Writing, working memory and dual-TASK
central executive then evolved (Baddeley, 1996) to a
more specified system performing functions such as
retrieval from long-term memory, selective and
divided attention, switching attention, task
switching, coordination of the slave-systems. As
such, the central executive is now conceived as a
system regulating and controlling information in
working memory. Whatever its conceptualization,
the central executive has been recognized as playing
a crucial role in complex cognitive activities and
being in the heart of the control of action. Finally,
whether or not the different functions assigned to the
central executive are separated is still discussed
(Baddeley, 2001).
Competing models of working memory have
been proposed (for a presentation of these models,
see Myake & Shah, 1999). For example, Cowan
(2001) regards working memory as an activated
portion of long-term memory and Just & Carpenter
(1992) conceived working memory as a single and
limited pool of cognitive resources in which all
processes draw and are in competition for sharing
that resources. Although at a rapid glance these
different approaches of working memory seem
antagonistic, they can be considered complementary.
Indeed, following Baddeley (1996) these competing
theories of working memory might simply reflect
the functions of the central executive and, therefore,
can be reconciled with its multi-component model
of working memory. Furthermore, according to
Baddeley and Hitch (2001, p. xvii) there is "a
general acceptance of the need to assume both of a
general executive system and specific verbal and
visual systems."
Working memory in writing
A large number of studies investigated the role of
working memory in verbal activities (e.g., Caplan &
Waters, 1999; Gathercole & Baddeley, 1993; Just &
Carpenter, 1992). Most of these studies explored
the impact of individual differences on verbal
performance as, for example, in text comprehension
studies. In writing, the impact of individual
differences in working memory on writing
performance was mainly investigated by Swanson
and Berninger and their collaborators who adopted a
correlational approach for investigating the
development of writing skill (for reviews, see
Berninger & Swanson, 1994; Swanson &
Berninger, 1996). By contrast, the precise
relationship between working memory and the
writing processes has been investigated only
recently. This can appear surprising given the
complexity of writing (see introduction). However,
this can be understood by looking closely to the
theoretical frameworks of writing that were proposed
in the literature: the implication of working
memory in writing was not sufficiently portrayed
and was still unspecified until 1996 where two
majors contributions on the role of working in
writing were published (Kellogg, 1996; McCutchen,
1996).
McCutchen (1996) adapted Just & Carpenter's
(1992) capacity theory of comprehension to account
for developmental and individual differences in
writing skill. Accordingly, the more efficient the
writing processes, the less they require resources
from working memory and the more resources are
available for activating other processes and for
coordinating goals. She surveyed studies bringing to
the front capacity accounts of writing. She reviewed
correlational and experimental studies showing that
a greater efficiency in the writing processes, in
particular in terms of fluency of operations, results
in a better management of working memory
resources and beside, in a better writing
performance. Surprisingly, among all the studies
she reported, only a few used a dual-task paradigm
(for example, Kellogg, 1987; see below).
The second influential contribution was the
model of working memory in writing that was
proposed by Kellogg (1996) that adopted Baddeley's
multi-component model of working memory.
Kellogg (1996) justified this choice arguing that
Baddeley's model has already been extended to
work on language production and comprehension
(Gathercole & Baddeley, 1993).
Further, Baddeley's model of working memory
has been influential in cognitive, developmental and
comparative psychology, as well as in neuroscience.
It stimulated and still stimulates a extremely large
number of studies and it accounts for a wide range
of data (Andrade, 2001; Baddeley, 2001).
In his model, Kellogg analyzed how the working
memory systems support writing and he clearly
delineated the relationship between each writing
process and the phonological loop, visuospatial
sketchpad and central executive. Like McCutchen
(1996), Kellogg first pointed out the capacity
limitation of working memory and the fact that all
writing processes place heavy demands on the
central executive (considered as a pool of cognitive
resources). Second, it analyzed how code-specific
components of working memory (the phonological
loop and the visuospatial sketchpad) are differently
engaged by the writing processes. Accordingly,
because in planning writers visualize images,
organize diagrams and plans, planning processes
require access mainly to the visuospatial sketchpad.
By contrast, both translating, reading and surely
editing impose large demands on the phonological
loop (for a review, see Alamargot & Chanquoy,
2001, chapter 5). Finally, the execution component
engages none of the slave systems. By contrast
with McCutchen proposals that focused on
individual differences and development of writing,
Kellogg's model aimed at describing the demands
of writing on working memory in adult and expert
writing. This model appeared heuristic and many
experiments, whose majority used the dual-task
paradigm, have now been conducted to test its
assumptions.
In this conceptual framework, how working
memory intervenes in writing can be studied
according to two points of view. Firstly, it can