7: Positive Theory
If we attempt to develop a theory about consciousness based solely on our experience of being conscious it is unlikely that we will ever break out of the subjective realm. Now this is not to say that our subjective experience can’t help us understand consciousness, indeed I will appeal to it at several points, its just that by itself it has no necessary connection to what actually exists. Instead I will begin from an understanding of what it is required for us to think that another person is conscious.
7.1: Definition of Consciousness
Let us begin then with the hypothesis that what if another being demonstrates that they have a first person perspective, and have experiences, that we are justified in saying that they are conscious. Of course in that form our hypothesis would seem to be a kind of behaviorism, because it makes consciousness dependant on what impressions the system gives us through its behavior. So let us say instead then that the system is conscious only if it really has a first person perspective, and experience. Obviously behavior that seems to demonstrate the existence of a first person perspective and experience is probably good evidence that there really is one (certainly that is the simplest explanation), but whether the system really has such things is a matter that depends solely on its internal workings (meaning that, conversely, some systems could display no evidence of a first person perspective or experience and still be conscious, although this too is an unlikely possibility).
But, before we proceed to define what it takes to have a first person perspective and experiences let us consider first if having both those things is always enough to be conscious, or, in other words, is it possible for a system to have both a first person perspective and experiences and not be conscious? I don’t think it is, primarily because we have no evidence that we aren’t such a system. Following Descartes, all we can say for absolute certainty about our own existence is that we think (we experience) and thus that we exist. Everything else, including whether our experiences really have a qualitative feel can be doubted. But a system with experiences and a first person perspective could carry forth the same argument for its existence (assuming it had sufficient intellectual capacity). So we can’t say for sure that we aren’t such a system, and that everything else is a delusion. But we are certain of our own consciousness, so we should be equally certain that such systems, in general, are conscious (since we could very well be one of them). Perhaps some will object to this, saying that even an unconscious program could produce the statement “I think” and draw conclusions from it. With this I agree, but it is only a system with a first person perspective and experiences that is able to construct that claim from its experiences, and thus these are the only systems that can make the argument from scratch, without having the premise hard-wired into them (although, again, not all of them are sufficiently smart or sufficiently logical to actually construct the argument).
7.2: The First Person Perspective
The next task then is to determine what kinds of systems can be said to have a first person perspective. Often we talk about the first person perspective as though it were a viewpoint from which our conscious mind “looks out” on our experiences. As intuitive as this idea may seem it is also misguided. If we really thought about the first person perspective in this way we could still ask all the questions we normally ask about consciousness about the first person perspective (e.g. what makes it have a first person perspective), and thus would have made no progress. In addition such a conception of the first person perspective leads either to an infinite regress, in which the first person perspective itself must have a first person perspective, or a denial that the question can be answered. Neither response gets us anywhere.
Thus, if we are to successfully understand the first person perspective (in terms of understanding why people have it and things like rocks don’t), we must define it without invoking a perspective. This might seem impossible, since having a perspective is essential to our conception of the first person perspective (the name gives it away I think). It would be like asking someone to draw a square without drawing something that was a square. Is this really impossible? Consider the picture below:
In this picture there is clearly a square, but at the same time there isn’t. The “square” in the picture is created by the features of other elements in the picture, and not by a separate element in the picture that has the property of being square. Attempting to explain a first person perspective in terms of an inner point of view is much like attempting to explain the square in the above picture by appealing to some elements of the picture as being square, you just can’t do it. But, just as the shape of the other elements in the picture define the square, perhaps we can define the first person perspective in terms of the properties of other aspects of the mind, which create a perspective without themselves having a perspective.
Since experiences and the first person perspective are tied so closely together it is natural to assume that the structure of experience somehow creates the first person perspective. But, before we can adopt such a view, we must address two possible problems. One is that we usually define experiences as being had by some consciousness, which in turn requires a first person perspective. This might seem like it would make determining which systems have a first person perspective impossible, since to see if they had a point of view we would need to know that they had experiences, which in turn would require that we already know that they have a first person perspective in order to know that they were conscious. The other potential problem is that we haven’t specified how experiences give rise to this first person perspective, and a failure to do so may leave some suspecting that it is impossible.
Let me first tackle the “chicken and the egg” problem. Although it may seem intimidating there is really a simple solution, instead of saying that the first person perspective is created by the structure of experience we can say that it is created by the structure of proto-experiences. A proto-experience is something that fills all the requirements to be an experience but isn’t necessarily part of a conscious system. Of course if the proto-experiences do contain a first person perspective then they can be considered real experiences, and if they don’t obviously they remain as only proto-experiences. This neatly resolves the problem, since we can still claim that all experiences necessarily have a first person perspective, but we don’t have to stipulate that something is a full-blown experience before we go looking for a first person perspective in it. I will handle the exact nature of experience, and thus these proto-experiences, in section 7.3, so for now let me just assume that a good description of them does exist.
