Language Dynamics of an Ethnic Minority Group: Some Methodological Concerns on Data Collection

| January 8, 2014

November 2006. Volume 1 Issue 3

Title
Language Dynamics of an Ethnic Minority Group: Some Methodological Concerns on Data Collection

Author
Dr. Francesco Cavallaro
Nanyang Technological University, Singapore

Bio Data:
Dr. Francesco Cavallaro is an Assistant Professor in Communication Skills at Nanyang Technological University, Singapore. He has worked at various Universities in Australia and he is now living and working in Singapore. His research interests are in sociolinguistics, bilingualism, second language acquisition and in how technology can be integrated into the language classroom.

Article: PDF version

Abstract
This article presents an analysis of the methodology used in sociolinguistic research. Reference is made to a preliminary study on the language dynamics of the Italian community in Australia carried out to identify the methodology necessary to obtain naturalistic data in four different contexts. In the study, four different methods were trialed: diary keeping, participant observation, questionnaire and tape-recording. One of the starting premises of this pilot study was that one method could be chosen to carry out the main study. However, since the data needed to be collected in four different contexts, each with varying degrees of formality, the study identified that certain methods were not compatible for differing situations. This paper points out that the use of one method is not appropriate to encompass all situations; rather, it shows that to obtain a clearer and more complete picture a combination of the different methods may be necessary.

Key words: Sociolinguistics, Methodology, questionnaire, survey, bilingualism, data collection, language maintenance.

Introduction
In any sociolinguistic study, before choosing the methodology one has to consider all the theoretical and practical issues involved in conducting such a study. Only after doing that, considerations can be made about the appropriate methodology. This article, will firstly consider the sociolinguistic research methodologies in the context of the language maintenance issues concerning the Italian community in Australia. It will then present an analysis of four methodologies trialed. Finally, the methodological concerns and considerations in how to obtain naturalistic language data will be discussed.

Linguistic repertoire of Italians in Australia
Any sociolinguistic research in the Italo-Australian community conducted must consider the complex nature of its linguistic repertoire. The simplest way to categorize this repertoire is to say that it may include an Italian dialect, Italian and English. The complexity lies in the fact that the categories of dialect and Italian both represent a continuum of varieties. The dialect continuum can have a local dialect at one end and a regional koiné at the other. The Italian continuum can begin where the regional dialect ends and reach the standard variety. Standard Italian, however, is a literary and written form and never found in spoken language (cf. Alfonzetti, 1998; Berruto, 1987, 1989, Pellegrini, 1960, Sanga 1981, Trumper, 1984, Trumper and Maddalon, 1982).

In Australia the linguistic repertoire of the Italian community is further complicated by the addition of English. However, trilingual people are slowly disappearing. That dialects are not maintained is not too difficult to explain. Even in Italy the use of the dialect is now more and more a function of social groupings and age (Trumper, 1989). The case of standard Italian in Australia is slightly different. Researchers have proposed various reasons for the shift away from Italian. Clyne (1987) from his study of census data posits that the high rate of exogamy in the second generation and the low rate of literacy in the first are two very important factors aiding the shift. Bettoni and Gibbons (1988) suggest that the shift is further aided by the fact that very few first generation Italians speak Standard Italian. They describe the community as being fragmented by the many different regional or popular varieties that are spoken within it. According to Bettoni (1985, 1989) one of the main reasons for its decline is that Italian in Australia is low in importance and prestige when compared to English. This is borne out by attitude studies carried out by Bettoni and Gibbons (1988, 1991) and Callan and Gallois (1982), which show that all Italian varieties in Australia hold low status positions, in comparison to the English language, in many of the traits tested for.

It is also a well documented fact that within two to four generations most Australians of non-English speaking background become not only native speakers of Australian English, but will also have lost their language of origin. (For a detailed analysis of language shift in Australia see Clyne, 1988, 1991, 2003; Clyne and Jaehrling, 1989; and Clyne and Kipp, 1997; for language shift within the Italian community see Bettoni, 1981, 1985, 1986,1989; Rubino and Bettoni 1996; Cavallaro 1998; for an international perspective see Hakuta, 1986 and O’Bryan, Reitz and Kuplowska, 1976). Every new generation born in Australia of Italian descent speaks more and more English, at the expense of the mother tongue of their parents or grandparents.

When second generation children approach school age they can be divided according to the range of their linguistic abilities. These abilities can be defined according to what their mother tongue is:
1) dialect; or
2) popular or regional Italian if

their parents speak different dialects, or
b) if the area they live in has many speakers of different dialects; or
3) Italian1 if the parents have made the effort of teaching them; or
4) any mixture of these languages with the addition of English.

