Given the rigidity with which certain writers have interpreted this taxonomy, it should be remembered that it is the basis of a number of controversies, either over the cataloguing of the research areas and their implicit aim, or arising out of domestic priorities when defining the scope of the discipline as a whole. In the ambit of Catalan sociolinguistics, this second type of issue has recently provoked discussion which I would like to stop and look at in more detail. It has been said many times that our tradition was -and is- essentially sociological. This choice of direction is entirely understandable, given that in a situation of language contact in which the very future of the language is under debate, the majority of researchers have concentrated their intellectual efforts on guaranteeing a comfortable communicative ecosystem for the Catalan language. Nonetheless, macrosociolinguistics has also had the opportunity of making its mark, and has brought about a good list of work, which though slight in comparative terms, is far from negligible.

A. Mas and B. Montoya, in an article (in press) on the state of the question of variation sociolinguistics (that is, taking stock up until the present) in the area of Catalan, the writers express regret -in my opinion with every justification- over the carping criticisms made of variationism by adherents of the self-styled "soberania sociolingüística" (sociolinguistic sovereinty) or "sociolingüística del conflicte" (the sociolinguistics of conflict) (Ruiz, Sanz and Solé 2001: 252, 257-258). The line of argument designed to minimise the contribution of sociolinguistics centres on, on the one hand, the negation of the social implications of the phenomena that it studies, and on the other the alleged total lack of "emancipatory usefulness" of an essentially erudite scientific model.

Taking quite the contrary view, Mas and Montoya conclude: "We have no doubt that the three types or tendencies we have observed in linguistic change under way in different parts of the Catalan-speaking Countries will, apart from constituting a map of speakers' language loyalty, prove useful information for centres of language planning. In synthesis, these three types of change are: genuinely Catalan changes, changes in the direction of Spanish and "atrophy"). This will be true of both the corpus and status information, in orientating work on the normalisation of the Catalan language. As I see it, what we have above is a restricted interpretation of the multifaceted interests of sociolinguistics. Interests which are constantly intertwined, with totally permeable boundaries to the disciplines, and which are striving to consolidate, by means of solid epistemological debate and a constant drip, drip of research, the much sought after paradigmatic status.

3. The study of linguistic variation

"[...] linguistics has striven to establish abstract, model units, to provide solid ground in the instability of the unvarnished facts of variation, which dialectology and sociolinguistics have brought to the fore. However, it is the field of pragmatics and discourse analysis where variation, without any doubt, takes centre stage. Indeed, the entrance on the scene of pragmatics and both discourse and conversation analysis has raised a series of new issues, arising in the main from the pressing need to incorporate factors of context and use into the proposed models - factors which it is essential to introduce. These amount to a large number of variables that need to be carefully managed to accurately reflect a particularly complex dimension, additional to the other difficulties." Salvador (1997: 204)

3.1 A multidimensional approach

The central objective of this section and its title is to take stock of the forms taken by sociolinguistic variation studies here in Catalonia. According to the scheme arrived at as if by consensus, the current within sociolinguistics most frequently adopted was that of strict sociolinguistics or, as it has come to be called in our nomenclature here, the sociolinguistics of variation. (5) The question, then, was to evaluate the most significant "deviations" from the reference model, either through pressure from more traditional disciplines or thanks to the infllux of new approaches and new issues.

The type of approach that concerns me here is to be found in Moreno (1990) and its flawless methodology. Moreno defines linguistic fact as multi-dimensional in character and places its study in the contextual framework of the linguistics of speaking/acting. Its multidimensional nature consists of four vertices occupied by geolinguistics, historical linguistics, pragmatics and sociolinguistics -with intermediate positions being possible. This proposal recognises that variation brings together diverse factors, but allows for the possibility of studying each of these factors separately, minimising the rest.

In the subsections which follow, I shall set out to make an initial attempt to reduce to order the variationist kaleidoscope of research on variable phenomena in Catalan. Let me say at the outset that this will not be an overview of research as such, with subsequent compartmentalisation. My intention here is to provide a qualitative (and initial) evaluation of theory-cum-methodology of the writers of the articles that appear in this monographic issue of Noves SL. Not surprisingly, the selection reflects the multidisciplinary hypothesis.

3.2 The sociolinguistic vertex

In the ambit of sociolinguistics strictly defined, a number of different models have been developed designed to provide an explanation for systematic variation. Two of the most important are the dynamic (C. J. Bailey, D. Bickerton and D. DeCamp) and the quantitative (W. Labov). At the same time, the fact that the majority of our researchers opt for the second option -also given the name urban dialectology, social dialectology, sociolinguistics ("strict defined", "properly speaking"), quantitative research on speech, the variable rule model, correlational (socio)linguistics, variationism and the sociolinguistics of variation- explains why I will be devoting my attention exclusively to this branch. This is all the more justified if we consider that this branch, methodologically speaking is thepoint of departure for transferences and transactions as announced in the title of the article.

William Labov's idea of doing sociolinguistics can be seen as the most successful development to have come out of the article by Weinreich, Labov and Herzog (1968) "Empirical Foundations for a Theory of Language Change". His work constitutes nothing less than a quantum leap in the social understanding of language.

At the beginning, Labov drew on the structuralist model defended by his teacher Uriel Weinreich. Nonetheless, at the end of the sixties he adopted aspects of the generativist approach, even as he proposed an extension which would incorporate a social component. As we have already said, Labov's approach was essentially linguistic: to study language as used by speakers in everyday communication, that is, to study its structure and evolution in the social context formed by the speech community. From this realist perspective, the object of study would be observed in all its diversity.

