Above is a high level picture of the structure of the mind. The diagram and its commentary are derived from two primary sources: the global workspace theory of consciousness, and Kant’s view of synthesis as explicated in the chapter “The Deduction of the Pure Concepts of the Understanding” (A95–130) in his Critique of Pure Reason. My own views play a role too. And, I connect the diagram to my earlier thoughts as described in “Consciousness and Mathematical Realism”. I do not claim (of course!) that the diagram is exhaustive, only that it lays out a core set of mechanisms and phenomenal contents.
Phenomena are the things that appear, or can appear if attention is turned towards them. The executive performs the sequential top-level activity of the conscious mind. At any moment, it directs attention towards some particular phenomena, and queries information that is relevant to the current task. The contents of consciousness consist of those phenomena towards which attention is directed.
As diagrammed, the phenomenal realm contains representations (in Kant’s terminology) of one sort or another such as perceptions of physical objects, fantasies, memories, etc. However this realm contains nothing in the way of a process. The mechanisms, together with synthesis are responsible for populating the realm of phenomena. It is meaningful to separate the three, because consciousness has access to the phenomena, but the mechanisms and the process of synthesis are opaque to it.
Synthesis is the process by which the outputs of perception are gathered into phenomena. Consider, for example, the activity of watching a tree blow in the wind. At each moment the mechanism of visual perception produces a representation of the tree, and at each moment this representation is different. After perception, the first stage of synthesis (which Kant refers to as “synthesis in imagination”) is the process of gathering these momentary representations into a whole and identifying the separate representations as representations of the same thing — the blowing tree. The next stage of the synthesis is the binding of this whole to the Kantian category of “existence” (as opposed to, for example, “possibility”) and then to the concept of tree (from semantic memory). Finally, what I am calling synthesis locates the tree within the realm of all phenomena.
And what is “a location within the realm of all phenomena”? The mind considers not just one world, but many possible worlds: fictional, anticipated, speculated, etc. The first aspect of a location is: in which possible world (“actual world” being an option) does the phenomenon reside? In Kant’s system of categories,this attribute fits under the category of “modality”. Then within that possible world, where in space and time does the phenomenon occur, and what other identifying attributes does it possess? In the case of direct perception, the possible world is “actual, now”, and the spatial position is given in qualitative terms relative to other objects and to the observer, and will vary greatly in precision. In the case of memory or anticipation, time as well as position needs to be specified (again in relative, often imprecise terms). In these cases, the mechanisms of episodic memory or scenario construction are involved, rather than perception followed by synthesis . Fictional worlds come with their own systems of space and time. So, if I am watching a Star Wars movie, my mental representation of the death star during a particular scene will be located within the space and time of the fictional world associated with that movie. If I am a detective, considering various scenarios for how a crime might have been committed, I bring to mind possible objects and events associated with each scenario — and these phenomena too have a position and temporal aspect, but in this case within the same system of space and time as pertains to the actual world. Anticipation and planning involve bringing to mind various possible future worlds.
Consider your own experience. If it is like mine, you will find that phenomena “just show up” in integrated and located form — demonstrating that the mechanisms involved do their work invisibly to consciousness. Each and every phenomenon, produced by whatever mechanism, takes its unique place within the overall structure of all active phenomena organized by what I have called location. (They are organized also by “identity tagging”, so that mechanisms can keep track over time of their outputs-but these are weeds I’ll not dive into.) It is in this sense that consciousness is unified. This is the structure known as “the global workspace” in the eponymous theory (I treat the terms “global workspace” and “realm of phenomena” as equivalent).
The contents of the workspace change continually. Perception and the other mechanisms (driven by commands from the executive) are always introducing new phenomena or modifying existing phenomena. The lifetime of any given phenomenon is only a matter of seconds if it is not refreshed by its mechanism or brought to attention.
Perception and synthesis involve parallelism, and other mechanisms might be active at the same time. That is, many phenomena might be entering the workspace simultaneously. But only a few can be attended to at any given time, and only those in the focus of attention enter full consciousness where they can be queried by the executive, or remembered. However, the rest take their place in the global workspace, where they are potential targets of attention.
Notes about Kant’s Terminology
What I have labeled “perception” in the diagram, Kant calls “intuition”. Intuition produces spatial representations, but representations which appear only as individual fragments, and are bare of concepts. According to Kant, the input to the process of intuition lacks spatial and temporal structure — it is a buzzing confusion upon which the intuitions of space and time impose order. But I do not follow him in this — I take it that spatial and temporal structure exist also in the external world, though the process of representing that structure in the mind (that is the process of perception) is very complex.