It is also true that, at least for the first-born migrant children, the first extensive active use of English is in the classroom. Importantly, however, for those children who attend Italian classes, going to school also entails their first active exposure to Standard Italian. The effect is that they find themselves in a situation where their “Italian” background offers no guarantee that they will cope with successfully learning Standard Italian, which for them could be a foreign language (Smolicz and Secombe, 1986; Ware, 1981). So, functionally it can be argued that, upon entering the school system, these children are acquiring literacy in two second languages.

Language roles
Research within the Italian community has not been able to come to definite conclusions on how each language is used. This indicates that the situation in Australia is a dynamic one. Language use is not constant and more research is necessary to understand the factors that lead to choosing a particular language.

The most commonly held view is that Italian is reserved mainly for more formal situations and that the dialect is the language of the home domain. There are, however, conflicting views; for example, Bettoni (1989) claims that “in the Italo-Australian community the functions of the three languages are not kept rigidly separate,” while Rubino (1993) argues that there are diglossic relationships between Italian and Sicilian, but not between Sicilian and English. If we accept Bettoni’s view that in Australia the Italian language, or any of its varieties, do not have the stable domains of a classical diglossic situation, we can predict that this lack of stability is the main cause of language shift.

Methodology in sociolinguistic research
Over the years different methods have been employed in sociolinguistic research. Different ways of collecting data have been tried out each with varying degrees of success. Since the basis for this article is the issue of methodological concerns, it is worth looking into how some sociolinguistic studies have been carried out and how certain methods were pioneered.

Poplack (1980) in her study of Puerto Ricans in New York used a member of the Puerto Rican community to collect the data. Using an ‘insider’ does have its advantages in that there are no problems as far as being accepted in a closely-knit community. Thanks to the presence of someone known and trusted by the group the data collected using this method was natural and spontaneous.

Labov (1972a) in his study on phonological variables did not use any ‘insiders’ nor did he try to have himself accepted by any group. What he did was to observe what people were saying in a department store in New York. With this method the researcher himself collects the data by observing, taking notes and, where possible tape-recording. With this method the researcher is able to study a number of linguistic phenomena but is limited by the fact that all observations can only be taken in very short time spans.

Labov (1972a) in his New York study also relied on the interview method. While being aware that this method would elicit ‘careful speech’ he also developed the technique to elicit casual or spontaneous speech. He did this by paying attention to those interactions that occurred outside the interview, such as: greetings, interruptions and if a third person was spoken to. He also believed that introducing certain emotional topics would elicit spontaneous speech, such as: childhood events or hypothetically dangerous situations. The interview method for the collection of spontaneous data has been criticized over the years. Indeed Labov himself did not rely on it any more in his Harlem study (1972b).

Labov et al. (1968), in their study on the use of Black English Vernacular, were the pioneers in multi domain research. In their study they used field workers who were participants in the vernacular culture and language. They also tape-recorded as many situations that were as far removed as possible from interview-type situations, such as: gatherings, parties and even a bus trip. While they could not totally dismiss the observer’s paradox in those situations, they found that, in their case, peer pressure was so strong that nothing inhibited the use of the vernacular.

Blom and Gumperz (1972) and Gal (1979) chose to live within the communities they were studying. While never accepted as members of the communities, they acquired enough familiarity with the people they were studying to elicit natural and spontaneous data. The data collection was done through ‘participant observation’. That is, by following people around or simply being present when these people interacted with each other and with the researchers. These researchers were able to observe, take notes of and, where possible, tape-record the language being used.

Milroy (1987a and 1987b) used a similar method in her Belfast study. The method differed from Gal’s or Blom’s and Gumperz’s in the status she acquired within the group and in the type of data she had access to. She joined the group not with the status of a researcher but as “a friend of a friend”. This gave her the freedom to observe and record both formal and spontaneous data. These last three studies mentioned all have numerous advantages. The quality and the amount of data collected in varying degrees of formality speak for the success researchers have had with them. Milroy herself suggested, however, that there is an incredible emotional strain on the researcher through the bond created with the informants. She also pointed out the large amount of data that was unusable due to poor recording conditions and the many hours of recorded speech that never got to be analyzed because of the lack of time and resources.

In Australia the studies on the Italian situation have used different methods according to the data that was needed. Bettoni (1981) used interviews in her study on English transferences into Italian. Interviews are, in general, formal situations and, therefore, they have proved to be a reliable method to elicit only formal language. Rubino and Bettoni (1990, 1996) employed a questionnaire given to a large number of members of the Italian community. However, questionnaires are restricted to the self-reporting of the participants. As such, the data obtained through their use is liable to be influenced by attitudinal factors. For instance, people tend to over-report the standard language to disguise the use of dialect, which might be seen by some as a sign of ignorance or lack of education. Rubino carried out detailed ethnographic work on data collected within a Sicilian-Australian family (Rubino, 1993, 1996, 2000, 2004). From personal experience, and in the opinion of other researchers, like Rubino and Bettoni (personal communications), Italians, as a group, are aware of the different status of the languages that make up their linguistic repertoire. This means that, when asked, they are able to distinguish between the varieties they use and are able to report reliably on what language they are using (although they are still susceptible to the over-reporting described above). One other problem, however, might arise when the participants would be asked to distinguish between two very similar varieties (for example, Regional Sicilian and Regional Italian).

