Combining text level metadata with tier data in queries/CLAN training

[Patrick McConvell]

Dear Brian & colleagues,

We are a group of  seven researchers (including four Ph.D students) in the
ACLA (Aboriginal Child Language Acquisition) project beginning work this
year on children¹s language and the input they receive in
bilingual/multilingual Central Australian communities. We are collecting
digital video data and beginning to transcribe using CHAT & CLAN. We are
interested in the CLAN analysis tools but have not done much coding or used
that part of the program at all yet.

We also see a need for a database of information separate from the CLAN
program for information (metadata) on a-v resources, transcripts,
participants and sessions. We are working out how this will interact with
CLAN, particularly the headers, with the help of Melbourne University
Computer Science department.

As far as we can see from the CLAN Manual,  CLAN tools (programs) operate
either on headers or tiers, but not on both together. We are particularly
concerned with answering such questions as Œwhat combinations of
words/syntactic elements occur with children between certain ages eg 18 and
24 months?¹. If using CLAN alone this would involve a combination of
information from headers and tiers (the latter analyzed by COMBO or a
similar tool). If we are correct that no single tool would bring these two
sets of data together (participant age and some combination of linguistic
elements on a tier), then presumably one process could be Œpiped¹ into
another to filter the participants as to age at the time of recording.
However we have not found any such filtering tool described in the CLAN
manual.

In section 7.7 (pp.146-7) there are exercises which refer explicitly to
comparing children at age 20 and 32 months in various ways using MLU. It is
not clear to us, however, what mechanism is used to filter out these cohorts
of children from others, unless perhaps it is a manual process.

It is probable that we have missed something basic here, so please let us
know. The answer to this question and similar ones is important to us in
designing our database and its interface with CLAN.

We are also needing assistance with learning the coding and analysis levels
of CLAN. We are managing with a brief self-run workshop to bootstrap
ourselves for the transcription level in February 2004. However we would be
glad to correspond with an advanced user who might act as a tutor for a
training session in Australia later in 2004.

Season's greetings

Patrick McConvell, AIATSIS, Canberra mcconvell at ozemail.com.au
Jane Simpson, U Sydney jhs at mail.usyd.edu.au
Gillian Wigglesworth U Melbourne, gillianw at unimelb.edu.au