The Audible Corpus of Spoken Ruaral Spanish (COSER) is a dialectal corpus based on interviews with informants who have been object of interest in the traditional dialectology: rural native speakers, elderly and with a low education degree. So far, 1,497 informants have been recorded, with the following distribution by sex:
Males 662 (44.2%) Females 835 (55.7%) Total: 1.497
The informants' global average age is 72.9 years old. the COSER deals with a survey oriented towards informants, who have been born in the first third of the 20th century, and who have not received much instruction. On the whole, they have attended some years of elementary school learning, according to their declarations, "to read and write, and four more rules [on elementary mathematics]".
The recordings within the COSER have been regularly obtained since 1990 up to recently in a series of surveys campaigns. This fieldwork has been organized by the support of several projects of research and as part of the field practices of the subjects "Hispanic Dialectology" (years 1988-1996) and "The Spoken Spanish. Peninsular Variants" (1996-2011), optional matters belonging to the Degree on Hispanic Philology of the Autonomous University of Madrid (UAM).
Until now, 801 rural places of the center and north of the Iberian Peninsula have been interviewed. The final objective is to obtain recordings of Spanish, spoken in rural areas of the whole Iberian Peninsula. The geographical zone of the localities surveyed so far appears in the map:
Map: Localities surveyed in COSER (2011)
The audible materials include, for the time being, the central band of the Iberian Peninsula. Besides, the density of the network of points is comparable to that of the regional atlases or, even, thicker.
In general, the COSER has nowadays circa 1,000 recording hours. All of them has been digitalized. Some have been also transcripted as text files, thanks to the support obtained by several research projects and the participation of numerous generations of students of the degree in the UAM, who have transcripted recordings that they collected themselves, as a part of their works in the academic course. These recordings and transcriptions are available at www.uam.es/coser.
The methodology used in the COSER belongs to that of the sociolinguistic interview, whose topics deal with the traditional life in the field. The fact that the interview focuses on these thematic modules does not prevent the interest to change towards others, such as the education, the desires, the personal experiences, the own life or the family, depending on the degree of comfort and spontaneity shown by the informant. The decision to focus the interview on thematic modules related to the rural life "from the past" has to do with the fact that, in order to accept the interview, the potential informant has to believe to have certain knowledge on a lifestyle already passed. This awareness is the result of his personal experience and of his age, and that knowledge grants the interviewed informative "authority" before the urban interviewer. The informant accepts the interview encouraged by our interest for the testimony about a lifestyle in decadence, of which the interviewed is known as an expert. We think that the spontaneous collaboration of the informant would be much more difficult to manage, if he/she were asked at the beginning about opinions or personal experiences, or on questions foreign to the rural life. In fact, in not a few occasions the informants have mentioned the condition of the university students to try to elude the interview, invoking "if you already know everything better than me!". The insistence of the team towards their interest in the strictly local traditions, in contrast with that of other rural places, has been in many occasions determinant so that he/she accepted the interview.
The contact with the informants realizes always at random, without previous negotiations, between the local individuals that the requirements assemble before mentioned. The experience few pleasing of some interviews owed to the scanty informative capacity of the informant (people of little to speak, that it answers with very short phrases or monosyllables) did that was added later the condition of talkativeness ("that she should like to speak") to the protocol of selection of the informant. Since it must turn out to be evident for anyone that has realized fieldwork, the success is never assured, and an interview can be ideal or terrible with the same conditions of item. This way, not all the interviews are neither equally suitable nor informative, depending on the disposition of the informant, on the skill of the interviewers and of the interaction among both, but not for it we believe that no testimony should be despised.
With regard to the number of informants of every place, in the COSER has preferred interviewing in general in depth an alone person, be man or woman. Nevertheless, the conditions of the recording prevent from avoiding, sometimes, the interruption on the part of other individuals (in general, members of the family or acquaintances who, attracted by the extraordinary circumstance of the interview, feel tempted to intervene and give his testimony). For it, though in the COSER they are registered up to 1.497 informants, the immense majority of the times only has been polled one by enclave by the desirable thoroughness (approximately the half).
In a beginning, the protocol of the interview was designed in order to document certain linguistic phenomena (in I make concrete, the employment of the unaccented pronouns), but immediately there was revealed that always there was obtained information about other many aspects apart from the waited. From this moment the development of the conversation has looked for the creation of contexts that were favoring the appearance of dialectal information of all kinds, some alluded in the specializing bibliography and others that scarcely had called the attention. In this respect, it is necessary to highlight that the interviews of the COSER have proved to be specially useful to document dialectal phenomena relative to the grammar, aspect this one traditionally little represented in the dialectal monographs and in the questionnaires of the linguistic atlas.
