known as contention scheduling, which selects one of the available schemas
on the basis of environmental information and current priorities. There is
generally more conscious awareness of the partially automatic processes
involving contention scheduling than of fully automatic processes. Finally,
there is a higher level control mechanism known as the supervisory attentional
system. This system is involved in decision making and trouble-shooting, and it
permits flexible responding in novel situations. The supervisory attentional
system may well be located in the frontal lobes.
Section Summary
The theoretical approach of Norman and Shallice (1986) incorporates the inter-
esting notion that there are two separate control systems: contention schedul-
ing and the supervisory attentional system. This contrasts with the views of
many previous theorists that there is a single control system. The approach of
Norman and Shallice is preferable, because it provides a more natural expla-
nation for the fact that some processes are fully automatic whereas others are
only partially automatic.
Automaticity as Memory Retrieval
Logan (1988) pointed out that most theories of automaticity do not indicate
clearly how automaticity develops through prolonged practice. He tried to fill
this gap by making these assumptions:
.Separate memory traces are stored away each time a stimulus is encoun-
tered and processed..Practice with the same stimulus leads to the storage of increased infor-
mation about the stimulus, and about what to do with it..This increase in the knowledge base with practice permits rapid re-
trieval of relevant information when the appropriate stimulus is presented..‘‘Automaticity is memory retrieval: performance is automatic when it is
based on a single-step direct-access retrieval of past solutions from mem-
ory’’ (Logan, 1988, p. 493)..In the absence of practice, responding to a stimulus requires thought
and the application of rules; after prolonged practice, the appropriate re-
sponse is stored in memory and can be accessed very rapidly.
These theoretical views make coherent sense of many of the characteristics
of automaticity. Automatic processes are fast because they require only the re-
trieval of ‘‘past solutions’’ from long-term memory. Automatic processes have
little or no effect on the processing capacity available to perform other tasks
because the retrieval of heavily over-learned information is relatively effortless.
Finally, there is no conscious awareness of automatic processes because no
significant processes intervene between the presentation of a stimulus and the
retrieval of the appropriate response.
In sum, Logan (1988, p. 5l9) encapsulated his theoretical position in the fol-
lowing way: ‘‘Novice performance is limited by a lack of knowledge rather
than by a lack of resources.... Only the knowledge base changes with prac-
tice.’’ Logan is probably right in his basic assumption that an understanding
388 Michael W. Eysenck and Mark T. Keane