This article is based on a study of morphological variation in the third person verb endings in the Lleida dialect (Lleidatà) of north western Catalan. The study analyses different registers of the spoken language used in the municipal district of Alguaire and in programmes on TeleLleida, the Lleida local television service. The study shows that the morphological variation traditionally associated with lleidatà speech is tending to be lost in formal situations. (1)

Summary

1. Introduction

2. The phenomenon in question

3. Methodology

4. Results
4.1 The Alguaire data
4.1.1 Elicited responses
4.1.2 The interview and conversations
4.2. Television data

5. General observations

6. Bibliography

1. Introduction

The importance of distinguishing linguistic behaviour conditioned by space, time, sociocultural aspects and communicative situation, is a constant in the linguistics literature of the 20th century together with recognition that these variables give rise to both constant and variable phenomena.

The study presented here sets out to analyse patterns of realisation of the third person verb endings in samples of Lleidatà speakers, based on the Labovian segmentation into different speech registers or styles and predicated on the need to approach dialectal studies from a functional angle (Biber, 1994).

The work comprises the following sections:

a) a linguistic and social description of the linguistic phenomena;

b) an overview of the methodology used;

c) a presentation of the results obtained in terms of a range of social and functional factors.

Some observations of a general nature on morphological variation in the third person in the dialect of Lleida (Lleidatà) will provide a lynch pin for the study.

2. The phenomenon in question

In north-western Catalan, the development of the verb inflections of the third person singular imperfect indicative, conditional and present indicative of the first conjugation was as follows: IMPERFECT: -ABAT> [aje], [ae], [ae]; -IBAT> [ie], -EBAT> [ie], [je] // CONDITIONAL: -IAT>[ie] //PRESENT: -AT> [e].

However, in the Lleida dialect of north-western Catalan (Lleidatà), variation in the verb endings of the imperfect indicative of the first conjugation has not followed quite the same path. Thus: 1) [aje] ethis has become sharply recessive, now relegated to the speech of the very elderly, as Solans (1990) and Creus (2002) have pointed out; 2) [ae] [ae] may be beginning to lose ground among the youngest speakers and is used basically in informal situations (see Solans, 1990; Creus, 2002); 3) lastly, [ae], and sometimes [a], are the  forms that dominate in the most formal contexts, and are furthermore the forms that come closest to the written standard (2)

Traditionally, these sets of verb endings were kept apart from other verb endings thanks to the open-close distinction in the vowel (vowel height distinction). Thus, we find: jo cantava [aj], [a], [a] alongside ell/a cantava [aje], [ae], [a], jo corria [i] opposing ell/a corria [ie], and ell treballa [e] vs. Treballa! []. (Gili i Gaya, 1931; Badia i Margarit, 1950; Martí, 1970; Pueyo, 1976; DeCesaris, 1986; Veny, 1982; Solans, 1990; Turull, 1990; Viaplana, 1996; Lloret and Viaplana, 1997). This contrast [/e], distinguished each of the persons of the verb and their endings, something which does not happen in the eastern dialects. (3)

At the present time, however, the endings of the first and third person of the conditional, of the imperfect indicative and the present indicative –in first conjugation verbs– are beginning to lose their distinctiveness: the resulting merger going to []. The impact of this change in progress is still quite slight and is not acknowledged systematically in descriptions of Lleidatà (4)

The causes for this change need to be sought in the written language. Note that in north-eastern Catalan, unstressed a in word-final position is pronounced [] (5) (See Gili i Gaya, 1931; Pueyo, 1976; Veny, 1982; Turull, 1990) the only exceptions being firstly nouns ending in –ista (the phonological distinctions make it possible to distinguish gender: fem. artist[], masc. artist[e]) (6) and secondly, the endings of the third person of the imperfect indicative, the conditional and the present indicative of the first conjugation (the so-called –ar verbs). Consequently, there is an analogy between the latter verb endings and the noun endings in []. The result of this analogy in the case of the verb forms is the initiating of a sound change now in progress, which tends to level phonetically all graphemes of unstressed a in word final position to the sound [].