Choosing a methodology
A pilot study was carried out to evaluate the methodologies best suited for exploring the linguistic diversity and dynamics of the Italo-Australian community. The study was aimed at a middle ground between Rubino’s ethnographic work (1993, 2004) and Rubino and Bettoni’s (1990, 1996) statistical work on data collected with questionnaires widely distributed across the community. Taking all of this in consideration the study intended to analyze how individuals use their linguistic repertoire, that is, which of the three languages is actually used in different domains, and what are the motivating factors that lead to the choice of a particular language.

In trying to choose the appropriate method for my studies I initially intended to draw heavily from the ethnographic methods of data collection as used by Blom and Gumperz (1972), Gal (1979) and Milroy (1987a, 1987b). Rubino (1991) had done as much in her study. However, in planning the study several discrepancies in situational demands surfaced and I had to reappraise my choice of methodology. Most of the studies carried out in the past have relied on the development of a special and close relationship (cf. Milroy’s “friend of a friend”) with the community or group to be studied. Blom and Gumperz (1972) and Gal (1979) had to observe the communities they studied for a long time (two months in Blom’s and Gumperz’s case and one year in Gal’s case) in order to allow them to form testable hypotheses. Being a member of the community I intended to study I was in a privileged position to know enough about the Italian (and the Sicilians in particular) community in Australia not to need too much initial observation. This meant that I would not need to spend too much time getting to know the cultural background of my participants.

The people chosen to participate in the pilot study are a Sicilian- Australian couple (see next section) and their networks. These participants lived in an inner suburb in the Sydney metropolitan area and the couple were observed in four different domains. The situations chosen for investigation were the ones that would most reflect the Italian community’s trilingualism (Rubino and Bettoni 1996). That is, (more details below):
at home within the nuclear family
in the extended family domain;
in the transactional domain (work or shops); and
in the friendship domain.

The method used in this preliminary study would then have to give clear indications that it could yield useable data. Three different approaches were chosen at first: having the data collected by the participants (through tape-recordings), participant observation and a questionnaire. It had been anticipated that one of these methods would then be chosen for the main study that was to follow. The questionnaire to be administered after all data was collected was included to evaluate its usefulness in comparison to the other two methods and to see whether it would show anything else the other methods could not.

One other major factor that influenced the choice of methodology was the data quality requirements that the chosen method would need to yield. Normally in-depth studies in Labovian type research require very high quality recordings or observations. It was, however, decided that the language data needed to be only of a high enough quality to be able to identify the base language of each interlocutor in each situation.

The participants
Having identified the tasks and domains where the data would be collected the next step before choosing the methodology for the study was to identify the target group that was to be studied.

The main theoretical framework of the study is language maintenance within the Italo-Australian community. Therefore, it was decided that the participants had to be the people who, in this context, were most at risk of losing their mother tongue. This would be true of most members of the Italo-Australian community. So, to narrow the choice of study group I relied on my own background. I was born in Sicily and Sicilian is my first language. I also have been living in Australia for all of my adult life and am quite knowledgeable of the local situation. I am also competent in all Sicilian, Italian and English. This led me to surmise that my knowledge and experience would aid me in the relationship with Sicilian-Australian participants and with the transcribing and analysis of the data. For this project, therefore, I chose to study the Sicilian community in Sydney. This meant that the three languages to be studied would be Sicilian, Italian and English. Therefore, the participants had to be trilingual. That is, they had to be from a home environment where they have been exposed to the three varieties that make up a Sicilian-Australian’s linguistic repertoire. Their degree of competence in any of three codes was not important as the lack of full competence in any of the three codes would in itself be an indication of the state of the language within that particular group.

The group chosen, therefore, had to be made up of second generation Sicilian-Australians, that is, of nuclear families where both husband and wife were born in Australia from Sicilian parents or were born in Sicily and moved to Australia when very young. The couple should also have children of their own and, preferably, their parents should be in Australia (within reach of the researcher).

The pilot study
Angela and Santo2 were a couple known to the researcher and who agreed to take part in the pilot study. Their immediate family contains three generations: grandparents, parents and children. Angela is 27 years old and was born in Australia of Sicilian parents. She has a 33 year old brother who is married to a Greek-Australian woman. Angela was finishing her degree in Italian at the University of Sydney and hoped to become a teacher. Her husband Santo is 33 years old was born in Sicily and migrated to Australia with his family when he was 13 years old. He has two older brothers. He works for an airline at Sydney’s airport. They have two children, Claudio, a 5 year-old boy, and Grazia, a 3 year-old girl. Both Angela and Santo also have their parents living in Sydney. They have lived all their married life in Sydney, and have frequent family get-togethers. Their social network seems to be made up almost entirely by the family. Their friends outside the family are not numerous and are of mixed ethnic background; they do not include a large number of Italians or Sicilians.

