• No results found

RQ 4. Putting together the quantitative analysis (RQs1 & 2) and the qualitative analysis (RQ3) what does this tell us about the nature of bi-varietal language use in young

4.4 Data Collection 1 Method

Variation theory is concerned with the language produced by speakers in everyday interactions. As such, it is necessary to capture data that best reflects the language used in such contexts. The traditional variationist method for obtaining such language samples is the ‘sociolinguistic interview’ (Labov 1966, 1984), wherein an informal conversation takes place between two people, the interviewer and the consenting participant. The interviewer is generally known to the participant, and a speaker of the same language variety/-ies. They invite the participant to share autobiographical stories as a means of distracting them from focusing on their speech. The goal is to collect ‘unmonitored’ speech, considered to be as close as possible to the vernacular, which is “the most systematic data for linguistic analysis” (Labov 1984:29).

In the present context, the sociolinguistic interview was not a realistic method for the collection of language data. The participants in this study are children aged between 5-8 years, who do not engage in the kind of lengthy conversation about one’s life that adults might be expected to perform and do with a sense of ease and informality. This is not a

88 verbal activity that the participating children typically engage in and it was hard to imagine a way that such a performance could be achieved.

Other studies of child language variation in young children have encountered similar problems in other settings, and have, like the present study, instead used recordings of children at play (in both home and institutional settings), taking their naturalistic and spontaneous utterances produced in these contexts as the primary linguistic data (e.g. Tagliamonte & Molfenter 2007, Roberts 1997).

Data for this thesis come from the corpus of recordings made for the Aboriginal Child Language Acquisition 2 (ACLA2) project (http://arts.unimelb.edu.au/soll/research/past- research-projects/acla2), a longitudinal study of the interaction of home and school languages in Aboriginal communities. This project has several other field sites, and multiple aims. The design of the recording schedule was focused on capturing naturalistic language use in a range of different home and school contexts. Archiving and permissions protocols for the corpora built under this project are managed by the Chief Investigators Prof. Jane Simpson (The Australian National University) and Prof. Gillian Wigglesworth (University of Melbourne).

In July of 2009, I undertook an initial introductory visit to Ipmangker with Samantha Disbray, who had participated in the first ACLA project, recording children in Tennant Creek (see Disbray 2008a). Disbray is a personal friend of Linda Dobbs, one of the senior women at Ipmangker who was enthusiastic about the project. I returned in September to conduct a preliminary field trip. The aims of this trip were to conduct a few recordings with age-relevant children, and transcribe them, so that parents and other participants could have a better sense of what the project would entail. It was also during this trip that two research assistants who were able to assist with transcribing emerged. The three main field trips were then conducted over the next two years, at six- monthly intervals:

Field Trip 1: April-May 2010 (5 weeks) Field Trip 2: Oct-Nov 2010: (7 weeks) Field Trip 3: April-May 2011 (4 weeks)

89 4.4.2 Corpus description

During each field trip I made at least 2 video recordings of each focus child at home, and at least two at school. Each recording was around 60 mins long and about half of this was usually transcribed. The Ipmangker corpus now consists of 50+ hours of

naturalistic video recordings of six focus children aged 5-8 (plus their relatives and classmates who happen to appear, with permission, ‘in the frame’).

There are also recordings that form a small sub-corpus of adult Alyawarr English speech. One of the research assistants allowed me to (audio) record several episodes of

her telling a story to her son using a wordless picture book11. The several Alyawarr

English-speaking teaching assistants and pre-school teachers also appear in the school recordings, speaking in both Alyawarr English and SAE. From this some tentative observations about the use of Alyawarr English by adults are made in later chapters. Recordings of the children made in the pre-school and school follow the normal routines of the day. As such the corpus contains recordings of a variety of school activities: literacy (e.g. reading from texts, talking about texts, writing practice, phonics activities, sight words recitation), numeracy (e.g. counting games, maths worksheets), science (e.g. learning about trees, volcano making), music, art, and various other informal interactions. The corpus also contains recordings from a number of ‘bush tucker’ excursions within areas surrounding the school. Recordings in the pre-school contain some formal interactions, such as the ‘welcome’ activity (involving discussion of the day’s date and weather), but given the play-based pedagogy of pre-schooling, the corpus also records much informal teacher-student interaction.

In the home context, recordings were typically made in small groups (3-4 children) around a set of toys provided by the researcher (detailed in the following section). The provision of interesting new toys was largely a logistical hack: it made the children more likely to stay within the range of the video camera. The corpus does contain some recordings in more spontaneous contexts (at the cubby house, at the basketball court) though these are in the minority and required re-positioning the recording devices more often.

11 There are some restrictions on the use of this material since the tragic death of this woman. For example, I no longer use the recordings themselves, but rather work from transcriptions.

90 4.4.3 Equipment

The basic recording kit consisted of:

Video camera: Canon Legria HF21 High Definition digital camera, usually placed on a tripod off to the side of the play area.

Radio/lapel mics: Sennheiser ew100 G2. This consists of a small lapel mic that was clipped to the child’s t-shirt, and a transmitter pack (about the size of a deck of cards) that was attached to the waistband of their trousers/shorts. The receiver plugs directly into the Zoom H4 unit.

Audio recorder: Zoom H4n recorder, on a mic stand. Recording input from the built-in mic and also from the radio/lapel mic (using an XLR input jack).

Toys: For the recordings at home it was necessary to provide the small group of children being recorded with a set of toys to play with. This was primarily done with the aim of keeping the children in the one spot, to allow recording to take place - normal play activities can involve a good deal of roaming around the community. For this purpose I always took several kits of toys, including a couple of new ones for each trip. These included:

Doctors dress-up kit Cooking/food kit Cash register/shop kit

‘John Deere’ station machinery figurines ‘Medieval Battle’ figures

Small dolls with changeable clothes Meccano set

‘Guess Who’ game

Wordless picture books: At each field trip I recorded each of the children telling the story from a wordless picture book. Sometimes this elicited more SAE-like speech and

91 sometimes Alyawarr English. For this purpose I used a series of culturally-relevant books created by Carmel O’Shannessy for her work with children at Lajamanu (Northern Territory) associated with ACLA, and focusing on a new mixed-language

variety12. Data collected in these activities was ultimately not used in any of the statistical

analyses carried out in this thesis as narrative text production often contains its own internal temporal structure that would make determining the temporal and aspectual semantics difficult without specific study of the way this genre is constructed by the children (see for example, Disbray 2008a).

4.4.4 Processing and transcribing

The raw video was first backed up and then imported into iMovie and saved at a more manageable size (in .m4v format) for transcribing. The superior-quality audio

recordings from the Zoom H4 (both internal mic and signal from the lapel mic) were synced to the video using Audacity. The two separate signals (internal and lapel) were merged to create one single audio file, but the separate signals were also retained so that if, for example, the child wearing the lapel mic moved away from the Zoom H4 unit and the camera followed the child, the competing ‘noise’ from the ZoomH4 unit could be ‘switched off’ by switching to the ‘lapel mic only’ signal. The transcribing software ELAN allows this to happen very easily, by switching to a different master media file. Transcribing in ELAN was done with both the superior audio recording and the video displayed. Initially I would run through the recording in ELAN and transcribe what I could, or at least enter the fields where speech was occurring. This was later added to and checked by one of the research assistants, Ms G Kelly (now deceased) or Ms Michelle Dobbs. Both of these women speak Alyawarr English, and were intimately familiar with the speech of young children in the community being mother/aunty to several young children.