3. Methodology

When it comes to analysing language production, it is of fundamental importance to observe: 1) the elements that are systematically held constant in the speech of individuals; 2) the differences between different speech events and their outcomes. In the project which concerns us here, we base ourselves, on the one hand, on the Labovian concept of speech styles (see Labov, 1972) (7) and, on the other, on the traditional classification of factors that characterise and delimit registers (8) (see Halliday, 1976; Gregory and Carrol ([1978]1986); Spillner, 1987; Biber, 1994, etc.). With these methodological postulates, we proceed to analyse variation in the third person morpheme in: a) 30 Catalan speakers from the town of Alguaire born between 1932 and 1998 (9); b) broadcasts in north-western Catalan put out by Lleida television, at two levels of formality.

a) Alguaire (10)

The 30 speakers analysed were chosen based on the census data for Alguaire and constitute l.146% of the population born in the municipality or in the surrounding county (11). As the phenomenon in question affects Catalan speakers, I have only selected speakers who have Catalan as their first language and who were born locally. The social variables that I consider in this stratified sample are: age, schooling, knowledge of Catalan, socio-economic status and sex. Segmentation by age of the Alguaire sample gives us eight age groups, as follows: (1) 61 to 70 years -born between 1933 and 1942; (2) 51 to 60 years -born between 1943 and 1952; (3) 41 to 50 years -born between 1953 and 1962; (4) 31 to 40 years -born between 1963 and 1972; (5) 30 to 21 years -born between 1973 and 1982; (6) 12 to 20 years -born between 1983 and 1990; (7) 6 to 11 years -born between 1991 and 1997- and (8) from 4 to 5 years -born between 1998 and 1999. (12) Even though the number of interviewees was the same in each group, the generational groups are not evenly segmented nor are they sized in proportion to the segment of the population they represent. Thus, with regard to the younger groups (those from 4 to 31 years), the criterion for dividing up the population has to do with educational level and the fact that these are the first generations to receive schooling in Catalan. While I am aware that the division or cut-up by ages is quite different to those applied to the middle aged and older age groups, I think this is necessary if we want to observe the effect of Catalan-medium education and how different states of language development are related to the adopting of new linguistic habits and practices.

On the question of data collection, given that the variable under analysis here does not occur at a high rate of frequency, we analysed firstly the responses elicited by a corpus of questions on a variety of variable phenomena in the Lleida dialect. These included 12 verbs representing all conjugations different tenses of which are elicited: the present, imperfect indicative and the conditional. The exact verb forms in question were: canta, (she/he/it sings); canviaria (I/she etc would change); comença (she etc begins); dormia, (I/she etc was sleeping); estudiava (I/she etc was studying); fregava (I/she etc was scrubbing); jugava (I/she etc was playing); naixia (I/she etc was being born); perdria (I/she etc would lose); presumiria (I/she etc would presume); sortia (I/she etc was going out) and temia (I/she etc feared (impf.)

We collected a total of 15 hours of less formal speech from two protocols: a semi-structured Labovian-style interview, plus the conversation generated during the meeting with each informant, following proposals by Briggs (1986) and Hazen (2000).

b) Lleida television broadcasts

The language we analysed from Lleida television exhibits two levels of formality: one where the material is read (news broadcasts) and another consisting of spontaneous debate where neither the degree of preparation nor the news-reader type of formality is evident (televised debates). As a result, the corpus for the present study is divided into two parts:

1) Phonological realisations occurring during the news broadcasts. The oral discourse is in the prepared speech mode (written to be read), with an informative tenor and a formal tone.

2) Broadcast debates, more specifically the programme El despertador. In this programme, four discussants and a moderator comment on the news from day to day. In this case the oral discourse is in the unprepared oral mode, with an interactive tenor and a formal tone.

In all, 22 hours of the programme were recorded between 9th May and 7th June 2001. Out of the total of 22 hours, 12 were from different news broadcasts (morning, midday and night) and 10 were from daily debates in the programme El despertador. The total number of verb forms collected was 318. The variables that were controlledwere situation, social class, and one othersocial variable: the sex of the nterviewees. The statistical processing f data was carried out using theGoldvarb 2001 programme (13) In the cae of this data from the Lleida television  the SPSS programme was also used toobtain an ANOVA of the results in general.

4. Results

4.1. The Alguaire data

4.1.1. Elicited responses

The responses obtained from all interviewees responding to the questionnaire show  in the word-final vowel in the forms of the third person. In the case of the imperfect indicative of first conjugation verbs this is almost always reduced to the forms [ae] and [a]. (14)

Taking the range of data from the statistic analysis as a whole, we see a high probability of general maintenance of the third person [e] (0.819). This overall observation points to the fact that the variation we have detected is in its beginning stages and as such appears in the most formal language and then only to a very cautious extent.

The results of this analysis have shown that in terms of the words included in the study, neither the tense of the verb, the conjugation, number of syllables in the verb nor the environment following the final vowel are significant when seeking explanations for the incoming variation in the third person morpheme.

In terms of social factors, three variables emerge as significant: age, educational level and knowledge of written Catalan. (See table 1). Contrary to what might be expected from the well-established variationist tradition, neither sex nor social status provide us with relevant explanatory data to explain the variation in the final vowel.

Table 1. (Logarithm of similarity: -151,691 [Logarithm of maximum similarity: -125,727]; X² total = 51,927; p= 0.0000).