Cognitive Theories of Consciousness
V. de Gardelle, S. Kouider, in Encyclopedia of Consciousness, 2009
Attention and the Central Executive
Various influential models developed in the 1960s referred to a central processor, a central executive system, or a supervisory system. Processing within the central system can be considered as analogous to conscious processing, even if the word consciousness was still largely banished in the scientific community. This system is at the top of the hierarchy in the cognitive architecture: it is involved in higher-order computations (decision, monitoring, planning, etc.) and leads to selection and control over lower-level subsystems. As in many contemporary accounts of consciousness, the central system was considered the most integrative element of the cognitive system, granting flexibility and control over behavior.
Another key element was the simple but powerful metaphor of attention as a filtering mechanism that was put forward by Broadbent. In a nutshell, peripheral processors in this theory provide sensory information to the central system dealing with control and decisions. Because multiple sensory channels are continuously acting in parallel, a huge quantity of information becomes available to the rest of the system. However, the central system is very limited in terms of computational resources. Hence, a selection mechanism is needed to prevent overload. As such, attention operates by selecting the most relevant information and by filtering out that which is irrelevant. Then, the most relevant information, which is under the focus of attention, becomes the target of the central system and can thus benefit from deeper and more enriched processing. Once again, although consciousness was not the main concern, one consequence of attentional selection was that it allowed the target information to become conscious. In this perspective, attention and consciousness are two tightly related notions.
The notion of short-term memory put forward by George Miller and later extended to the notion of working memory is also an important precursor. For example, in their model of working memory, Baddeley and Hitch relied on a central executive system, which has top-down control over the distinct specific subsystems, namely the phonological loop and the visuospatial sketchpad. Here, the content of working memory may be roughly equated with the content of consciousness, an aspect that will also be important for future cognitive theories of consciousness.
Norman and Shallice, in turn, proposed a model of action selection implicating a supervisory attentional system. This central system receives sensory evidence and determines the appropriate behavior by selecting instruction schemes for action mechanisms. In addition, the supervisory attentional system can be modulated by the goals of the organism, and it is primarily involved when a new or critical situation appears. Here too, the central part of the model shares some properties that are associated with consciousness, namely flexibility, reactivity regarding unexpected situations, decision, and control over behavior.
In sum, these influential early models depicted the global architecture of the cognitive system by emphasizing the following components: sensory inputs in the periphery that are processed in parallel in multiple channels, attention that performs selection upon these sources of information, a working memory component that keeps tracks of the selected information, and finally a central system that acts as a supervisor. But one major limitation of this view is that it falls into the homunculus trap, when it comes to the question of consciousness. Indeed, if this central supervisor is governing the whole cognitive system, one may ask who is in turn governing the central supervisor! That is, if we were to rephrase this question by focusing on consciousness, it would be problematic to rely on a hypothetical little man in our head (i.e., a homunculus) that has consciousness, which is the same property we are supposed to explain. This approach unavoidably leads to an infinite regression. Because consciousness was not the main issue for these early models, this crucial issue was left out or even denied during the development of early cognitive models with a central supervisor. As we will see below, current theories of consciousness will overcome this limitation by proposing various cognitive architectures, sometimes including a central system, that take into account the homunculus issue.