This article situates historical sociolinguistics as a part of variationism and an integral part of micro-sociolinguistics. It goes on to describe the theoretical principles which provide its basis for the analysis of language change in progress. It constitutes a double alternative: firstly, within historical linguistics as a tool for the study of linguistic change and, secondly, within variationism, as a way of submitting written documents to diachronic analysis. It also reviews the social factors have been investigated by this branch of sociolinguistics, within the small amount of research work that has been published in this area to date.

Summary

1. Sociolinguistics

2. The sociolinguistics of variation

3. Historical sociolinguistics
3.1 Analysis of linguistic change
3.2 Analysis of the written code
3.3 Social factors
3.3.1 The diatopic factor
3.3.2 The diachronic factor
3.3.3 The diastratic factor
3.3.4 The diaphasic factor

4. Conclusion

5. Bibliography

1. Sociolinguistics

The interrelatedness of language and society has been evident throughout the history of the different disciplines of linguistics. But it was not until around the middle of the 20th century that sociolinguistics accepted this relationship as axiomatic to its approach and established the analysis of these two elements as intrinsically involved in a systematic way. There are more than a few examples in linguistics which ratify this link between language and social environment. Witness to this, for example, are the studies carried out by Ancient Greeks, by the German philosophers of language Herder and Humboldt, or indeed all the historical grammatical works of the 19th and 20^th centuries –which necessarily had to make reference to factors external to language. And lastly, there are the notions put forward by Saussure to differentiate internal and external linguistics, and the works of the semiologists Sapir and Whorf, the authors of the theory of linguistic relativity. Nor should we forget the work done in dialectology, the majority of which include a reference to the socio-historical framework of the geographical area in question.

Nonetheless, however, other disciplines within linguistics have sought to remain aloof from this interrelation, and this is true for example of structuralism and transformational generative linguistics These disciplines approached language as a homogenous and invariant system, thus achieving a totally abstract notion of language, disconnected from its environment. In opposition to this idea, sociolinguistics came into being, a multidisciplinary field of study that looked at the use made of language in a specific community.

Undoubtedly, sociolinguistics has emerged with the support of other disciplines, the ethnography of communication, which analyses language as an element inseparable from its cultural context, and of course sociology. And this same viewpoint enabled William Labov to regard the discipline as divided into two different facets: sociolinguistics in the broader sense and sociolinguistics strictly defined. In line with this approach, and within this referential framework, we can also distinguish macro-sociolinguistics and microsociolinguistics. In the first we would include sociolinguistics and areas of the above-mentioned broader sociolinguistics, and remark that it is concerned with analysis of the language as a projection of the social sciences, sociology, anthropology, etc. The second would include sociolinguistics strictly defined, together with the ethnography of communication, and be concerned with the study of the linguistic fact.

2. The sociolinguistics of variation

Within microsociolinguistics we include variation sociolinguistics which studies the correlation between linguistic variation and social variation. The great contribution of variationism was to turn the point of interest around to focus on speech and external linguistics, safe in the knowledge that what was being analysed here was not the ideal, utopian speaker-listener, proposed by generativism. Instead, it sought to see within language the system that was shaped by society. Variation exists from the moment that one accepts the influence of the social context –which, obviously, is also variable– upon language. This is sufficiently ductile to accept modifications induced by the social context. From this new theoretical perspective, therefore, languge is defined as an orderly, heterogeneous and dynamic system (Weinreich, Labov and Herzog, 1968).

We know for sure that free variation, as proposed by generativists, does not exist. Rather, variation responds to correlative social patterns conditioned in turn by such external factors as geography, which produces diatopic variation; by temporal factors, responsible for diachronic variation; by social factors, which generate diastratic (stratified) variation, and lastly contextual factors, producing diaphasic variation. Having accepted this variation, therefore, we must necessarily take it that variation cannot ever be considered random, but rather operates in a totally systematic way.

On the other hand, the study of hetereogeneity of language is not only done on the basis of qualitative analysis but is complemented by quantitative analysis. For that reason, this discipline has provided an empirical model, adopting the same approach as other social sciences. This analytical model which adheres to that of the surveys questionnaires of sociology and of statistical mathematics, is based on the analysing and processing of linguistic and social factors by means of a computer package known as VARBRUL. The latter provides the indices of probability of different variants occurring in a context, by means of logarithmic calculation.

3. Historical sociolinguistics

3.1 Analysis of linguistic change

The historical perspective within linguistic research becomes especially notable from the mid 19th century onward. Using the comparison of languages to work out genetic relationships and the subsequent formulation of language families led on to a historicist perspective and the development of historical grammars. From then onwards, the principle of the evolution of languages become a clear axiom informing much historical language research. What's more, this epistemological agreement within linguistics led F. de Saussure to set up the famous dichotomy between the synchronic and the diachronic. In principle the fact that one could now differentiate studies with synchronic paradigms from those with diachronic ones was in itself an important step above all toward recognising the dynamic nature of linguistic systems. Nonetheless, structuralism and generativism were to develop solely the synchronic facet, and indeed were not able to overcome the Saussurian dichotomy. This was because they were not able to handle the temporal correlation of discrete linguistic states without ignoring the intermediate states in their entirety. It was within this conceptual approach that the American structuralist Bloomflield affirmed in 1933 that it was not possible to study language change except by analysing related languages or comparing different states of a language.