The average duration of the recordings is of an hour and quarter (75 min.) for place, but it can range from only half an hour even more than two hours and a half. The quality of the recordings is not directly proportional to the duration, since there exist excellent and very informative recordings of only half an hour, which results are comparable to the obtained ones in a more long session. Otra traducción: The methodology used in COSER has consisted in sociolinguistic interviews, aimed by part of the interviewers at some subjects of traditional country life. The fact that the interview is focussed on such specific subjects does not prevent that, after some time and having gained the informant’s confidence, interest is aimed at other subjects, such as education, personal hopes and experiences, life or family, depending on the level of easiness and spontaneity shown by the informant. The decision of focusing the interview on specific subjects related to rural life “of former times” has much to do with the fact that, in order to accept to be interviewed, potential informants must prove to have some knowledge about a way of life in decline. This knowledge is a product of their own personal experience and age and gives them informative "authority" in front of the urban interviewer. Informants accept the interview as they realize that we are interested in a testimony on a way of life in decline about which very few have hardly any memory at all and which they know they are expert on. We think that the informants’ spontaneous cooperation would be much more difficult if they would be required at first to be interviewed on personal views or experiences, linguistic matters or other aspects beyond rural life. The fact that the interviewing team has insisted on their specific interest in the strictly local tradition, in contrast to that of other rural enclaves, as well as in the exclusive informant’s condition as recipient of such tradition, has been on many occasions a decisive factor for accepting the interview. Informants are always randomly contacted, with no previous actions, among the local inhabitants fulfilling the above mentioned requirements. Due to the experience, not much gratifying, of some interviews on account of the informants’ low communication ability (people not much willing to speak, who answered with very short sentences or just in monosyllables) led us to add subsequently the condition of loquacity (“that the informants like talking”) to the informants’ selection protocol. As it will be obviously well-known to anyone who has ever carried out fieldwork, success is never assured, and an interview starting under the same conditions may be optimum or dreadful. Thus, not all interviews are equally suitable or informative, depending on the informants’ willingness, the interviewers’ skills as well as the interaction between them; however, no testimony should be disregarded for that reason. This methodology can not avoid the problem of accommodation between the informant and the interviewer, or the challenging representativeness of the informant ramdomly chosen. Nevertheless, we think that the quantity of the data allows to circumvent these potential problems, since the data always show geographical coherence and make it possible to discard those informants who could be considered anomalous with their area. Regarding the number of informants of each enclave, in general one single person has preferably been thoroughly interviewed in COSER, either a man or a woman. Nevertheless, recording conditions have sometimes not allowed to avoid interruptions from other individuals (generally members of the family or acquaintances who, drawn by such an extraordinary event as the interview, cannot resist the temptation to take part in the interview by giving their own testimony.) Thus, although up to 1,408 informants have been recorded in COSER, most of the times only one informant per enclave has actually been thoroughly surveyed as desired (almost the half). The average duration of the recordings is one hour and fifteen minutes (75 minutes) per enclave, although it may range from just half an hour up to more than two hours and a half. The quality of the data recorded is not directly proportional to the duration, since there are excellent and very informative recordings of just half an hour, whose results are comparable to those obtained in a longer session.
III. Utility, innovation and limitations of the COSER COSER is a corpus aimed to measure the differences which may be found in the speech of sociocultural groups with a lower education in rural areas. It is therefore a complement to both linguistic atlases and to the different corpora of cultivated and urban speech which have been compiled or are planned to be so in the Spanish-speaking world. The uniformity in the methodology used makes it useful to measure both the linguistic distance which separates different areas (physical distance) and the linguistic distance which separates this social group from others, like for instance, that of speakers with a higher sociocultural level or that of younger speakers (social distance). Although the proportion of men and women interviewed is not identical (55,7 % women vs. 44,2% men), the number of speakers of each gender is statistically representative and also allows to investigate linguistic differences associated with gender. The fact that the media are the sources of most Spanish oral corpora lends some singularity to the COSER, since the interviewed speakers for COSER are rarely recorded in this field. The comparison between the data obtained in COSER and in other corpora of spoken Spanish enables thus to point out clear sociocultural differences. In this regard, COSER has proved especially useful since it provides the study of non-standard grammatical solutions, which are usually systematically avoided in written language and in the speech of sociocultural groups of higher education. For that reason, Chambers (1995) has proposed, as a sociolinguistic universal, the qualitative character (presence/absence) of grammatical variables in the social scale, in contrast to the quantitative character of phonetic variables.
IV. Researched topics and publications Los materiales del COSER han hecho possible investigar un conjunto de aspectos de variación grammatical del español y han dado lugar las publicaciones siguientes: (traducer)
(aquí copiar lo que aparece en la página de publicaciones del COSER, de momento, a falta de completarlo).