Even though Santo was born in Italy the family is fairly representative of second generation Italo-Australians (Ware 1981). Angela’s and Santo’s parents, all from working class background, arrived in Australia during the 1950s and 1960s. Their initial efforts were centred on adapting to a new way of life and to give their children a better life than they themselves had; all with little regard as to what language to pass on to the subsequent generation. In both cases no special efforts were made by the parents to encourage the learning of Italian. In both families language only became an issue with Angela’s and Santo’s decision to have children of their own. Both of them communicated to the researcher a desire for their children to learn Italian. This was also one of the overriding factors for Angela’s decision to study Italian at university.

Domains and tasks
In the nuclear and extended family domains
As has been already stated, the aim of this pilot study was to identify a method of data collecting which could be used with a larger group of participants in the main study that was to follow. In adopting a methodology for the pilot study the researcher had to take into consideration the fact that it would not be possible to establish a close relationship with all the participants in the main study. Rubino (1990) was able to do so and, therefore, make sure that the presence of the researcher in the home would not affect the language used (Labov’s observer’s paradox). For this reason, in the family domains participant observation was excluded from the very start. To this end it was decided that tape-recording would be trialed. Rubino (1993) successfully encharged a member of the family she was studying with a high-quality tape-recorder. It was to be switched on during the family meals. Rubino states (1993, p.115) that in this manner she was able to access the “core” language of the family. I also felt that a tape-recorder left out of sight during family gatherings would provide the best possible results within this domain. The participants, therefore, were observed through an analysis of the tape-recording as they interacted with their children (at home) and their relatives (in family gatherings). To make sure each voice on the tape was identified and to clearly indicate with whom the participants were using each language, the researcher listened to the tape together with one of the participants following each recording.

The use of a tape-recorder did give rise to some ethical concerns. It is simply unethical to tape-record people without their knowledge. However, drawing attention to the tape-recorder contrasted with the need to have everyone involved forget about it as they interact in a normal way. It was decided that everyone would be told about the tape-recorder before it was switched on. It would then be placed in an unobtrusive place and that the researcher would ignore the first ten minutes of the recordings or until it was clear that the conversation seemed not to be affected by the tape-recorder.

In the transactional domain
In this domain the participants were observed as they interacted outside the home. This domain was included to try and see how the participants interact with other Italians away from their family domains. Aiming, at first, to keep the same method across domains, the use of a tape-recorder was trialed. However, as will be discussed later, this proved problematic.

Participant observation and note-taking were also trialed. In this way, the researcher accompanied the couple while they did their shopping in a few Italian shops. The danger in this domain was that interacting with the researcher before and during the shopping trip might establish a language trend that the participants would then keep while he was present. However, the decision to have the researcher present was made with the thought that outside of the home and in crowded shops any effect the presence of the researcher might have on the participants’ language choice and use would be minimized.

In the friendship domain
The aim of investigating the language use in this domain was to see to what extent language choice and use is governed by one’s social network. Since it had been decided that the researcher could not establish a close relationship with the couple in the home and family domains it also became obvious that the same would hold true in their friendship domains. It was also important to keep the same ethical considerations in mind. The two people that agreed to take part in the pilot study refused, at first, to bring a tape-recorder to a meeting with their friends. They felt that bringing a tape-recorder to a restaurant, bar, or their friends’ homes, would not have been practical or ethical. More importantly, they also thought it might upset the friendly atmosphere of the meetings and therefore affect the choice of language variety. Since participant observation had already been ruled out, a different method to those initially proposed was tried out. After receiving detailed instructions (see Gibbons 1987), the participants were asked to keep a diary of the language they used while interacting with friends. The decision to trust the participants’ own judgement on reporting the correct code used necessitated from the lack of any other reliable method.

The participants used diary-keeping at first. Fortunately, however, after their successful experiences with the tape-recorder in the home and family domains the participants decided that they would make a tape-recording of a reunion with the same friends. I, therefore, had the opportunity to compare two different methods used in the same domain and, more importantly, with the same friends. The only drawback being that the two sets of data were not taken at the same time or in the same place.