And it is here that the theoretical corpus of historical sociolinguistics offers a more convincing alternative for the analysis of language change. As we stated earlier, sociolinguistic theory sets out from the premise that language is an ordered system which is heterogeneous and variable. This premise, which accepts the coexistence of heterogeneity and variability in a language, on the synchronic level, constituted a new way of looking at historical linguistics, thus enabling us to obtain a diachronic perspective from within a synchronic study.

Work by Labov (1966), and later by Weinreich, Labov and Herzog (1968), demonstrated it was possible to carry out diachronic research within a synchronic framework. By looking at intergenerational differences among a given sample it would be possible to see, on the one hand, the existence of variability –and thus of language change– and, on the other, the direction of this change based on the variability produced by the oldest generation compared to the youngest. These new theories have given rise to the concepts, on the one hand, of apparent time, which involves the synchronic approach but combined with the use of data from different social strata and age groups. Here, we first have to take into account the fact that the linguistic system learned in adolescence remains more or less stable for the rest of one's life. Were this not so, we could not usefully study the differing use of variables made by people aged 60, 40 or 20, for example, to be able to determine the variation that was used, is used and (probably) will be used. On the other hand, the other possibility is to study change in real time, that is, the diachronic study of language change by means of analysis of the speech of individuals at different times in history.

In this way, variation and language change become two parallel phenomena, since one and the other concept constitute two sides of the same linguistic process. Even so, it is important to note that where we see variation we cannot always assume that this is linguistic change, given that within synchronic variation we have to distinguish between stable and unstable variables. On the other hand, whenever we find language change it is because we have language variation. It is only in this second instance, then, that we can argue that variation –in this case unstable variation– means linguistic change is in progress.

Basing itself on these principles of language change, historical sociolinguistics is able to approach the language of the past from two perspectives. If it assumes a synchronic perspective, it will observe variation as a possible model of language change, previously checking on whether the variables are stable or not. If, on the other hand, it approaches the variation from a diachronic point of view, both variation and change can be investigated thanks to the unstable variables recorded in the different synchronic sections. Developing this line of thinking, Labov concluded that the analysis of the current linguistic situation could serve to explain what was the situation in the past -and vice versa, the linguistic situation in the past could help us to understand the current one.

3.2 Analysis of the written code

The majority of studies on the sociolinguistics of variation use oral speech data. With this limitation, linguistic analysis has always to be synchronic (even taking advantage of apparent-time data), since if we wish to go back in time, we cannot go further back than the 19th century, when we start to get phonetic recording. Thus, another possibility is the methodology offered by historical sociolinguistics, which offers the alternative of working with historical documentation, that is, with written texts. This option involves the inconvenience of having to vary some of the analytical co-ordinates as compared to oral texts but, on the other hand, we will enjoy the advantage of being able to shift our gaze to any stage in the history of the language.

Within this dichotomy, between oral and written language, we have to state, as an essential axiom, that language is basically oral production. Once this principle is accepted, therefore, we have to recognise the inconvenience of studying language at this level alone, since (as we have said) oral language can only be studied by means of synchronic studies of more or less contemporary language. Conscious, then, of this disjunction between the two codes, we also have to bear in mind the nature of written language. As Martí (1989: 21) puts it: "I do not think it is right not to accept that written production, irrespective of its type, simply by virtue of being written, involves a special tension, out of which comes production which may be more or less close to oral expression".

The diaphasic difference, then, between written and spoken language is obvious, but we should add that in written production it is perfectly possible to distinguish greater or lesser levels of formality depending on document type, such that lesser formality will tend bring us close to the spoken language. Having said that, we should recognise that in any analysis of historical linguistics we will have to remember that written expression has certain features that we need to be aware of when drawing conclusions from a piece of research. These can be summarised succinctly under two headings:

a) Firstly, while observation of frequency of occurrence of given variants is central to the research methodology for sociolinguistic investigation of speech data, we should be aware that written data, in contrast, tends to reduce considerably the amount or frequency of variants that occur. That is, to put a brake on their occurrence, as a result, obviously enough, of the conservative characteristics that the written medium assumes. Thus, we can conclude that written language presents a certain resistance to the variationist computation in any study.

b) Secondly, written language always poses a problem for the dating of any linguistic phenomenon, since noting the appearance of a given form in documents does not mean that it can be situated chronologically by the date of the document in question –even if we know that this is the first known occurrence– because, owing to the conservative nature of the written medium, already discussed above, any linguistic form appearing in writing can be assumed to have been in use for some time in the spoken language.

3.3 Social factors

As we stated earlier, the analytical paradigm used by variation sociolinguistics in synchronic studies of oral language is entirely applicable to historical sociolinguistic studies. Remember that the factors influencing variability have always been the same and, as such, they are available to the methodology of this discipline, too. That having been said, it is also important to note that geographic, temporal, social and contextual factors have not always been exploited to the same extent in the few diachronic sociolinguistic studies that have seen the light until now.

3.3.1 The diatopic factor

The distinctive origin of the historical document can help us to analyse this diatopic factor. In reality, what we are interested in here is the origin of the author of the document in question -since it will be a dated document from a particular place; but where the author is from outside the area, with dialectal features not typical of the place the document was produced, analysis of this factor will not be possible. By the same token, the analysis could be valid if the author were from outside the area but shared the dialect features of the place where the document was produced.