This project, providing
systematic descriptions
of the tasks used in all
languages in the KPG
examinations, is aimed
at the collection of the
essential meta-data for
standardisation of the
exam specifications and
control of the exam
battery. So far, it has
involved the creation
and the continuous
enrichment of a database
in which tasks are
stored for each
language, exam period,
and proficiency level
along with information
regarding the task type
(e.g. multiple choice,
multiple match, gap
filling, etc.), what a
given task checks (e.g.
reading comprehension,
language awareness,
production of written or
spoken language, written
or oral mediation) as
well as descriptions of
the texts associated
with a given task (the
source text to which the
candidate is asked to
respond and the target
text which is to be
produced). In the theory
of language underlying
the KPG examinations,
the notion of text is
understood as the
material configuration
of various aspects of
the communicative
context in which
language functions. In
the Task Analysis
database, texts are
described in terms of
bundles of features
specifying their
communicative function,
or the so-called text
genre. These features
include the topic and
the domain to which a
text pertains (e.g.
environment,
entertainment, travel,
sport, etc.); the
discourse environment of
the text (e.g.
newspaper, magazine,
webpage, encyclopedia,
dictionary, etc.),
incorporating
distinctions regarding
the communicative roles
or identities of the
author and addressee of
the text; the text type
(e.g. article,
announcement, report,
advertisement, email,
etc.), identified in
terms of the
communicative goals for
which the text has been
produced as well as its
structure; and the text
process (e.g.
description, narration,
explanation,
argumentation,
instruction) by means of
which the communicative
goals are fulfilled. The
values that these
features may take are
specified in pre-defined
typologies, i.e. the KPG
typologies of domains,
discourse environments,
text types and
processes. Treating
genres in terms of
combinations of
individual parameters
has significant
advantages over their
conception as atomic
categories. The model
that has been
implemented supports the
classification of texts
on the basis of
well-defined criteria
designating coherent
categories. This
classification is
readily extensible to
new categories.
Furthermore, it enables
identifying how
linguistic choices
attested in texts
correlate with specific
aspects of their
communicative context.
