Qualia is a word invented to describe the character or “feel” of an experience. For example, when you see a red object you experience the qualia of that shade of red, which seems to be a part of your experience that exists in addition to the red light entering your eyes and the surface properties of the object. The qualia thus seem to be best described as a feature of experience, a way of describing it, and not a part of the natural world; while the natural world is made up of objective facts qualia are inherently subjective. Of course we would like to say that the two are in some way connected, ideally by explaining how qualia reduce in some sense to objective facts, but it is not obvious how. And thus by being inherently subjective qualia pose problem for materialist theories about the mind. If they were based on the physical facts then shouldn’t we be able to take our materialist theory about the mind, say about vision specifically, and use it to explain why red feels like red and not green to a particular subject? But so far none of our physical theories have been able to give us such an explanation, it seems like they might never be able to. Certainly our theories will provide explanations of why someone describes an object as red, why they behave as though it were red, and possibly even why they think about it as if it was red, but what is wanted is an explanation of why it “feels” red. And because we can’t give an explanation of why it “feels” like red some take this as evidence that qualia exist in addition to the physical world, as something extra.
4.1 Mary, The Economics-Blind Social Scientist
Before I address qualia directly let’s consider a thought experiment that shares some features in common with the ideas that lead us to reject the possibility of qualia being explained in physical, objective, terms. Consider a social scientist, Mary, who is completely cut off from contact with other people, she does not interact directly with them in any way. Thus Mary does not share our language, our social conventions, ect. For the sake of convenience let us assume that Mary has her own language. Now let us suppose that Mary is given access to all the basic physical facts about a society. She knows the location of various items and people, and how they move over time. For Mary’s convenience we can grant her the ability to zoom into different levels of detail, observing, if she chooses, how individual atoms interact with each other, or choosing to see how whole objects move around, or anything in-between. Because we don’t want Mary to adopt our ideas let us give this knowledge one limitation, that she isn’t allowed to know any facts about our communication using language, either written or verbal.
Let us assume that this society is our society. Does Mary know that the GDP is rising? No, Mary does not, and not because she is ignorant of linguistic facts, because whether the GDP is rising or falling is independent of what we say. Certainly Mary can observe all the individual transactions that we think of as economic, but she doesn’t think of them as economic, because she doesn’t share our economic conceptualization of the world, no one has ever told her what money is or how it works, and certainly they haven’t told her how such transactions can be aggregated into the measure we refer to as the GDP. Of course it is reasonable to suppose that Mary has her own theories about why certain individuals and certain items are together at some times and apart at others (i.e. ownership has been transferred), but they are not in terms of a larger economic system.
But we could give Mary an economics textbook (translated into her language), and then she would be able to deduce that the GDP was rising, as well as a large number of other facts about the economy. So what does this tell us about economic facts? Not much, I would say. We use economic terms to describe the world because they provide concise descriptions of aggregate behavior that allow us to make reliable judgments and predictions about this aggregate behavior. But we don’t think of them as existing independently of aggregate behavior, independently of the physical facts. Rather they are just one way of looking at those physical facts, at a higher level of abstraction; given that Mary wasn’t led to use that particular abstraction it is no surprise that she didn’t know the economic facts.
4.2 Mary, The Colorblind Color Scientist
But there is a similar thought experiment that is supposed to lead us to the opposite conclusion, namely that the economic facts can’t be reduced to the physical facts, and thus exist in addition to the physical facts. In this thought experiment we are asked to imagine a color-blind woman, named Mary, who studies color vision. By hypothesis Mary knows all the physical facts there are to know about how color vision works in the brain, and how it gives rise to certain thoughts and actions. One day Mary’s color-blindness is fixed by a new kind of surgery. Now that she can see again doesn’t she learn something new, specifically what colors feel like? If this is true then clearly the physical facts don’t reveal all there is to know about the mind, or so it is claimed, since she already had access to all the physical facts before.
Obviously the situations in which Mary the colorblind color scientist and Mary the economics-blind social scientist find themselves in aren’t quite the same. The question we need to ask ourselves is why the situations are different, and if those differences matter in terms of leading us to the conclusion that qualia, facts about what experiences feel like, exist in addition to the physical facts.
