Bernard Baars’ Global Workspace Theory (GWT) of consciousness is arguably the “hottest” theory of consciousness on the market right now. The essence of the theory is that most mental contents are nonconscious and localized to specific sensorimotor circuits e.g. a nonconscious visual mental content is primarily localized to the visual cortex. In order for a mental content to be “conscious”, the GWT says that the localized nonconscious content must be made available to a globally distributed neuronal workspace that integrates and associates that content with other localized content to form a unified, multimodal representation that is our conscious experience and which enables certain functions which can only operate on the basis of such global information. For example, when a human becomes conscious of a blue coffee mug, the GWT says that the localized circuits specific to different sensory modalities must become “globally available” for use by a distributed network such that each different sensory modality becomes “unified” into a conscious percept. The key idea of the GWT is thus the transition from nonconscious, local information to conscious, global information. The global information is said to now be in a “global workspace” whereby it can enable functions like (1) the integration of novel information into prexisting circuits (2) working memory functions such as inner speech and visual imagery (3) diverse kinds of learning (4) voluntary control enabled by conscious goals (5) access to the autobiographical self and self-referential articulation, etc.
The GWT has been born out by numerous experiments whereby consciously reported percepts are correlated with a wider, globally distributed brain activation whereas nonconscious percepts are correlated with less widely distributed activation in the localized sensory cortices. But as a philosopher, I am interested in questions of ontology. A question that interests me about the GWT, and which seems to have garnered little attention from proponents of the GWT, is whether or not the global workspace really exists. Now, in one sense of the term, the global neuronal workspace really exists insofar as there really is a “widely distributed network of brain circuitry” that has the functional properties associated with consciousness. But notice the terminology of the global workspace. What is the nature of this “space”? Is the space inside our skulls? Can we open up the brain and point to this workspace? Evidently not, for the “workspace” seems to refer to a functional space that is only experienced as a literally internal space. The only real “space” in the brain is the physiological space of brain tissue. But if you described to a layperson the nature of working memory such as inner speech and visual imagery, and then asked them “where” in space such functions are located, they would likely tell you the workspace is inside their heads. Now ask them how they know this and they will likely tell you they know it because they experience inner speech and visual imagery as taking place inside their heads. But as philosophers, we should be skeptical about deriving the reality of the workspace from our normal experience of the workspace. And it is this distinction between the “real” global distribution of neuronal circuitry and how we experience the workspace that captures my worry about the “reality” of the workspace. For it is one thing to say that there is a global distribution of neuronal activity in the brain, and it is another to say that there really is a “spotlight” in our heads that “shines an attentional beam” onto a theater in our brains.
Proponents of the GWT far too often fall into the trap of confusing the metaphorical nature of such theater talk with the “globally distributed” talk of neuron populations. It is important here to keep distinct what we are trying to explain with the GWT and the mechanisms which we use to explain that phenomenon. What we are trying to explain is the experience of ourselves as having a unified conscious experience that is coherent, stable, richly detailed, and continuous across time. This is what needs explaining. But a neuronal explanation of this phenomenon based on global distribution of neural information is fully compatible with the unified conscious experience being an illusion. I think this is a point that Dennett has been trying to make for decades. But when most people hear the word “illusion” they think that Dennett is trying to dismiss the phenomenon trying to be explained, namely, unified narratologically continuous experience. But what Dennett is actually trying to do is to get people to realize that one of the capacities of the human brain is indeed the capacity to generate illusions. And the “unified” Cartesian theater “in our heads” is the biggest illusion of all. For it is equally possible that those globally distributed neural mechanisms could make us experience ourselves as being outside our heads.
That this is true is evidenced by both clinical and experimental evidence. On the clinical side, we have reports of out of body experiences. Julian Jaynes recounts such a report:
That there is no phenomenal necessity in locating consciousness in the brain is further reinforced by various adnormal instances in which consciousness seems to be outside the body. A friend who received a left frontal brain injury in the war regained consciousness in the corner of the ceiling of a hospital ward looking down euphorically at himself on the cot swathed in bandages, Those who have taken lysergic acid diethylamide commonly report similar out-of-the-body or exosomatic experiences, as they are called. Such occurences do not demonstrates anything metaphysical whatever; simply that locating consciousness can be an arbitrary matter.
On the experimental side, we have people like Thomas Metzinger who can induce out of body experiences through virtual reality setups where you are given visual feedback through a camera of your own backside. After adjustment, your brain will make you feel like you really are floating outside your body. This tells us that the “location” of the experential global workspace is more or less arbitrary. If the brain wanted to, it could make you feel like the global workspace exists 3 feet above your skull. Now, of course, there are good reasons why the brain generates the illusion as taking place in your brain, for this is tied into the closeness of volition and internal sensations with our bodily experience. But it is important to remember that thousands of years ago the “internal mind-space” was experienced as being in the heart, not the head.
So the moral of this post is that if we are going to develop an adequate scientific theory of consciousness like GWT, we must be clear on the distinction between what is being explained and the mechanisms we posit to explain the phenomena. The phenomena to be explained is the experience of a unified Cartesian theater in our heads. The explanation is a global distribution of localized, nonconscious information. But when asked whether the global workspace “really exists”, we have to be clear between the workspace as experienced by us and the workspace as hypothesized by science. As we experience it, the “location” of the workspace inside our heads is arbitrary. As we explain it, the location of the workspace is the precise network of globally distributed brain* activity. So does a global workspace really exist? Yes. But it exists both as an illusion we experience, and as a brain distribution.
*I say “brain” activity and not “neuronal” activity because there is growing evidence that astrocytes play a role in modulating neuronal information processing through modulation of glutamate uptake in the synaptic cleft.