Epiphenonema: can't live with 'em, can't live without 'em

Hampshire College

Philosophy of Mind

Introduction

Frank Jackson, in his paper “Epiphenomenal Qualia”, argues that we could know all the physical facts about the world, yet we would never know of qualia, and that therefore qualia are not contained in the physical world. Furthermore, he claims that qualia are epiphenomenal: that is, that they are caused by physical processes, but do not have any effect on the physical world. In this paper, I will argue that this claim that qualia are epiphenomenal destroys the Jackson’s first argument, and yet, this same argument does not work without epiphenomenalism. I will conclude by discussing the reasons I think that we should continue to believe a physicalist thesis regarding qualia.

Epiphenomenalism

In the second half of his paper, Jackson argues that there is no good reason that one should not accept that qualia might be epiphenomena—caused by physical processes, yet with no causal power whatsoever in the physical world. I do not believe that he establishes the epiphenomenalism as a sound hypothesis.

Jackson lists three major objections that he feels philosophers often have against epiphenomenalism:

1. “It is supposed to be just obvious that the hurtfulness of pain is partly responsible for the subject seeking to avoid pain, saying ‘it hurts’ and so on.”
2. “According to natural selection the traits that evolve over time are those conducive to physical survival. We may assume that qualia evolved over time—we have them, the earliest forms of life do no—and so we should expect qualia to be conducive to survival. The objection is that they could hardly help us to survive if they do nothing to the physical world.”
3. “…how can a person’s behavior provide any reason for believing he had qualia like mine, or indeed any qualia at all, unless this behavior can be regarded as the outcome of the qualia…And an epiphenomenalist cannot regard behavior, or indeed anything physical, as an outcome of qualia.”

The first objection, as Jackson phrases it, is silly, as all arguments resting on “obvious”-ness are, and so his reply to it is not very interesting for our purposes. The second objection betrays a substantial misunderstanding of evolutionary theory, and I think Jackson’s reply to it is correct[1]. The third objection, however, is very interesting, and it is in his reply to this objection that Jackson makes the argument that I now wish to deconstruct.

Jackson’s reply to this third objection is in the form of an analogy, which I will paraphrase. Suppose I read in the New York Times that the mayor of NYC is cutting funding for police. I can reasonably infer that the New York Post also reports on this fact, even though the New York Times and the New York Post may not have any influence on each other. I can do this because I know that the New York Times and the New York Post both report on issues of interest to New Yorkers, that the New York Times reporting that the mayor is cutting funding for police is a good indication that the mayor is, in fact, doing such a thing, that this is an issue of interest to New Yorkers, and that therefore the New York Post has also probably reported on this fact. The analogous case for qualia (which Jackson never explicitly lays out) is that given that my nervous system produces both qualia and certain types of behaviors (analogous to how the mayor’s actions “produce” the stories in both the New York Times and the New York Posts), and that your behaviors are similar to mine, I can infer that you have the sort of nervous system that would also produce qualia.

The difficulty with this argument, as Daniel Dennett (“’Epiphenomenal Qualia?”) points out, comes from the fact that, according to Jackson’s account, qualia have no causal powers in the physical world. Because of this, qualia cannot influence my behavior in any way[2]: otherwise what we have is not epiphenomenalism, but interactionist dualism. For this reason, it is impossible for me to know that I myself have qualia, as their existence or non-existence would make no difference to the workings of my brain.

“Ah,” it may be replied, “but just because qualia do not have any effect on the workings of your brain, they may still have an effect on the workings of your mind, say, your belief that you have qualia.” This is, strictly speaking, a valid move. However, it has troubling consequences: for example (referencing Dennett), if I lost my qualia, I would presumably no longer believe that I had them, but I would still act exactly as if I did. Mental states could no longer be said to influence behavior, as those mental states are potentially influenced by qualia and qualia cannot influence behavior[3]. It seems, then, that the epiphenomenalist is left with two options. One, they could say that qualia influence mental states, in which case mental states cannot influence behavior and my mind is completely severed from the physical world, or two—if we want to preserve the influence of mental states on behavior—that qualia do not influence mental states, in which case it is impossible for me to know whether I myself have qualia. The first option is unpalatable (as I will discuss below), yet if the epiphenomenalist takes the second position, he/she must admit that there is no way for me to infer the existence of qualia in others, since I cannot even use myself as an example!

