and self-identification p<.000 / FMorocco-Latin America(2,306)=13.205; p<.000). Self-identification with Spain produces no differences.
These general results point to both variables having individual effects on attitudes towards Catalan. Only the place of birth has effects on attitudes towards Spanish. It is therefore not possible to elucidate their joint interaction, or which of them, if any, has more weight in the explanation of the differences found.
The General Linear Model (GLM) ( Table 2) carried out for attitudes towards Catalan indicates that the model itself explains a significant proportion of the variance attitudes towards Catalan (for the signification of the Model Corrected; p < .000). It also informs us that the variables place of birth and self-identification with Catalonia have an individual effect, as has been seen above. However, the most interesting is that the interaction between place of birth and self-identification with Catalonia is not significant, which implies that they do not have a joint effect.

TABLE 2 NEAR HERE
An Interaction Graph explains the articulation of the influence of the variables ( Figure   1). As a general rule, the pattern of attitudinal behaviour is rather similar; although attitudes towards Catalan may be more or less positive among groups, those who identify more with Catalonia show, in all cases, more positive attitudes than those who do not.

FIGURE 1 NEAR HERE
Regarding Spanish, results from both the GLM and the Interaction Graph (Table 3 and Figure 2) indicate yet again that only the country of birth has an incidence on attitudes towards Spanish,. The model on the whole is significant, as is the effect of country of birth, while self-identification with Spain and the intersection between the two variables are not.

Catalonia and Spain. Young people with the MOST favourable attitudes
Among those showing the most positive attitudes towards Catalan, the first difference in respect of the total sample is that their attitudes are not influenced by the place of birth. 6 However, the introduction of self-identification with Catalonia does produce some variations. In this case, those who do not identify at all show a mean attitude value of 8.3750 (SD=1.0080), while those who do, show a mean attitude value of 8.9541 (SD=1.0218), the difference being statistically significant (F(1,139)=7994; p< .005).
The GLM including both variables confirms these results, self-identification with Catalonia being the only significant variable (Table 4).

TABLE 4 NEAR HERE
The Interaction Graph also confirms that, among those with the most positive attitudes towards Catalan, these attitudes are more related with their identification with Catalonia than with their place of birth (Figure 3). (SD=1.0254) and the mean of those who do identify with Spain is 9.0435 (SD=1.0034). 7

Attitudes towards Catalan and Spanish, Place of Birth and Identification with Catalonia and Spain. Young people with the LEAST favourable attitudes
In a manner similar to the subjects with the most favourable attitudes towards Catalan, those Prevalence of self-identification is stated again in the GLM (Table 5), as well as in the Interaction Graph (Figure 4).

Conclusions and Discussion
The effect of the place of birth and self-identification on attitudes is obvious and complex. preferences and self-identifications of the Latin American group (Alarcón, Parella and Yiu 2014;Portes, et al. 2012, among others). This issue certainly requires further in-depth qualitative and quantitative studies.
Moreover, self-identification with Catalonia produces better attitudes towards Catalan, something which does not happen with self-identification with Spain and attitudes towards Spanish.
Looking into the case of attitudes towards Catalan at the level of the whole sample, the effect of both conditions can be proven separately, but when considered jointly, selfidentification with Catalonia is the one that modulates attitudes towards Catalan. This is more obvious in the case of young people with the most and the least favourable attitudes.
All this seems to suggest that the place of origin may produce attitudes towards Catalan to be more or less favourable, but what makes attitudes improve is a high degree of selfidentification with Catalonia. Obviously, additional studies are required to further investigate this issue.
Nevertheless, when delving into the case of attitudes towards Spanish, only the place of origin produces this effect and only at the level of the whole sample. Self-identification with Spain or the place of origin has no effects on subjects who show the most and the least favourable attitudes.
These points have important repercussions at the theoretical level and in the Catalan context.
The fact that the improvement of language attitudes towards Catalan is tied to selfidentification with Catalonia is connected to the role given to this language within the project of coexistence to be constructed; a project with a multilingual approach but with Catalan as its symbol par excellence (Fishman 1977), with a communicative function and also a participatory function (name deleted to maintain the integrity of the review process 2006; name deleted to maintain the integrity of the review process 2008). A possible effect is that the development of a feeling of belonging is also expressed through the language and is manifested in the language attitudes (Trenchs and Newman 2009;Trenchs, Patiño and Ng 2008;Woolard 1991).
Language attitudes towards Spanish find themselves in a completely different situation.
In spite of the lack of the necessary in-depth studies on this process, elements such as the international prestige of the language, its instrumental value, its important social presence in Catalonia (Galindo 2008;Vila-Moreno 2004) and family language factors (Madariaga, Huguet and Lapresta 2013) seem to produce positive attitudes towards it.
Taking into consideration these conclusions, and evidently without overlooking the place of origin, work needs to be continued to enable young people's self-identification with Catalonia, especially, a self-identification that implies the development of a sense of belonging and emotional ties with the country (Haines 2007;. In the case of the children of immigrants, research carried out both in the Catalan context and in others, shows that this process is more effective the moment a social value is given by the host society to the languages and cultures brought by these young people (Coelho, Oller and Serra 2013;Cummins 2001;Huguet, Janés and Chireac 2008;Lapresta et al. 2010;Trenchs and Patiño 2013).
Moreover, given the weight of language attitudes in any policy regarding languages (Lewis 1981), it would be a simplification to focus only on one language and only on strictly linguistic measures in multilingual contexts. The present time is not unfavourable in Catalonia.
Largely thanks to the policies on language, education, reception and integration currently in force, attitudes towards Catalan are mainly favourable. But this means that work needs to be done at the non-linguistic and linguistic levels; the latter being as important as those seeking selfidentification with a project of multilingual coexistence, both for the children of immigrants and for the rest of the society.