One difference is in the way the Marys have access to the relevant facts. In the case of economic facts Mary has to deduce those facts from the physical facts she is already aware of, while in the case of qualia Mary is seemingly just presented with them the first time she is able to see color. But this difference alone isn’t sufficient to motivate the conclusion that qualia exist in addition to the physical facts while economic facts do not. We could imagine that in the case of Mary the social scientist instead of teaching here what the GDP was we simply did some weird surgery on her, so that now when she is exposed to certain situations she immediately forms the thought that they are an instance of an economic exchange, and that from her knowledge of patterns of economic exchanges she also will immediately form the thought that the GDP is rising or falling. In this version Mary the social scientist also comes to know economic facts without deducing them from physical ones, but it doesn’t make the economic facts seem like they must now exist in addition to the physical facts. Clearly then the immediacy of the facts, by itself, is not what motivates the different conclusion in the two cases.
The real difference seems to be that the qualitative facts are defined only subjectively, while economic facts are not. And thus we can accept the possibility that economic facts can be explain in objective terms because they themselves are objective, but that qualitative facts cannot, because they are subjective. As compelling as it may seem we have to watch out here for the possibility that this response is begging the question. What if qualia were some feature of the physical world, and we knew what this feature was? Couldn’t we then explain them to Mary the color scientist using this explanation? Possibly not, because if qualia are inherently subjective Mary can only really know about qualia by experiencing them, even if qualia could be identified with some feature of the physical world. So defending the idea that qualia aren’t in fact part of the physical world seems to turn on the inability in general for anything that is inherently subjective to be identified with some physical event. If they could be then the mere fact that qualia are subjective couldn’t rule out the possibility that they are really something physical.
So, if we can find subjectivity in a completely physical system then there is no reason to resist the idea that our inherently subjective qualia can be identified in some way with a feature of the physical world on the basis of their subjectivity alone. Consider then a completely physical system that can be put into a number of states. Let us further suppose that this system is intelligent and rational, although not conscious, since that would possibly be to presuppose that it had qualia. Now there are basically two ways to know about this system. One way is to look at it from the outside, and to say in objective terms what state it is in. But the system can also know about itself, in a fundamentally different way. To make things simple let us further suppose that the states of this system can be described as a collections of propositions, and that this description captures all the interesting facts. Some of the propositions are about the world, but some are about the system itself. We will say that the system “knows” these propositions. By know in this case we mean that the propositions affect the behavior (possibly verbal) and future states of the system, such that future propositions may be deduced from, or be about, those propositions. Since this system is not conscious we don’t have to worry about knowledge meaning any more than this. Again let us make one more arbitrary choice and say that what the propositions that are about the system itself “tell” the system is that it is in a certain kind of state, state A, state B, and so on, all the way to state Z. Now this system, given that it is rational, will “notice” that the proposition that it is in a certain state is correlated with certain kinds of behavior, certain other propositions, ect. Thus the proposition that it is in a certain state is a kind of subjective knowledge. And assuming that we can’t look into it and read its internal state off, which is something we certainly can’t do with people, the system knows more about itself than we do.
Now it is important to keep in mind that the system we are discussing here does not see its own internal state as being structured around a number of propositions. All it “knows” about its internal state is that it can come in one of 24 flavors: A, B, C, and so on. The system cannot arbitrarily reflect on the propositions that make up its state, and so these internal propositions might be described, anthropomorphically, as “unconscious”. Which is simply to say that the system is never directly aware of its inner goings on as a collection of propositions being repeatedly operated on. The process by which it learns that the “I am in state Q” proposition is correlated with a lack of energy (inability to do certain activities, reduced strength and perceptual clarity, the need to sleep soon, ect) is not done by realizing that the proposition “I am in state Q” occurs with these other features, because that would require our system to have some knowledge about how its state were structured, which it does not. (To explicitly form the proposition “The ‘I am in state Q’ proposition is correlated with proposition X”.) Rather the correlation is noted by an automatic process, which causes “I am in state Q” propositions to now be shortly followed by “I will need to rest soon” propositions after the “I am in state Q” proposition has been followed by the need to rest sufficiently often. And note that the system here does not realize either that the “I will need to rest soon” proposition is present either; the existence of this proposition manifests itself only by giving rise to certain other future propositions, and certain behaviors. Remember the point of this thought experiment is to consider a system that is not conscious, and making it aware of its own internal states in that way might be too much (not to mention that it adds an additional, and more complex, form of subjective knowledge).