The Knowledge Argument

“Alright,” says the epiphenomenalist. “Let us assume your analysis is correct. If qualia have no effect on the physical world, then it is impossible for us to know whether we ourselves have qualia. However, there may be many things that it is impossible for us to know about, but which nevertheless exist. For example, there might be creatures that live beyond—and will always live beyond—the area of the universe that we can observe. It would be impossible for us to ever know about them, yet that does not change their existence.”

Against this, many would argue that, in the absence of any sort of evidence, it is useless to speak of some thing’s “existence”. There is a much larger philosophical issue at stake here. Happily, we do not need to discuss this issue now. For if qualia are truly epiphenomenal, then Jackson’s Knowledge Argument—the central part of his paper—completely falls apart.

This argument is as follows. Imagine that there is a neuroscientist, Mary, who for some reason has been forced to live her whole life inside a black-and-white room[4]. Despite this, Mary becomes a specialist in color vision and, eventually, learns everything that it is possible to know about how color vision works: complete descriptions of the physical mechanisms, the functional roles that color perception and various colors play, etc. One day, Mary’s captors decide to let her out of the room. She gets handed a tomato—the fruit of choice for released prisoners—and finds that, despite her omniscience regarding color vision, the color of the tomato is a completely new experience for her. “Aha,” she exclaims, “so this is what ‘red’, the color I know is associated with tomatoes, actually looks like!” She has learned a new fact about color that was not amongst any of the physical facts: what such color is like. This would seem to imply, then, that this fact about what red is like—the quale associated with red—is not part of the physical world.

However, this story does not make any sense if we accept that qualia are epiphenomenal. Firstly, if qualia have no influence on the physical world, it cannot be the new quale that caused Mary to make any sort of exclamation. She would have said the same thing, quale or not. This objection does not mean much, though—it can easily be countered that even though we cannot know from her actions whether or not Mary has learned a new fact, it is still the case that her subjective experience obviously now includes that fact. Perhaps we can imagine that Mary merely thought about the fact that she now had this new qualia, so we don’t have to worry about its effect on her actions. This does not help, though, as we must now ask whether Mary’s thoughts are reflected in the workings of her brain. I.e., could some sort of neuroscientific Laplace’s Demon, who knows everything about the physical structure of Mary’s brain at any time, be able to tell us what she is thinking? I think Jackson would want the answer to this to be yes. Yet, if it is true that qualia can influence Mary’s thoughts, and that thoughts are reflected by brain states, then one of two options become must be chosen: either qualia can influence brain states, and are therefore not epiphenomenal, or thoughts—and this would seem to include all thoughts, not just those specifically about qualia, as any given thought could always have qualia as part of its causal history—are entirely distinct from those brain states, even though they are reflected by them. If we choose this latter option, then the causal relationship between thoughts and brain states, if one exists, must be a one-way street: thoughts are caused by the brain states, but thoughts cannot influence brain states, and are, in fact, epiphenomenal. However, if thoughts are epiphenomenal, then anything that can be influenced by thoughts would have to be epiphenomenal too (otherwise, thoughts could exert some sort of influence on the physical world). This would likely include any mental state Mary might have (thoughts are, after all, quite influential in the mental world). In fact, if the entity called Mary can learn a new fact from experiencing a new quale—Mary can be influenced by qualia—then Mary herself must be an epiphenomenon: something caused by the lump of matter that has just been released from its black-and-white room, but with no ability to influence that lump of matter in any way.

Does the Knowledge Argument work without Epiphenomenal Qualia?

We have seen that if qualia are truly epiphenomenal, and yet I can learn a new fact from experiencing new qualia, then I myself must be epiphenomenal—otherwise, qualia can influence the physical world through me. While there is no immediately apparent logical reason why this view could not, possibly, be true, I think very few would be willing to accept it. If I am epiphenomenal, then, by definition, I have no power over the physical world—not even my own body. I am merely experiencing the effects of processes over which I have no control. In my mind this consequence is a sort of informal reductio ad absurdum, and I think most philosophers would agree[5].

