a) Effect of age factor

Analysis of the results in table 1 shows that almost all interviewees have a high probability factor when it comes to maintaining the traditional form of the third person morpheme, which is clearly not dropping out of use. At the same time, it can be seen that the interviewees aged 6 to 11 are the only ones who show a low probability factor for maintenance of the [e] (0.28). These results can be explained, as I have already observed in other studies of the Lleida dialect (see Creus, 1997; Carrera-Sabaté, 2002b), by the influence exerted by the schools' speech model on the speech of their students. This influence becomes less strong when the speakers move on to higher or further education, as can be seen in the studies by Aniansson (1979) and Portz (1979), on Swedish and English respectively, and this is evident too, in the results from the respondents aged 21 to 30 in our Alguaire study, who have had schooling in Catalan.

b) The influence of schooling and knowledge of written Catalan

Taking these two variables overall, neither alters the probability of maintenance of the traditional [e] variant to any significant extent. It should be noted, nonetheless, that there is a group of speakers with a greater probability of alternating between [e] and []: they are the respondents with secondary school education (0.54) and those with knowledge of written Catalan (0.60).

If we cross the data from the variables age, knowledge of written Catalan and schooling, we observe straightaway that all interviewees with knowledge of the written language maintain the [e] less consistently than those who have none. Having said that, if we analyse the behaviour of speakers aged from 41 to 70, some with knowledge of written Catalan and others not, we see little divergence and only in the 51 to 60 age group do the differences in maintenance of the [e] approach 20 %:

Figure 1. Percentage maintenance of the [e] realisation

Apart from this, the combination of educational level with age group brings the following to light: 1) In the 51 to 70 age group, and more especially in the 51 to 60 age group, the interviewees with secondary education produce least [e], followed closely by those with higher education. 2) In the 31 to 40 age group, interviewees with higher education use significantly less [e] than those with only secondary education. 3) Among speakers aged from 21 to 30 years, who were the first to be actually educated in Catalan (so older speakers obtained knowledge of written Catalan outside school, if at all), these show a maintenance of the [e] variant at a different level of frequency from the other age groups (see figure 2): speakers with higher education maintain the [e] more than 90 % of the time, and in contrast, those with secondary education, do so at rather more than 10 %. Clearly then, on this showing, speakers with higher education are more conscious than others about the issue of conserving native dialect features. 4) The interviewees aged from 12 to 20 are the ones that least use the [e] variant, although analysis of the probability indices do not show them to be the most innovative speakers.

Figure 2. Percentage maintenance of the [e] realisation

4.1.2 The interview and the conversations

Analysis of less formal speech from the 30 inhabitants of Alguaire point to very different conclusions from those to be drawn from the questionnaire with its closed questions. In general, the final vowel of the third person of the imperfect, the conditional and the present indicative in verbs of the first conjugation, practically always gets realised as [e]. Only in the case of two interviewees (a young girl and a young boy) in the 6-11 years age group did we note the variant []. The variants recorded from these two youngsters were as follows: 1) [e] was used in 19 verb sites by the girl and in 23 by the boy; 2) [] was used twice by the girl and twice by the boy.

Apart from this observation, the analysis of the verb inflection for the 3rd person of the imperfect indicative of the verbs of the first conjugation show great variability among the interviewees, in parallel with the development of this morpheme in Lleidatà: 1) [a j e], the least commonly used form, noted in only two interviewees: 2 realisations in a man in the 61 to 70 age group and 1 realisation in a woman in the 51 to 60 age bracket, both with primary education and no knowledge of the written language. 2) [a e] occurring at quite a high level of frequency, in contrast to what was observed in the questionnaire and its closed questions, where it did not occur at all. 3) [ae] alternating with the previous variant ([a e]) with approximately similar rates of occurrence. Lastly, [a], belonging rather to a more formal style of speaking, did not occur either in the interview or in the conversations.

We have seen that linguistic factors themselves are not relevant when it comes to explaining or characterising variation in the 3rd person ending. If we analyse the incidence of social factors we see once again how the data can be ordered in terms of age, education and knowledge of Catalan. It should be noted, however, that the results did not all reach significance. (15)

In the case of the age factor, we note a higher incidence of [a e] than [a B e] in all age groups except among speakers aged 61 to 70, who use both forms at an approximately equal rates, and those aged 6 to 11, who produce  [ae] some 20 % more than they do [a e]. Speakers aged from 21 to 30 show the highest percentage of [a e]: at 83 % (see figure 3). Detailed analysis of the results show once again that they tie in with the predictions made earlier and concur with the studies by Aniansson (1979) and Portz (1979): the school influences children’s' speech, but subsequently, speakers adapt to the forms of their own native dialect, especially when the formality of the communicative act is not very great.

Figure 3. Realisations of the -va morpheme by age

In the case of educational level, there is increasing use of the [a e] covarying with increased education:

Figure 4. Realisations of the -va morpheme by educational level

In terms of knowledge of (written) Catalan, the variant [a e] occurs 24 % more than [ae] in speakers with notions of the written language:

Figure 5. Realisations of the -va morpheme by knowledge
of written Catalan

4.2 Television data

In samples culled from the Lleida television broadcasts, the use of the [e] and
[] variants are totally conditioned by the situation in which the speakers find themselves. Thus, in the news bulletins, the vowel that coincides with the written form, [], occurs 77 % of the time and [e] the remaining 23%. In complete contrast, in the debates the percentages are reversed: the [e] variant is used 83% of the time as against 17 % for the formal /written []. variant. (All the results reach significant level: F=131469, sig=0.000):

Figure 6. Realisations of the 3 variant on Lleida television