The vision of language as an orderly heterogeneous and dynamic system is achieved by correlating linguistic, social and stylistic variables. And the methodological instruments developed to provide an empirical model are the sociolinguistic interview and statistical quantification, first in terms of straight frequencies, and then later on with the contribution of the Canadian school (Sankoff, Cedergren and colleagues), probability values. This last development is in no wise trivial, since it is based on the epistemologically essential idea that the performance data is merely a statistical reflex of competence, which could now be quantified by the application of a multiple logistic model, the latter making it possible to convert observed frequencies into theoretical probabilities. This processing meant, in short, going from treating performance data as rigorously as possible, to developing theoretical models that take account of speakers' sociolinguistic competence. (6)

The sociolinguistics of variation arrived in Catalan speaking regions (and in Spain as a whole) for the first time thanks to the Alacant-based researcher F. Gimeno. As a teacher at the Universitat d’Alacant, the first result of his influence was Montoya (1985), a doctoral thesis that approached the study of linguistic change both from the synchronic and diachronic standpoint (historical linguistics). At the same time, Maria-Teresa Turell, first at the Universitat de Barcelona and then at the Universitat Rovira i Virgili, began to bring together a group of collaborators who would focus their work essentially on the synchronic aspects of changes in progress. The publication in 1995 of her volume La sociolingüística de la variació (The sociolinguistics of variation) would make available a part of the research that was carried out.

Despite the interest here at home in historical sociolinguistics, quantitatively insubstantial but qualitatively of exceptional interest, enthusiastic reception of the Labovian models meant the adoption of a clearly synchronistic approach inspired by Labov. The paper read in 1972 at the XI International Linguistics Conference held in Bologna, "On the use of the present to explain the past" (Labov 1974), or the work presented in Montreal at the conference on New Ways in Analysing Variation in English (NWAVE) entitled "What can be learned about change in progress from synchronic descriptions?" constitute two good doctrinal samples. Thus centred on this approach, and with numerous small methodological differences in detail (see Pradilla 2002 for the phonic variables and Mas and Montoya –in press - for an overview of all levels of analysis), the "orthodox" variationist approach is accumulating a good long list of successful research projects carried out. And looking at the contributors to this monographic issue of Noves SL, it is worthwhile noting that Josefina Carrera, who read her doctoral thesis in 1999 (Carrera 2002) is also present.

The line of investigation taken by B. Montoya deserves a separate mention, apart from his work on the breakdown of intergenerational transmission of the Catalan language in Alacant (Montoya 1996), where, from the variationist perspective, he describes the phenomenon known as linguistic atrophy or linguistic shrinkage. This, then, is a study of structural attrition or disintegration of the recessive language (Catalan), the latter in the process of extinction, or of convergence with the expanding language (Spanish) (Montoya 2000).

Lastly, I would like to bear witness to the growing mutual interest that is occurring in the methodologies of variation sociolinguistics and experimental phonology. At the 15^th International Congress of Phonetic Sciences, held in Barcelona in August 2003, it was evident that there were a good number of studies that come under the label Sociophonetics. Since Labov (1966, 1972) established correlations between voice quality and social class of the speakers in his samples, using technical instruments, linguists working in the same tradition have produced a steady flow of work, slight at first but more prolific over the years. And, at the same time, the results obtained have obliged phoneticians to rethink some theoretical-methodological questions, having to do with socio-stylistic effects and drastically underresearched up to that point. In the area of the Catalan language, the present writer has repeatedly underlined the need to work towards a common vision of methodology shared with those working in experimental methodology (Pradilla 1997, 2001a, 2001b and 2002). from the beginnings of his work with the quantitative treatment of prepalatal intervocalic disaffricativisation, the variable looked at in his doctoral thesis – (Pradilla 1993). Also, with the prioritisation of acoustic analysis, the recent work by Planes and Pradilla (in press) and by Carrera, Planes and Pradilla (in press) also corroborate the influence of variationist postulates.

3.3 The pragmatic vertex

As we have just seen, the variables studied by sociolinguistics are not discrete, but defined in terms of statistical correlation. Note that, while language use can be discrete, and language choice is one of the most cited examples, yet what also should be born in mind are the different phenomena that emerge with contact and interference, codeswitching, or the emerging of interlanguages (pigdinisation and creolisation). Pragmatics, seen as the suiting of language use to the sociocultural context, will concern itself with these questions.

At the heart of a terminological fog, there emerges, as an alternative to variationism and with a diaphanous connection with pragmatics, the ethnography of communication (Dell.H. Hymes, and John J. Gumperz). Starting out with an anthropological orientation, it conceives the study of linguistic variation in a community of human beings, "as the analysis of how its members manage the phenomenon heteroglossia in their lives, while based on an active and dialoguing concept of human communication" (Argenter 1998: 16).

By utilising an architecture of units of analysis to form a dimension of more concrete types– the communicative act, speech event and speech act- the ethnographer's task will be to correlate patterns of language use with the sociocultural context of a speech community. This, then, will involve the identification of current speech acts, the speech styles associated with that and the sociocommunicative contexts that they evoke.

Having completed this first round of characterisation, we turn our attention now to points of conflict with the sociolinguistics of variation: a) while the latter is interested primarily with spontaneous speech (the vernacular), the former the ethnography of speaking) gives priority to the formal end of the stylistic contínuum; b) while the Labovian procedure of compartmentalising the stylistic continuum establishes a priori segmentation based on the degree of attention paid by a speaker to their speech, ethnographers on the other hand search for natural language varieties in given contexts, that is, ways and forms of speech which are often ritualised, marked and laced with aphorisms; and c) the gathering of data and information for studies of the ethnography of speaking are made by prioritising qualitative techniques, such as participant techniques, while variationism uses techniques which are arguably less natural such as the sociolinguistic interview.