Questionnaire
The purpose of administering a questionnaire was to test the reliability of self-reporting against “real” data, and to try and gather socio-economic and attitudinal data on the participants. The questionnaire was a variation of the one used by Rubino and Bettoni (1990) and was administered a few months after all other data was collected. It is divided into four main parts. The first part is aimed at eliciting demographic and self-assessment data about the participants. The second part puts the participants in different domains contextualized according to the interlocutor, the topic of the conversation and where it takes place, and asks to provide what language they would use in those situations. For comparison, the situations included were similar to the ones covered in the other phases of the research. The third part aimed at establishing the degree of contact each participant has with the Italian community and Italy. The fourth part, which is an innovation over the questionnaire used by Rubino and Bettoni, elicited attitudinal data through open-ended questions. In this part a discussion was prompted by the researcher with questions on what the participants’ thoughts were on language maintenance in general, on the three languages in question and who should be responsible for their children’s language education.

Effectiveness of methods used
The following discussion will be an analysis of each situation, the method used and the type of data obtained. This analysis will identify whether the methodology used in each domain succeeded in obtaining useable data.

At home
The recording in this domain is of the couple and their two children sitting in the kitchen having their evening meal. Angela, who was in charge of the tape-recorder, had earlier told the rest of the family that the dinner would be taped. She then switched the tape-recorder on a few minutes before everyone came to the table and the tape-recorder was left on a shelf near enough the table to make the recording possible but not obvious enough that it would be a constant reminder.

The recording in this situation was clear enough to produce useable data. For example, it provided clear indications of code-switching. The following are typical examples of such instances (Italian is shown in bold and Sicilian is shown in bold italics):
(1) At the end of a long conversation in Italian:
Angela (to her son) hurry up Claudio!
Angela (to Santo, her husband) you know on Thursday morning I
could drop Grazia (the daughter) off at school
Santo Thursday I work
Angela (to her son) Claudio, you know those couple of mothers?
Cal’s mother and Jessica’s mother?
Claudio yep
Angela (to Santo) stiamo pensando di ire o parcu pi fari iucari
‘we’re thinking of going to the park to let the
i piccirilli. OK? So I’ll be home by the time you get up
children play’
Angela (to Claudio) u nonnu ha telefonato
‘grandfather phoned’
Claudio: ah?
Angela: ah what?

(2) After a few minutes silence while eating:
Santo (to Grazia) cosa hai fatto a scuola?
‘what did you do at school?’
Grazia: ho fatto testing
‘I did testing’
Santo: che testing?
‘what testing?’
Santo (to his son) tu c’iai bisogno d’un volantino vero?
‘you need a napkin don’t you?’

With other relatives
The family was recorded at a family dinner. The recording was done in the dining room as the extended family sat down to eat. In this family gathering were present: the children, both sets of parents, Angela’s brother and his wife (who is of Greek origin and is learning Italian; however, neither speaks nor understands Sicilian), and one of Santo’s brothers. As far as the actual recording is concerned, the same procedure as the one described for the home recording was followed. That is, everyone was warned that the get-together would be recorded with the exception this time that they were not told the actual time when the tape-recorder was to be switched on. A 45-minute recording was obtained.

This recording was considerably more difficult to transcribe than the one taken at home. The differences being that, while before there were four people, now there were nine; and that while previously they were sitting down to a meal, now they were all milling around the kitchen and dining area adjacent to it. However, the tape-recorder and microphone were of a high enough standard to capture most of what was said. Enough was captured on tape to make analysis of the language use possible. In this context what was needed was to be able to identify what language Angela and Santo were using, with whom and on what topic was discussed in a particular interaction. Language use was clearly identifiable from the recording. Changes in topic and codes-witching were also identifiable, except those that were away from the immediate vicinity of the tape-recorder. However, the nature of the get-together meant that enough interactions happened within range of the microphone to safely say that this method had a high degree of reliability needed for naturalistic data collection. The following are examples of the interactions recorded (Italian is shown in bold and Sicilian is shown in bold italics):

(3) In a general discussion with everyone
Angela ammeno era pulito però. He didn’t break anything.
‘At least he was clean though.’

(4) To his brother
Santo eh? Ma che come back! I don’t have to pay?
‘eh? But what (what do you mean) come back!’

(5) To everyone
Santo u patri figlio spiritu santu e così sia
‘the father son Holy spirit Amen’

(6) To her daughter then to everyone
Angela come on Grazia veni a mangià
‘come on Grazia come and eat’
ebbè allora amma aspittare a mangiare.
‘well then we have to wait to eat’

In the transactional domain
In this task two methods were trialed: tape-recording and participant observation. However, the recording proved to be very difficult to transcribe. Shops full of customers and shopkeepers produced very muddled and unclear recordings, making it impossible to distinguish who was talking to whom.

Participant observation proved a better method. The researcher accompanied the participants on a routine shopping trip in the suburb of Haberfield, Sydney. This particular shopping centre is made up largely of shopkeepers from an Italian background and is a popular shopping place for the Italo-Australians living in the inner suburbs of Sydney. To lessen the impact of the researcher’s presence the participants and the researcher socialized over a cup of coffee in a local Italian Café before the shopping began. This allowed the participants to get used to the extra presence. All note-taking was done immediately after each interaction and done discretely away from the participants so as to not distract them during each interaction and not remind them they were under observation.

