Processing Novel, Yet-to-be-Learned Material

Processing Novel, Yet-to-be-Learned Material
In contrast to a traveler at the highly learned end of the learning continuum,
consider someone at the other end of the continuum, represented by
the left side of the matrix of continua of Fig. 1. This person is traversing the
route for the first time and so has few or no schemas to coordinate
the elements of information. There is no well-defined, schema-based
central executive to deal with the information. In the complete absence
of a schema-based central executive, problem-solving search to ascertain
a suitable route will be required. As indicated earlier, when engaged in
problem-solving search, at certain points there is no choice but to
combine and test elements randomly. In this particular case, that
requires choosing roads on a random basis and testing the consequences
of the choice either mentally or physically. That means while we can
consider the consequences of choosing a particular direction, we can only
do so after deciding to consider that direction, not before. In the absence of
knowledge, the decision to choose a particular direction for consideration
must be random. More frequently, partial executive functions can be
provided by other sources (e.g., a map) and, indeed, precise, ongoing
instructions from someone else can provide full executive functions.
Nevertheless, in the absence of suitable domain-specific schemas to
coordinate elements of information, the person normally will need to
engage in problem-solving search using a general problem-solving strategy
such as means-ends analysis. Using this problem-solving strategy, the
problem solver must attempt to find problem-solving operators that will
reduce the diVerences between the current problem state and a goal or
subgoal state. These operators must be chosen randomly but can be
tested mentally for their consequences using means-ends analysis, a process
that is very expensive in terms of the limited working memory resources
available at this extreme of the matrix of continua (Sweller, 1988).
The left side of the matrix of continua applies to a wide variety of
intellectual tasks. When listening to or reading unfamiliar, high element
interactivity material, various aspects of the material need to be related in
order to derive meaning. If the relations are not incorporated in schemas,
they will need to be processed in working memory, which will require a
234 John Swellerproblem-solving process to determine which relations are appropriate.
Initial attempts to establish connections between referents, for example, will
contain random components and so some attempted relations will be
inappropriate and fail, resulting in a comprehension failure. To understand
the statement ‘‘Initial attempts to establish connections between referents
will contain random components,’’ the listener or reader must establish that
‘‘random components’’ refer to the ‘‘attempts’’ and not the ‘‘connections’’
or ‘‘referents’’ directly. To understand text, the number of such attempted
relations must be limited in order to prevent a numerical explosion of
possible relations that would permanently prevent comprehension. A
limited working memory reduces the number of possible relations allowing
the prospect of comprehension. Nevertheless, if there are too many possible
relations not previously incorporated in schemas, comprehension will fail
(e.g., Britton & Gulgoz, 1991). In contrast, schematic control determines
which relations between interacting elements are appropriate and embeds
them within schemas. A schema for a statement includes all of the
interacting elements within it and can be processed readily in working
memory. As a consequence, large amounts of information can be processed
with a limited working memory load, allowing very complex relations to
exist and thus ensuring comprehension