|
3.3.
Analysis according to linguistic group
The next step is to segment the
sample of pupils into linguistic groups. These groups are defined
by the language spoken between the parents when the pupil was small.
So, the school population is segmented into four groups according
to whether the parents spoke only or above all in Catalan with one
another, spoke in Catalan and Spanish equally, spoke only or above
all in Spanish or spoke other languages. Each of the linguistic
groups is analysed independently and an in-depth look is taken at
their linguistic usages with friends. The main interest is to interpret
differentiated usages between those who above all use their parents'
language and those who use their second language to a more or less
extensive degree.
These new multivariable models we are producing are much less robust.
In the case of the Catalan-speaking group (the parents spoke Catalan
with one another when the pupil was small) a 27.7% fit for the variance
of the dependent variable is obtained, while in the case of Spanish-speakers
the figure is a little higher: 38.3%. By contrast, both in the case
of children of couples who spoke Catalan and Spanish and for those
of couples who spoke other languages, the model is better and is
provided with fewer variables. The children of couples speaking
both languages achieves a 63.1% fit and the second, despite its
internal diversity, achieves 64.9% with just two variables. The
fact that the model for the children of Catalan- or Spanish-speaking
couples should be less powerful provides evidence that family usages
play a vital role in usages with friends. And, in delimiting the
language spoken with friends according to family language, we do
not have variables that are sufficiently correlated to give a joint
explanation.
The factor that most conditions
the language used by the Catalan-speaking group – whose parents
spoke Catalan when the pupil was small – with friends is,
precisely, the language the pupils speaks with his/her parents.
See Table 7. We must highlight the fact the inclusion of the language
spoken with parents seems like redundant statement when the dependent
variable of the model is the language spoken with parents. But in
no way is this the case; as we have seen in Table 1, part of the
population are the children of parents who speak one language but
who have been transmitted a different one. And the children of Catalan-speaking
parents who have been transmitted Spanish maintain a very high correlation
with the greater use of Spanish with friends. Following the lines
of usages with parents, a set of ideological-type questions follows,
focusing on currently desired linguistic usages in the immediate
area ("at your school, would you like to speak Catalan...?")
and not so immediate area ("in Catalonia, Catalan should be
used...?") Finally, in less important positions, come competence
in Catalan, language spoken with brothers and sisters, informal
usages at school and the mother's place of origin.
Table 7. When you were little,
what language did your father and mother speak with one another?
= Only or above all Catalan
|
Standardised coefficients
(beta) |
Significance (standard
error) |
Language with parents |
-,0243 |
0,000 |
In Catalonia, Catalan should be used.... |
0,198 |
0,003 |
At your school, would you like to speak Catalan...? |
0,167 |
0,016 |
Catalan/Understanding |
0,167 |
0,004 |
Language with brothers and sisters |
-0,156 |
0,007 |
Average language outside class with teachers |
0,152 |
0,009 |
La teva mare va néixer |
0,137 |
0,018
|
As for the other big group – the children of Spanish-speaking
couples – the order of the variables is quite different. See
Table 8. In first place, the variable that most correlates with
their language usages with friends is the language spoken with teachers
in more informal contexts. Quite a way behind appear family language
usages (with brothers and sisters) and the group belonged to. And,
now in relatively unimportant positions, are the language spoken
with parents and representations on the use of Catalan. Finally,
in last position, appears self-assessment in oral competence in
English.
Table 8. When you were little,
what language did your father and mother speak with one another?
= Only or above all Spanish
|
Standardised coefficients
(beta) |
Significance (standard error) |
Average language outside class with teachers |
0,280 |
0,000 |
Language with brothers and sisters |
-0,198 |
0,001 |
I'd like to belong to the group who speak... |
-0,182 |
0,000 |
Language with parents |
-0,132 |
0,027 |
Catalan is used... |
0,121 |
0,003 |
English/Speaking |
0,106 |
0,008 |
In the case of the children who
use Spanish and Catalan, the variables that most correlate with
their linguistic usages are the language spoken with teachers in
class, competence in Spanish, the language spoken with parents and
the desire for the use of Catalan in the non-immediate area.
Table 9. When you were little,
what language did your father and mother speak with one another?
= Catalan and Spanish equally
|
Standardised coefficients
(beta) |
Significance (standard
error) |
Normal language of pupils with the teacher in class |
0,517 |
0,000 |
Spanish/Understanding |
0,425
|
0,000 |
Language with parents |
-0,389 |
0,000 |
In Catalonia, Catalan should be used.... |
0,297 |
0,002 |
|