The data obtained shows that this method as well can provide useable data on language choice. Angela and Santo each interacted, in 3 shops (a delicatessen, a bakery and a café) with four people. The researcher was able to ascertain that the people were 3 women and 1 man, all in their 40’s. Two of the women were Sicilian and the rest were Italian but not Sicilian. The following is an example of the interactions observed (Italian is shown in bold):

(7) To the Sicilian-Australian shopkeeper
Angela  I’d like some ham quello lì (‘that one’)
E’ buono? (‘Is it good?’)
Grazie, arrivederci (‘Thank you, good bye’)

The advantage of this method is that since the researcher is present while the participants are interacting any queries can be clarified immediately after the shopping trip while everything is fresh in the participants’ mind. For example, the participants were able to inform the researcher that they were aware of the regional background of the people they interacted with. In example (7) above Angela was aware that the shopkeeper was of Sicilian background, but still chose to speak to her in Italian and not Sicilian. For a detailed discussion on the choice and use of language in the Italo-Australian community see Rubino and Bettoni (1996) and Cavallaro (1998).

In the friendship domain
In this situation Santo, Angela and their two children met with two friends and their daughter. Of the friends, the husband is 27 years old and was born in Sicily. He came to Australia at the age of 20. The wife, instead, is 35 years –old. She was born in Australia but has been back to Italy many times. They have a 1 year-old daughter.

In the first meeting with these friends Santo and Angela kept a record of what language they spoke with their friends during the evening. This was mostly done after they got home. They also took the notes together so that they could consult each other if unsure what language they had used or what the topic of the conversation had been.

Given the right conditions it can be a useful tool for sociolinguistic research as Gibbons (1987) has shown. This pilot study showed that the participants’ self-reporting, while accurate as far as the language use is concerned, was not very detailed.
In their second meeting Santo and Angela brought a tape-recorder with them. The approach here was the same as that used by the participants at home and within their family. That is, Angela again was in charge of the tape-recorder, warned their friends that their talk would be taped, and switched the tape-recorder on before they sat down to eat. This produced a recording of about 50 minutes.

This recording was, again, clear enough to show great variation between codes. It was also interesting to note that the data obtained in this domain were clear enough not only to identify that Italian was the base language when the group spoke all together, but Sicilian was used more whenever the group divided up. The explanation for this variation is not within the scope of this article and will be investigated further in a follow-up study. The following are examples of the interactions recorded (Italian is shown in bold and Sicilian is shown in bold italics):

(8) To everyone
Angela ah il cugino di Giuseppe. Non ha fatto il militare?
 ‘ah Giuseppe’s cousin. Didn’t he do national service?’

(9) To his male friend
Santo No sacciu chi evi. Na vota uncuntrai
I don’t know who he is. I met him once’

The questionnaire
A comparison of the language observed or recorded and the answers given to the questionnaire shows some interesting points. On the subject of what language the participants used with their children both participants said in the questionnaire that in all situations they would use only Italian, except when helping them with their homework and then English would be used. The questions on their attitudes towards the three languages also showed that while they feel English and Italian are the more important codes, they wish their children to grow up with a good knowledge of Sicilian as well. This is reflected in the tape-recording, which did show that whenever the children are around the base language is indeed Italian and that English and Sicilian play secondary roles at home. This is a very important notion: it corroborates the studies on language attitudes which indicate that logical, rational and status conscious thinking favours the standard language while emotions and ethnic solidarity favour the dialect.

The recordings, on the whole, corroborated the questionnaire’s answers on language choice with the rest of the family. However, the recordings and the answers to the questionnaire differ in the amount of Italian spoken with the family. The recordings indicated that Santo and Angela used a substantial amount of Italian with their family. In the questionnaire both participants had stated that Italian would have been used much less than the recordings showed. The recordings were also able to give a more complete picture that the questionnaire. In the questionnaire the participants answered that they would switch to Italian whenever the children or Angela’s sister-in-law would come into the conversation. However the recordings show that things are not so clear cut. The recordings showed that Santo would at times not change from Sicilian into Italian whenever his children became involved. It is these types of situations that the questionnaire did not handle very well. The recordings also indicated that Angela and Santo were very aware of certain people whose presence required a switch to Italian, in this case the children and sister-in-law. For a questionnaire to consider all the complexities that can arise in any given interaction its questions would have to be redesigned to allow for them and many more would need to be added. As such this questionnaire took nearly an hour to administer. Any redesigning, therefore, would entail an increase in size and of the time needed for all the questions to be answered. This would make it very impractical.

