Whereas the IP model
found considerable scholarly agreement over the initial stages
of information processing, significant inquiry has been debated
over the mechanics of the last stage, long-term memory. After
three decades, how humans retain and retrieve information is still
a hot topic, but recent emerging models have synthesized of the
convergent principles from earlier theories.
Connectionist theory, for example,
the recent but highly contentious theory many scholars view as
an alternative to classical IP models of cognition, is consistent
with the most recent neurological studies of the human brain,
as well as artificial intelligence research. The theory posits
that information is stored in multiple locations throughout the
brain in networks of connections--the more connections there are
to a concept, the more likely it is to be remembered.
Earlier cognitive models, especially schema
theory, have influenced the development of connectionist theory
significantly. Schema theory evolved as an extension of many associative
memory theories, including David Ausubel's
meaningful receptive learning, which proposed a structural arrangement
of knowledge where learners subsumed
new knowledge into existing cognitive structures. Effective teaching
then, according to Ausubel, required the use of an advance organizer
to activate higher-level (prerequisite) cognitive structures--presumably
shared by all learners, before presenting new instruction. This
notion of sequencing was extended
by Reigeluth's Elaboration
theory, which emphasizes the arrangement of instruction by increasing
order of complexity, thereby enabling learners to establish meaningful
contexts into which new information may be assimilated.
Schema theorists criticize the rigidity of Ausubel's model, especially
the presumption that information is stored in hierarchic
taxonomies, and propose instead that knowledge is stored as propositional
networks, where specific information units are not internalized,
only their general meaning. Meaning,
thus, is stored in memory as a set of relationships.
Schema theorists also criticize Piaget, arguing that there is
not just one body of knowledge available to learners at any given
developmental stage, but rather a network of context-dependent
schema that learners apply to different situations. It is this
situational specificity that distinguishes expert
from novice--the more experience
a learner has with a subject, the more developed their schemas
are, the more likely they are to function successfully than novices
with no schema or inadequate schema.
Connectionism attempts to explain human cognition using artificial
neural networks, which are computer-based
models of the brain that allow direct observation of the strength
of relationships between what was previously called schema.
There are no discrete representation units in connectionist models,
such as the imagens and logogens of Paivios Dual-Coding theory;
instead knowledge is viewed as a distributed
representation of patterns of activity.
Thinking about surfing does not activate your watersports channel;
rather, it activates multiple associated units distributed across
your neural net.
Artificial Intelligence research has validated this distributed
model convincingly, and it is important to note the challenge
it poses to classical cognition where thinking is a staged phenomenon
(encode-process-store). The theory continues to be hotly debated
in scholarly literature.
|