However, perhaps we have been too concerned with epiphenomena. Perhaps Jackson’s Knowledge Argument still works—perhaps he just made a mistake by attaching epiphenomenal qualia to it. Of course, there are numerous objections to the Knowledge Argument not related to epiphenomenalism: perhaps what Mary learns is not a new fact, but a new mode of knowing some fact—not a “what”, but a “how”, or perhaps, as Dennett points out, our reaction to the Mary story merely reflects that we cannot conceive of what it would mean for Mary to know everything physical that there is to know—perhaps, with that knowledge, she would be able to reconstruct the subjective experience of seeing red. According to Dennett’s objection, Jackson is merely begging the question by pre-supposing that knowing all the physical facts will not allow Mary to have the experience. However, I want to take a different tack. I want to argue that, without epiphenomenalism, the Knowledge Argument does not work. Of course, we also just saw that with epiphenomenalism the Knowledge Argument doesn’t work. This means that if I’m successful, the Knowledge Argument will have nowhere to go.

If qualia do exist, but are not epiphenomenal, that is, they do have some sort of influence on the physical world, how does this affect the story of Mary? Let’s go back to before Mary left the room, when she knew everything physical there is to know about color vision. If the qualia associated with seeing color have a physical effect, then this effect would have to be included in Mary’s knowledge. This would lead to one of two options: either qualia are non-physical, in which case Mary’s description of the physical mechanism of color vision would have to have some sort of causal hole in it—that is, she could say only that A causes B, and then B causes something or other, and that something or other causes D—or that qualia is included in the physical description, in which case Mary would already know the subjective sensation of seeing red before leaving the room. Obviously, the second option leads to physicalism (although it is possible that the physical description might include stuff that we have not yet incorporated into our neuroscience), so it is obviously the first option that anyone who wants to say that qualia are non-physical must choose. This option implies that the physical world is not closed under physical causation—that is, that there are non-physical things that can nevertheless have physical effects. Now, there is no prima facie reason why this isn’t a viable option. However, there is also no prima facie reason to choose one of the two options over the other (at least, not until one considers the arguments of those such as Dennett). Once again, we are faced with the problem of begging the question: Jackson must pre-suppose that qualia would not be included in a complete physical description. Of course, those who want non-physical qualia could say that the physicalist is also guilty of question begging. The question is empirical.

Qualia of the Gaps

The question is empirical, but we must ask: can it, in fact, be answered? We move now to the realm of personal opinion. It seems likely to me that there is no possible empirical evidence that could allow us to choose between the two options until such time as we do find a physicalist description of qualia. That is, there is no way that we could determine whether the causal gap we have encountered—the place where we throw up our hands and say “well, something happens”—is, in principle, impossible to bridge. The question now becomes: what assumption shall we make when we encounter a seeming explanatory gap? Do we assume that it is intrinsically inexplicable, or do we assume that we just don’t know enough, or may not even be smart enough? I think the latter is a much more satisfactory option, as it allows and motivates us to continue searching for an answer. Furthermore, the history of science includes numerous examples of the solution of seemingly insoluble problems, problems that were declared to be permanently insoluble. Physicalism has triumphed in the past, and I believe that we can reasonably assume that it will continue to do so.

---

[1] Namely, that the fact that some trait exists does not mean that it was evolutionarily selected for. It may, instead, be a necessary byproduct of some other trait that was selected for.

[2] To clarify, in all following discussion I assume that epiphenomenalism defines the inability to influence the physical world as a necessary property of qualia instead of a contingent one. I.e., I take it to mean not only that qualia do not happen to influence the physical world, but that they cannot influence the physical world (at least not in our reality).

[3] At this point a particularly legalistic reader might have begun to object to my characterization of “influence”. “Qualia influence mental states, and mental states influence behavior, but you have failed to show that qualia influence behavior,” that reader may say. To this I would reply: imagine that qualia were taken out of the causal chain. Would this not change the behavior of things further down in the chain? At least one property of influence/causal power is that if in order to give a complete causal account of the behavior of B one must describe the behavior of A, then A influences B.

[4] We may suppose also that, for whatever reason, she cannot see the colors of her own body.

[5] It may be objected I am pre-supposing the existence of free-will here, which would be ironic, as I do not believe in free-will. However, I think the objection is unfounded. Even though we may not have control over our volitions, it still seems very hard to hold that those volitions do not have the causal powers that we feel they do.