With Sicilian friends they reported in the questionnaire that they would use Sicilian (Santo) and English with some Sicilian (Angela). But, while the recording corroborated the individual interactions, the recording also shows that Italian was the base language when the group was together. This observation could not have been made solely by analyzing the questionnaire.

Another, seemingly, inaccurate self-reporting was on the question of what language they would use in a Sicilian shop. Santo reported he would use Sicilian, which he was not observed to do. He used Italian when he was observed talking to a Sicilian woman in a delicatessen. One possible explanation for this may be that the researcher’s presence affected Santo’s choice. However, Santo and the researcher interacted almost exclusively in Sicilian during the shopping trip. It seems, instead, that the knowledge that Italian is the code for the more public situations is something that Santo may have internalized. This will be the subject of investigation investigated further in the follow-up study.

Conclusions
No firm conclusions can be made on the actual data collected. Indeed it was not the aim of the pilot study to include an analysis of any data obtained. The sample is too small and the fact that Santo was born in Italy did seem to skew the data towards Italian and Sicilian as far as he is concerned. As far as the methodology is concerned, this study has shown that having the participants collect the data themselves with the use of a tape-recorder and participant observation are two very reliable ways of obtaining naturalistic data.

The lack of detailed information from diary keeping is something that could not be overcome without more lengthy instructions and more practice. It was felt, however, that forcing the participants to concentrate on the language they use at all times was to further formalize the situation with the subsequent impossibility of obtaining naturalistic data. There is no doubt that out of the two methods trialed in the friendship domain, the tape-recording yielded more information on the interactions; while at the same time dismissing any doubts on the accuracy of the data. The questionnaire’s greatest asset was that it provided a lot of background and attitudinal information.

The pilot study, in the end, did identify the best methods for the collection of naturalistic language data. However, far from preferring one method, the pilot study has shown that not one single method is suitable for all situations. What it showed was that a detailed analysis of the language dynamics of a minority group is best attainable if one makes use of all the means at one’s disposal, adapting each method to suit the different domains to produce the most complete picture possible.

Notes
1 The term Italian will be used from now on to indicate the variety opposed to the dialect (Alfonzetti, 1998:208).
2 Pseudonyms have been used in this article because of the wish by the participants to remain anonymous.

References
Alfonzetti, G. (1998). The conversational dimension in code-switching between Italian and dialect in Sicily. In P. Auer (Ed.), Code-switching in conversation: Language, interaction and identity (pp. 180-210). Routledge: London and New York,.

Berruto, G. (1987). Sociolinguistica dell’italiano contemporaneo. Roma: La Nuova Italia Scientifica.

Berruto, G. (1989). Main topics and findings in Italian sociolinguistics. International Journal of the Sociology of Language 76, 7-30.

Bettoni, C. (1981). Italian in North Queensland. Townsville: James Cook University Press.

Bettoni, C. (1985). Tra lingua dialetto e inglese: il trilinguismo degli Italiani in Australia. Sydney: FILEF Italo-Australian publications.

Bettoni, C. (1986). Italian language attrition in Sydney: the role of birth order. In C. Bettoni (Ed.), Altro Polo. Italian abroad: studies on language contact in English-speaking countries (pp. 61-85). Sydney: May Foundation for Italian Studies (University of Sydney),

Bettoni, C. (1989). Maintaining and developing Italian in Australia. Vox, The Journal of the Australian Advisory Council on Languages and Multicultural Education 3, 48-53.

Bettoni, C. and Gibbons J. (1988). Linguistic purism and language shift: a guise-voice study of the Italian community in Sydney. International Journal of the Sociology of Language 72, 15-35.

Bettoni, C. and Gibbons, J. (1991). L’influenza della generazione e della classe sociale sugli atteggiamenti linguistici degli italiani in Australia. Rivista italiana di dialettologia 14, 113-138.

Blom, J.P. and Gumperz, J.J. (1972). Social meaning in linguistic structure: Code switching in Norway. In J.J. Gumperz and D. Hymes (eds.) Directions in sociolinguistics (pp. 407-434). New York: Holt, Rinehart and Winston.

Callan, V.J. and Gallois, C. (1982). Language attitudes of Italo-Australian and Greek-Australian bilinguals. International Journal of Psychology 17, 345-358.

Cavallaro, F. (1998). The language dynamics of the Italian community in Australia. Unpublished doctoral dissertation. Melbourne: Monash University.

Clyne, M. (1987). Bilingualism and community languages: The educational consequences. L’italiano in Australia. Rome: Il Veltro Editrice, Il Veltro 31(1-2), 57-71.

Clyne, M. (1988). Community languages in the home: A first progress report. Vox, The Journal of the Australian Advisory Council on Languages and Multicultural Education 1, 22-27.

Clyne, M. (1991). Community languages: The Australian experience. Cambridge: Cambridge University Press.

Clyne, M. (2003 ). Dynamics of language contact: English and immigrant languages. Cambridge: Cambridge University Press.

