In the classical definition, automatic processes, as pointed out by Posner (1978) and Hashner and Zacks (1979) are effortless, unconscious, and involuntary. However, Neumann (1984) and Carr (1992) pointed out that there are rare cases where all three features hold simultaneously in an automatic process. Others have introduced a minimalist (common feature) definition of automaticity. Logan and Cowan (1984) introduced the concept of ballisticity. A process is said to be ballistic if it completes its task, once started, without involvement of conscious control (an act of will). This identification of conscious control with an act of will confounds consciously mediated action with voluntary action. Bargh (1992, 1994) pointed out that ballisticity is common to all automatic mental process and proposed that it be adopted as the definition of automaticity. Logan (1992) argues that a common feature for automaticity is that it depends on attention and is a postattentive process. We agree with Logan’s view. Each step of a novel task (before it is automatized) is processed with the involvement of consciousness. In global workspace theory, attention, which is initiated voluntarily or automatically, controls access to consciousness. At any given point in the processing of a
task, attention brings the necessary information content (knowledge, skill, etc.) to consciousness. The consciousness processes use the information content to recruit the necessary computing resources to perform the next step of the task. It is our view, whether a task is automatized or not, that a task is initiated with the service of the same attention system (an attention codelet). In our mechanism, a task is operated in the same sequence using the same set of computing processes (codelets or low-level action units) whether it is performed automatically or under conscious control.
Tzelgov (1997) proposed that automaticity be defined as processing without monitoring in the sense that monitoring is the intentional setting of the goal processing and the intentional evaluation of its output. We do not have a disagreement with this definition, except that we qualify monitoring to be conscious monitoring or consciousness. Also, we believe that an automatized task can be initiated voluntarily or intentionally. This comes from our assertion that if a task is initiated as a result of a voluntary attention, then the same voluntary attention (intentional setting) is needed to initiate the task after
automatization. That is, once automatized, the triggering intentional setting (“what to do”) is always consciously experienced while the triggered goal structure (“how to do”) that produce the actions is automatic.
In agreement with Logan’s definition, automatization in our mechanism is dependent on attention and, as such, is a postattentive phenomenon. In our mechanism, different attention components, in competition with each other, try to bring the corresponding pieces of information content (knowledge, skill, etc.) to consciousness. The attention
component that wins the competition brings its content to consciousness to be broadcast to other processes. Our belief is that automatization, as a postattentive process, reduces the competitive strength of the relevant attention components. As a result, the role of consciousness fades in the processing of the task. Automatized tasks are not affected by the serial and limited-access constraints associated with consciousness, and they can be performed in parallel.
Bargh (1997, 1999) stressed the importance of automatization, stating that much of psychological processing (perceptual, cognitive, social domains and implicit learning) is automatic. Automatization frees up one’s limited conscious attentional capacity/resource from automated tasks in which it is no longer needed (Barge, 1999 and many other before him). Automatization is itself implicit learning, which takes place automatically without our conscious awareness. Although implicit learning is automatic, it requires conscious information (Reber, 1989). Conscious regulation of one’s decision and action is effortful, requires limited resources that can be exhausted very quickly, and is a serial process that is relatively slow. In contrast, automatization allows an effortless, relatively fast, and mostly parallel control of one’s own behavior and decisions. If we were conscious of everything that we do in our every day regular activities, we would not have the time and resources to perform tasks that are deemed important. By not bogging down
consciousness with what we call trivial routine tasks, the freed consciousness is used to solve problems requiring informational input and for creativity. Without automatization, no task will be trivial.
One disadvantage of automaticity is that it is not easy to turn off automatized skills and automatized processes could produce action errors by being active in the wrong contexts.
Also, arguably a disadvantage, we lose the details of our automatic skills. If needed, the steps in automatized tasks can be observed voluntarily (be aware of what we do
automatically), but this defeats the main advantage of automaticity: to save the important resource of consciousness. So, automaticity is inherently inflexible.