Given that this system is fairly intelligent what would it do if it were placed in the situation of Mary (let us assume “I am in state Z” is only triggered when seeing green, and we don’t let it see anything green). Well, given how we have described it here it would react in basically the same way as Mary does. Yes it would understand all the facts about its own operation. It might even be able to realize that its internal state could be characterized as a collection of propositions, just like we do. I could even realize that there is a possible proposition “I am in state Z” that has never been triggered, and deduce all the consequences in terms of behavior and alterations to its internal propositional state that would result from it. Even so when we first expose it to green it will be put into a state containing “I am in state Z” for the first time. And as a result the proposition “I have never been in state Z before” may arise, or any number of similar propositions like it (or at least they should if the internal processing is sophisticated enough for the system to be called rational). And perhaps even “so this is what Z feels like”, if it were programmed to use “feels” talk to discuss its internal states, because now, the proposition “I am in state Z” having arisen at least once, its internal machinery can compare future occurrences of “I am in state Z” to this one, and so it “knows” for the first time “what it is like”, even though within the room it could have predicted these reactions, known how being put into that state would affect its processing, and so on.
But this is odd, because our hypothetical machine is purely physical. Nor does there seem to be a reason that we couldn’t construct it. Admittedly we don’t know how to make a machine reason practically yet, but there doesn’t seem to be any reason why it would be fundamentally impossible. But yet this machine has given us all the evidence that Mary has given us to support the idea that there is something in addition to the physical going on. What this hypothetical machine demonstrates is that subjectivity can exist, and seemingly pose a problem for our explanations, even without something extra in addition to the physical (couldn’t our machine reject the physical explanation of its own subjectivity, demanding that an understanding of its operation inform it what being in state Z “is like” without being in that state?). But, because this problem doesn’t require something in addition to the physical in order to arise, the fact that qualia are essentially subjective doesn’t prove that they can’t also be explained in objective terms.
4.4 What Are Qualia?
However demonstrating that qualia can have an explanation in objective terms, despite their essentially subjective nature, is different from giving that explanation. They still might be something over and above the physical; the fact that they might have a physical explanation doesn’t guarantee that they actually do have a physical explanation. So allow me to outline a brief proposal then as to what qualia really are, which has already been suggested by the construction of the hypothetical machine that I appealed to in order to explain subjectivity. As with intentionality, by showing that we can construct an explanation of qualia in physical terms I hope to dispel the worry that something non-physical is required to explain the mind.
Under materialist theories we can think of the mind as a system that receives inputs and produces outputs, a way of looking at minds that we have already used to explore intentionality. Let us specifically consider vision in one such system. This system receives an array (a two dimensional grid) of signals, each of which corresponds to some unique color. We can think of each of the raw signals as being a single letter inside the system (the letter stands for the signal itself, so there is no temptation to treat it as a number or some other complex construct). And the only processing the system can do on these letters is based on their position in the grid an on comparisons of them to each other (given two letters the system can tell how similar they are, although this is itself a primitive function, the workings of which are not exposed to the higher functions of the system). It is my hypothesis that this is all there is to the experience of sight (and in general all perception of the world), the processing of certain primitive signals..
Obviously the human mind has a more elaborate visual system, we favor certain colors over others, and think certain combinations are more pleasing than others, but this doesn’t make the signals themselves any less primitive. We could work these complexities into the system by arguing that these ideas are either themselves generated by another basic kind of processing that is done (like the comparison of similarity between signal), or it is built on top of the comparison, in the sense that certain degrees of similarity are pleasing while others aren’t. But these details are not really important to the explanation of qualia (or at least to our brief explanation of qualia).
The fact that a system deals with visual signals obviously does not, by itself, give rise to qualia. To introduce qualia we need to posit a certain way of reacting to the introspective gap that exists in these systems, specifically the one that exists because the signals, and the comparisons between them, are simple with respect to introspection, meaning that no amount of introspection can reveal their structure. When we think about our perception of a color introspection doesn’t tell us why the color is the color it appears to be, it just simply is that way. When introspecting on our visual phenomena we simply find ourselves having the thoughts that “it is red” or “it is green”, but without ever knowing why we are having those thoughts. Likewise we can’t say why two colors are similar on the basis of introspection. Of course we have constructed theories about color similarity, based on knowledge about how various colors appear when mixed, but these theories are not revealed to us by our introspective awareness of color similarity, which only tells us how similar they are, not why.