The second problem, explaining how experiences create or are responsible for a first person perspective, is a bit trickier. The first possibility to consider is that the first person perspective is created by some specialized part of experiences, but given how we are aware of the self this seems unlikely, since unlike vision or hearing there is no specific “mode of experience” that is associated with the first person perspective. It is reasonable then to assume that the first person perspective is something that all aspects of experience contribute to. Experiences, as we know them, are not simply information gathered from externally directed senses, even if we set aside the internal components of experience such as thought. Our experiences are filled with concepts and additional information by unconscious processes, for example when we see a tree we see it as a tree in addition to seeing it as specifically colored shape. My proposal, then, is that the first person perspective is created by information “added” to experiences that relates the contents of perception, thought, ect to a subject. Objects seen are seen as being in a certain position to the perceiver, feelings are construed as my feelings, thoughts as my thoughts, ect. The information about the perceiver, the me, the I, that is added to experiences is responsible for our observation that a first person perspective exists, since in any conscious mental act information that there is a subject, that the experience is from our point of view, is also presented to us. And because information about a subject is embedded in the experience itself, and it is this information alone that is responsible for the first person perspective in consciousness, there is nothing more to the first person perspective, just as there is nothing more to the square in our earlier picture besides the shape of the black areas.
I admit that, properly speaking, experiences are only found in conscious systems (or at least that is normally how we use the term). So what we really want to define is a proto-experience, which, when combined with the kind of first person perspective described above, results in a conscious system, and is thus a real experience. Initially we might simply think that input in general is what I am calling proto-experiences. Input into an unconscious system, like a thermostat, is simply data, while input into a conscious system is an experience. Unfortunately things aren’t so simple. Consider, for example, the phenomenon of blind sight. A patient with blind sight has lost the ability to have the experience of vision at least for part of their visual field. Consciously they are blind in that area, but the information is still reaching their mind; when guessing (randomly, or so they think) they demonstrate that they have, unconsciously, some information about what going on in that region, as they “guess” spectacularly well. Blind sight then is an example of input into a conscious system that is not part of an experience.
Obviously in blind sight the input only goes to the unconscious mind. But this does not answer our question, about what a proto-experience is. We might be tempted to say that it is input into the conscious mind, but the unconscious is part of the same system as the conscious mind, so what makes one kind of input a proto-experience and not the other? The key, I think, is how the information contained in the input is made available to the mind, and, more importantly, future mental states.
Consider, for example, a memory suddenly recovered of an event that one never had a memory of before (perhaps not a condition that occurs in real life, but definitely something that is possible). After such a recovery we would have memories of a time that we didn’t before, and of course, by the nature of memory, such events would be remembered as being experienced by us from a first person point perspective. Was this period of time conscious all along, then, and simply hidden from us, or did it become conscious upon our ability to remember it? Accepting either interpretation of events seems to lead to unacceptable consequences. If we accept that it became conscious when we were able to remember it we seem to be accepting that an experience being conscious to a subject is something that can be decided years after the subject actually experienced it. On the other hand, arguing that the remembered experience was conscious all along would seem to undermine our first person authority about consciousness, which is equally strange since consciousness is usually understood as the one aspect of the mind that we have authoritative first person access to. Even stranger, if we accept that a recovered memory was conscious all along then how are we to handle false memories or “borrowed” memories (the possibility of one person’s memories being given to another person)? It would seem strange to say that since they can be remembered as conscious they must have been conscious all along, since in the case of a false memory it never really happened, and in the case of borrowed memories it happened to someone else, not the person who it is now conscious for. (Of course there is also the option to deny that remembered events are necessarily conscious in any way, but this would lead to a strange position where one could rationally doubt that they were conscious before the present moment, and clearly this is an absurd position to be in.)
The underlying problem is not how we approach remembered events, but how we deal with the conscious subject. Because we naively assume that during a lifetime there is one conscious subject we assume that every experience must be conscious for that same subject, leading us to the problems above. However, it is well known that considering people as some single entity that exists for a lifetime (the person at one time is the same person at some later time) is paradoxical (because of branching and transitivity). One solution to this problem is to consider each instant to be a single person, who is very similar to the people in that body at subsequent and preceding instants, and less and less like the people at instants further away in time. In other words, we could call James now James(A), James five seconds later James(B), and James five minutes later James(C). James(A) is very similar to both James(B) and James(C), but he is more similar to James(B) than to James(C). The degree of similarity itself is determined by how many psychological (or mental) properties the person-instants have in common. Under this view the idea that there is a single person at all times who lives in the same body is simply a rough abstraction that arises because a person at any two moments in their life is generally more similar to “themselves” than to other people.
