Data Processing
Video-recordings made in
Chisasibi were transferred from the CamCorder onto the project
computer, compressed and uploaded to the CCLAS project computer at
Memorial University. The original tapes reside under lock and key
at the Cree School Board until one of the research team visits and
takes them back to Memorial University, where they are stored under
lock and key in the Linguistics Department’s Native Languages
Archive room. A DVD copy of each movie is made in Chisasibi for the
Cree School Board Archive.
When the recordings are on the project computer at Memorial University, the research team in St John’s is able to begin the data processing. Since the summer of 2005, CCLAS has employed several undergraduate students who are either doing a major or a minor in Linguistics. Student team members are selected from among those who express an interest in the field of first language acquisition and training is provided on an ongoing basis. Ideally, these students will seek to be integrated into the intellectual life of the project by continuing on to graduate studies, taking the CCLAS data as the focus of their personal research. All the students work in the Speech Sciences and Language Acquisition Laboratory under the supervision of Yvan Rose. For convenience, we refer to any member of this student team as ‘the data processor.’ The data goes through a number of stages in order to be ready for analysis. Some of the work is done at MUN, some of it is done in Chisasibi:
These nine stages are the essential precursors to preparing the data for linguistic analysis. Any discrepancies between the child’s IPA actual production and the target form will be apparent when we compare IPA actual with IPA target. This provides the data necessary for assessing both phonological and morphological development. Phon’s alignment module, already capable of automatic phonological alignment between target and actual sounds and syllables, will also facilitate the manual alignment of actual and target morphemes, allowing us to decide which (if not all) have been acquired at any given stage.
When the recordings are on the project computer at Memorial University, the research team in St John’s is able to begin the data processing. Since the summer of 2005, CCLAS has employed several undergraduate students who are either doing a major or a minor in Linguistics. Student team members are selected from among those who express an interest in the field of first language acquisition and training is provided on an ongoing basis. Ideally, these students will seek to be integrated into the intellectual life of the project by continuing on to graduate studies, taking the CCLAS data as the focus of their personal research. All the students work in the Speech Sciences and Language Acquisition Laboratory under the supervision of Yvan Rose. For convenience, we refer to any member of this student team as ‘the data processor.’ The data goes through a number of stages in order to be ready for analysis. Some of the work is done at MUN, some of it is done in Chisasibi:
- Stage 1: (MUN) Recordings are imported into Phon software
- Stage 2: (MUN) Segmentation
- Stage 3: (MUN) Double-blind transcription of child speech
- Stage 4: (MUN) Data validation (supported by acoustic analyses where clarification is necessary)
- Stage 5: (Chisasibi) Segment checking (contextualization of child speech)
- Stage 6: (Chisasibi) Oral commentary is provided (translation into English is provided, and target forms for Cree child forms are provided as necessary)
- Stage 7: (MUN) Transcription of audio of files from stage 6
- Stage 8: (MUN) Translation check, orthography and morpheme breakdown
- Stage 9: Updating of Phon's fields based on work done in stage 8
These nine stages are the essential precursors to preparing the data for linguistic analysis. Any discrepancies between the child’s IPA actual production and the target form will be apparent when we compare IPA actual with IPA target. This provides the data necessary for assessing both phonological and morphological development. Phon’s alignment module, already capable of automatic phonological alignment between target and actual sounds and syllables, will also facilitate the manual alignment of actual and target morphemes, allowing us to decide which (if not all) have been acquired at any given stage.
