To achieve the objectives outlined above, various research instruments are needed, including:
a) audio tape recorder
b) video tape recorder
If our aim were to provide a comprehensive description of language use in Budapest, then our data collection would have to be representative in a) sociological and b) linguistic terms. We know from Hungarian sociologists that sociological representativeness would call for the use of 200-300 informants. What we don't know, however, is how much material should be recorded and in what communicative situations from a single informant in order to enable us to make inferences for the entirety of their language system with only a minimal margin of error.
It follows from the above that the Budapest survey can only aim at sociological representativeness at best.
We cannot strive for a comprehensive analysis of individual language features such as for example, the so-called ``-ik '' verbal conjugation. (Hungarian verbs fall into two categories: -ik and non-ik verbs, e.g. alsz -ik `he sleeps' vs. áll -0 `he stands'. The two types of verbs used to be conjugated quite differently but today the differences are disappearing rapidly. The -ik or non-ik issue is a stylistic one today but speakers are sensitive to it, and this or that usage often invites a good deal of comments.) The Standard Prescriptive Grammar (=Nyelvmuvelo Kézikönyv 1980-1985; henceforth abbreviated as SPG) sorts these verbs into six groups ``from a descriptive point of view'', on the basis of their use. However, one finds variation not only in verbs belonging to the so-called ``fluctuating -ik '' and ``fluctuating pseudo -ik '' category but in those belonging to the ``stable -ik '' category as well. It is easy to realize that a comprehensive analysis of the issue would call for hundreds and thousands of relevant examples from a single informant in a great number of communicative situations ranging from careful formal style to the most casual relaxed speech situations. It is clearly beyond the means of our project to undertake such a detailed analysis, however justified and desirable it may be to compile such a close-up view of the use of these verbs. What the project can undertake is to provide solid empirical evidence to answer certain selected questions, including, for example, about the -ik conjugation. Our investigations will not yield a definitive answer to the question whether the classification in SPG is descriptively sound, whether the particular verbs are correctly assigned to their category in SPG etc. However, we will be able to answer the following questions:
Certain research questions of the interview will be set out below. The selection of these questions was made in the following way:
contains the interview itself in the following arrangement: the text of the three sections of the interview (phonology; morphology, syntax and lexicon; guided conversation) will be followed by a discussion of the research tools of the relevant sections together with the instructions to the field workers. Problems of transcription, coding and analysis are dealt with in the final chapter.
In Autumn 1987 Version Two of the Budapest Sociolinguistic Interview was used with a quota sample: 10 teachers of over 50 years of age, 10 university students, 10 sales clerks, 10 blue-collar workers and 10 vocational trainees (around age 16) were interviewed.
Version Three in 1988 was used with a stratified random sample, the (200 person) Budapest subsample of a national representative sample comprising 1000 persons. This sample has been used by Róbert Angelusz and Róbert Tardos to record more than a thousand sociological questions for their project ``Social stratification - communicative stratification''. In May 1988 the 1000 person sample was administered a linguistic questionnaire as a complement to the Budapest Sociolinguistic Interview (see Kontra 1995:10-11).
Therefore, it can be claimed that the sample used in 1988 meets any sociological standard. As an extreme example, should there be any correlation between a person's use of a linguistic variable and the date when their home was last remodelled, we would be able to precisely state that correlation.
Basic sociological data recorded about the informants include: age, sex, occupation, education, race (Gipsy or not), birthplace, in-migrant or not, teacher or not, mother tongue, time abroad, knowledge of foreign languages.
The Survey hired field workers who are not trained linguists. The field workers are experienced interviewers in sociological research but had to be trained by us to do this linguistic interview. This is why some of the things we say here may seem to the professional linguist too obvious to mention.
The interview must be started with every informant and should be taken up to `break'-point. If the informant turns out to be illiterate, then
If an informant should be unable even to read the letters A and K (for `same' and `different') or the figures `1' and `2', then the test must be done orally. In other words, the informant must be trained to give oral answers, the trial tests should be administered and when the test seems to be safely administrable, it should be started. The informant's microphone should record the stimulus on the test tape together with the informant's answers. The final result will be something like this:
test tape: ``one: gyanu - gyanú ('suspicion')''
informant: ``they are the same''
In doing the oral sentence completion tests, the informant's attention mostly focusses on the word to be inserted and the suffix to be used. It can be presumed that when reading the full sentence out the informant pays less attention to the rest of the sentence. Therefore, we will call the item that the card is explicitely focussed on 'primary data', the rest secondary. For example,
Ebben a ..... nem mehetsz színházba.FARMER `jeans'
`In-this the ..... not you-may-go to-theatre.'
The expected answer is farmer-ban/ben . Here the primary data are the vowel of the suffix and the presence or omission of the word-final nasal. On the other hand, presence or absence of the word-final nasal in Ebben as well as the short or long `i ' of színházba are considered secondary data.
The field worker should have a clear view of the location of secondary data as well although they are not explicitely marked on his/her cards.
In processing the results, primary and secondary data should be handled strictly separately. After the transcription and coding of the data is completed, a separate study must be made to establish whether the distinction introduced in 1988 as a working hypothesis was indeed justified, in other words, whether informants' responses are significantly different as a result of the amount of attention devoted to the linguistic task. (For a preliminary answer to this question see Váradi 1995/1996.)
Kontra, Miklós. 1995. On current research into spoken Hungarian.International Journal of the Sociology of Language 111:9-20.
Váradi, Tamás. 1995/1996. Stylistic variation and the (bVn) variable in the Budapest Sociolinguistic Interview. Acta Linguistica Hungarica 43:295-309.