If we accept this analysis of personhood then wondering about what makes an experience conscious to any future person-moment becomes a much simpler task. If that experience can be remembered or reflected on by at least one future conscious person-moment then it is conscious with respect to that person-moment (and whether other future person-moments can remember it or not becomes largely irrelevant). A tentative hypothesis about proto-experience then is that they are the collections of inputs that can be remembered and reflected upon by future states (although not that it necessarily must be reflected upon or remembered, just the potential is sufficient). If this hypothesis were to be false it would mean that it would be possible for a system in which no experience could be remembered or reflected upon by any future moment could be considered conscious, and I will simply take it as a given that this is an impossibility, since they would seem too unstable to support a first person perspective. Of course I am sure that someone will bring up the cases of people who can’t form new memories as an exception, but they aren’t really, since even these people have short term memories, meaning that they can think about and build upon their experiences for at least a short period of time.
Reflexes are a good example of how experience must be incorporated in a way that allows it to be reflected upon for it to be conscious. The input that causes us to, for example, pull back from something hot does not immediately become part of the conscious mind. We react before we are aware of why we are reacting. And thus the input that caused us to pull away is effectively unconscious and not part of an experience, as it immediately causes us to pull away, and is not incorporated into the mind in a way that allows it to be reflected upon. Of course that input does become part of an experience after the reflex has occurred, but until it does, while the reflex is occurring, it is unconscious, simply input.
In addition the input has to be incorporated into a reflexive structure in the “right way” in order to constitute a proto-experience. The “right way” for a particular kind of input, say sight, is that when that input is processed, it is treated as sight. For example, if a system has visual input from a tree, and if that visual input is to be part of experience, then any thoughts or responses that the system generates as part of the circular process that defines experience must be about it as visual, not as a hunch, or smell, or whatever. Or, in other words, the input-output correlations that I appealed to earlier as the basis of intentionality must be preserved, input of one kind can’t be recalled as input of another, and it must be the input itself that is stored, not its contents. What the difference might be in simple creatures is hard to imagine, but for people it is clear, it must be thought about, in every way, as vision. This is why the visual input in patients with blind sight is not a visual experience, even when they are correctly guessing about the contents of the region they cannot consciously see; that information is incorporated into their consciousness as a random guess, and not as visual input. When they reflect upon it they don’t think of it as sight, they think of it as a guess. Thus they have the experience of guessing, but not the experience of seeing, even if their guess and the visual input contained the same information.
So, in summary a proto-experience is input into a system that is unified, identified as input, and that is made available to the system not just when the input comes in, but also for some time afterwards. This means that a proto-experience will probably be incorporated into the immediately subsequent proto-experiences to a greater or lesser extent, which is probably corresponds to our awareness of the moments that immediately precede the current one.
7.4: What Kind of Theory is This?
That then is my theory about consciousness. So what kind of theory is it? Well certainly it is a materialist theory, since the explanations of both the first person perspective and experience are in objective terms. More specifically they are in terms of how information is handled by the system. To generate a first person perspective the system must maintain information about itself, and incorporate that information into experiences. And experiences themselves are simply input, a kind of information, unified and made available to the system not only when the inputs arrive but also when desired at later moments (that is, if something doesn’t cause the experience to be forgotten, although all experiences seem to make it into the short term memory, where they have the opportunity to be reflected on, at least as retentions in the margins of experience, for some time), The theory above then ultimately defines consciousness in terms of how information flows in the system, a view that might be described as informational functionalism. And of course how information flows in a system is ultimately determined by the basic physical properties, although systems of different structure may handle information in the same ways.
7.5: The Domain of Consciousness
I think then that I have at least addressed the central questions about consciousness, namely what it is and why some systems have it. As a closing remark allow me to touch on a question that always seems to reappear in the debates surrounding consciousness, namely whether animals and computers can be conscious. Certainly I see no barriers, in principle to either group possessing consciousness. However, many animals may not possess consciousnesses simply because their brains aren’t structured in the right way (fish and insects for example seem too simple to be conscious). But, many animals, especially some of the mammals, do seem sophisticated enough to be conscious, although we might not know for sure until we have the technology to study in detail the operation of their brains (since they can’t talk). Similarly a computer programmed in the right way might be conscious. But I don’t see what we would expect from one, because consciousness is not the same thing as intelligence, and I think the first goal of the computer scientists working in this area is to make a computer that acts intelligently, not one that is conscious, since consciousness by itself wouldn’t be very useful.