Clyne, M. and Jaehrling, S. (1989). Trends and factors in language shift in Australia. Vox, The Journal of the Australian Advisory Council on Languages and Multicultural Education 3, 63-71.

Clyne, M. and Kipp, S. (1997). Linguistic diversity in Australia. People and Place. 5(3), 6-11. Retrieved May, 4 2006 from http://elecpress.monash.edu.au/pnp/free/pnpv5n3/kipp.htm

Gal, S. (1979). Language shift: Social determinants of linguistic change in Austria. New York: Academic Press.

Gibbons, J. (1987). Code-mixing and code choice: A Hong Kong case study. Clevendon, Philadelphia: Multilingual Matters.

Hakuta, K. (1986). The mirror of language: The debate on bilingualism. New York: Basic Books.

Labov, W. (1972a). Sociolinguistic patterns. Philadelphia: University of Pennsylvania Press./Oxford: Basil Blackwell.

Labov, W. (1972b). Language in the inner city. Philadelphia: The University of Pennsylvania Press.

Labov, W., Cohen, P., Robins, C. and Lewis, J. (1968). A Study of the non-standard English of Negro and Puerto Rican speakers in New York City. Department of Linguistics, Columbia University, New York. 2, 76-152.

Milroy, L. (1987a). Language and social networks. (2nd edition). Oxford: Basil Blackwell.

Milroy, Lesley (1987b). Observing and analysing natural language: A critical account of sociolinguistic method. Oxford: Basil Blackwell.

O’Bryan, K.G., Reitz, J.G. and Kuplowska, O.M. (1976). Non-official languages: A study in Canadian multiculturalism. Ottawa: Supply and Services Canada.

Pellegrini, G.B. (1960). Tra lingua e dialetto in Italia. Studi mediolatini e volgari, 8, 137-153.

Poplack, S. (1980). Sometimes I’ll start a sentence in Spanish Y TERMINO EN ESPAÑOL: towards a typology of code switching. Linguistics, 18, 581-618.

Rubino, A. (1990). From guest to family friend: fieldwork progress within the Italo-Australian context. Australian Review of Applied Linguistics, 13(1), 77-99.

Rubino, A. (1991). Patterns of language choice and language contact in formal and informal settings. Rivista di linguistica, 3(2), 277-305.

Rubino, A. (1993). From trilingualism to monolingualism: a case study of language shift in a Sicilian-Australian family. Unpublished doctoral dissertation. Department of Italian, University of Sydney.

Rubino, A. (1996). The children’s role in parents’ language shift. A Sicilian-Australian case-study. In Kiernan, S. (Ed.), Altro Polo. Italian studies in memory of Frederick May (pp. 209-237). Sydney: University of Sydney (Frederick May Foundation).

Rubino, A. (2000). Playing with languages: language alternation in Sicilian-Australian children’s conversation, Rassegna Italiana di Linguistica Applicata, 32(2), 89-108.

Rubino, A. (2004). Trilingual women as language mediators in the family: A Sicilian-Australian case-study. In A. Rubino (Ed.), Using and learning Italian in Australia (pp. 25-49). Special issue of the Australian Review of Applied Linguistics, S18. Melbourne.

Rubino, A. and Bettoni, C. (1990). The use of English of Italo-Australians. Australian Review of Applied Linguistics, 14(1), 59-89.

Rubino, A. and Bettoni, C. (1996). Emigrazione e comportamento linguistico. Un’indagine sul trilinguismo dei siciliani e dei veneti in Australia. Galatina.

Sanga, G. (1981). Les dynamiques linguistiques de la société italienne (1861-1980): de la naissance de l’italien populaire à la diffusion des ethnicismes linguistiques. Langages, 61, 93-115.

Smolicz, J.J. and Secombe, M.J. (1986). Italian language and culture in South Australia: a memoir approach. In C. Bettoni (Ed.), Altro Polo. Italian Abroad: studies on language contact in English-speaking countries (pp. 27-59). Sydney: Frederick May Foundation for Italian Studies (University of Sydney).

Trumper, J. (1984). Language variation, code switching, S. Chirico (PZ) and the migrant question (Konstanz). In P. Auer and A. Di Luzio (Eds.), Interpretive sociolinguistics: Migrants – children – migrant children (pp. 29-54). Tübingen: Gunter Narr Verlag.

Trumper, J. (1989). Observations on sociolinguistic behaviour in two Italian regions. International Journal of the Sociology of Language, 76, 31-62.

Trumper, J. and M. Maddalon (1982). L’italiano regionale tra lingua e dialetto. Cosenza: Brenner.

Ware, H.(1981). A profile of the Italian community. Melbourne: Australian Institute of Multicultural Affairs and Italian Assistance Association (Co.As.It.).

Tags: , , , , , ,

Category: 2006