The next step, in Kant’s terminology, involves the processes of “understanding”, namely the conjunction of representations into unities (as in the example of the blowing tree), and the application of categories and concepts. There is also the matter of unifying all of the many unities produced by understanding into the one grand unity of the mind’s experience as whole. This involves the concept of “I” or “self” as the common subject to which all these unities appear. Kant refers to this as “the synthetical unity of apperception”.
A discussion of the many points of correspondence between Kant’s views of the mind and contemporary cognitive science can be found here.
Among the phenomena is the self. There are three levels of representation of the self. First, the experiential self, is the bare subject of experience. This is a pure representation without other properties. It is the abstract container for the mind’s own experiences. This is the self to which Kant’s term “synthetical unity of apperception” refers.
Second is the empirical self (Kant’s terminology). This is the self which appears when the gaze turns inward via the mechanism of reflection. It consists, however, only of what appears directly to that gaze. “I am angry now” or “I am now thinking of my recent trip to Colorado” would result from that turn of gaze, but not “I was born in Chicago” or “I am a forgetful person”. Kant referred to this inward gaze as “inner sense”.
The third is what Antonio Damasio refers to as the autobiographical self (and Sartre as the transcendental ego). The autobiographical self is the self regarded as something that has properties that are stable over stretches of time, and possesses a history. It is the self viewed more or less as if it were a thing among things in the world. An autobiographical self might indeed be described as a “forgetful person from Chicago”.
The empirical self appears to the mind only when the process of reflection is active, and the autobiographical self only when theorizing about the-self-as-object, but the experiential self is present at all times — for example even when concentrating fully on external activities (surfing, dance, chess,etc), that are too demanding to allow reflection to take place at the same time.
What are phenomena?
In “Consciousness and Mathematical Realism”, I propose that the brain be viewed as connected to a set of mathematical systems, each constituting a layer of description of the brain’s activity. Where realism comes in is that these systems are granted existence in their sort of world, without the need for a human or Godly mathematician. From this point of view the diagram depicts one of these layers of description, and the phenomena in it are mathematical objects in their structural aspect. If Tononi’s notion is correct that the degree of consciousness of a system can be measured by its degree of integration (by either Tononi’s Φ or Hoel’s simpler EI ), then the layer in which consciousness occurs would be expected to have a maximal measure of integration among the layers in the stack.
However, mathematics cannot capture all aspects of phenomena. We experience qualia, such as the redness of red, or the sensation of pain. An explanatory gap yawns between any purported non-mental substrate, and the “being there” of our own awareness. In this case the proposed substrate is mathematical rather than strictly physical, but the gap remains as wide. Clearly, neither the physical nor the mathematical can fully constitute consciousness. This is a problem that our minds seem ill-equipped to ponder. I have no solution to offer.
However, I do propose that the qualia, whatever they are, are painted onto the mathematical objects which constitute the core of phenomena. Restated: The dimension of quality, whatever it may be, is attached to mathematical objects in the mind, not to the things in the physical world which, via perception, populate the mind. This is not to say that the dimension of quality might not also attach to things external to minds, but that is none of our affair, since it plays no role in our experience.
Commands to the Mechanisms
This final section, low in detail, posits a way to conceptualize the commands by which the executive controls the mechanisms.
The commands take the form of instructions to build new phenomena, or in the case of the body, to perform actions. The instructions may refer to the current contents of consciousness. For example, an instruction to episodic memory takes the form, roughly, of a search term, such as “where have I seen X before?” , where is a X is a reference to a phenomenon already present in consciousness. As a result the episodic memory mechanism populates the workspace with relevant phenomena.
Similarly, in planning, the command might have the form “send back the scenario that results from performing action A in scenario S” where A and S again refer to phenomena already present in consciousness.
The mechanisms, it is clear, are integrated with each other in many ways. For example, perception and the control of muscles are processes that talk to each other at high bandwidth during physical activities such as walking, and this chatter takes place without the involvement of consciousness. The mechanisms also have access to the phenomenal realm. However, the executive receives no input directly from the mechanisms nor the process of synthesis. All communication from mechanism to executive is mediated by phenomena.
A word about memory. At all times during normal consciousness and REM sleep, the contents of consciousness, and only those contents, are continuously encoded and stored in episodic memory. Most such memories do not survive long, of course. Retrieval is a matter of decoding and reconstruction. Reconstructions are not always accurate, and errors are propagated back into the store.