So what does the mind do when some aspect of it is closed to introspection? Well in general minds could do anything, but human minds have a specific and well-documented response, they make things up. For example, if the mind can give us a solution to a problem without us consciously being aware of where that solution came from (because it was developed by the unconscious). If you ask a person in such a situation where that solution came from they will fabricate a story, a story that has nothing to do with how they really came to the solution. But they will believe that story, and be totally unaware of the fact that it is a fabrication. Similarly, if basically identical items are placed in a sequence people tend to prefer the rightmost one (we can show this by shuffling the items and proving that in each arrangement the rightmost is favored by a statistically significant number). This means that some people had no reason to pick that one except for its position. But when you asked people why they picked it every last one of them will have a story about why it is superior to the others, and will be totally unaware of the fact that their choice was determined largely because of the position. I propose that qualia are another one of these made up stories. When we introspect on our experience we come to certain things that are un-analyzable by the mind. The mind doesn’t know why the signal is the way it is, or why it is having the thought “it is red”, so it makes up a story, like every other time introspection can’t reveal the answer. And that story is called the “feel”, we say it is red because it “feels” red. What does it mean for something to “feel” red, well we can’t say (except maybe to make analogies to other colors – it’s a bit like orange – but we then find ourselves unable to say exactly how it differs from orange – well it’s more red). I think this is good evidence to support the idea that the “feel” is a construction in response to a failure of introspection, and not a an extra, non physical, feature.
A qualia then is simply a higher order thought, directed at the unexplainable simple phenomena that occur in perception which attempts to explain them. All there is to experiencing a certain qualia is having that certain higher order thought, or so I claim. And this is how Mary learned something about qualia for the first time upon leaving her room, she was in that higher order state for the first time. Even though she might have known that she would be placed into such a state, and the consequences this state would have in terms of subsequent thoughts, this is the first time that she has subjectively encountered it, and so it feels new to her, even though she really hasn’t learned anything new.
4.5: A Brief Detour Back To Intentionality
So, now that I have given a brief outline of a materialist theory of qualia, let me go back and completely wrap up the materialist theory of intentionality I had developed above. As you remember I said that it didn’t completely explain the experience of intentionality, but that the remaining work, explaining the feeling of a thought being about an object, could be explained by our theory about qualia. Now part of our experience of intending is expectations about perceptions. When I think about an apple I have expectations about what the apple will feel like, look like, and so on. And I also have a set of expectations about what will happen if I move in relation to the apple, if I bite into it, if I throw it, and so on. These reflect the input-output correlations that the theory claimed were the real basis of intentionality.
But there is more to the experience of intentionality than that. When I think about an apple I am not just thinking about those things, I also feel that I am thinking about an apple. It is natural to suppose that this is a qualia, and thus represents an introspective failure. But in the case of perception introspection failed when it attempted to analyze inputs that were basically un-analyzable, so what is introspection unable to uncover in the case of intentionality? I think it is the connection between all these expectations. Obviously the many and varied expectations we have about an object are connected to each other, otherwise they wouldn’t be all available to us when we needed to think about that particular object. I hypothesize that when we introspect on that connection our introspection fails, the connection just is there, without further details. But, as usual, introspection makes up a story for us, in this case that they have the “feel” of being about the same object.
4.6 Wrapping Up
So we can admit that qualia are essentially subjective, while at the same time explaining them in terms of objective facts, given that we allow subjectivity to be explained in terms of the way certain systems are constructed, and we allow that the qualia themselves can be explained by the way such systems may react to the introspective gap created by their own construction. Of course not every qualophile will be happy with this solution, because in a sense it does away with the idea that there is something extra “that it is to be like” a system experiencing a certain qualia, and replaces it with a certain kind of reflection on the system’s own workings. So much the worse for that brand of qualophile, in my opinion; we can’t expect our intuitive approach to be right all the time.