experience & reason

(Forthcoming in: Clotilde Calabi & Kevin Mulligan (eds.),

The Croaked Oar: The Illusions of Inner and Outer Perception, MIT Press.)

1. Introduction

In this paper we shall address some issues concerning the relation between the content and the nature of perceptual experiences. More precisely, we shall ask whether the claim that perceptual experiences are by nature relational implies that they cannot be intentional. As we shall see, much depends in this respect on the way one understands the possibility for one to be wrong about the phenomenal nature of one's own experience. We shall describe and distinguish a series of errors that can occur in our introspective access to our perceptual experiences. We shall argue that once the nature of these different kinds of error are properly understood, the metaphysical claim that perceptual experiences are relational can be seen to be compatible with the view that they are intentional.

Before presenting the argument, we should try to articulate some elements of an intentionalist approach concerning the role of experience in our relation to ourselves and to our environment. The picture should offer a motivation for the arguments that follow.

2. The role of experience

Experiences, occurrences in the flow of consciousness, play a central role in our engagement with the world. The way things and people appear to us in experience is intimately connected to the way we think about them and to the way we interact with them. The way we experience the world matters to us. There are experiences we enjoy and we are glad to expect. Other experiences are painful and we would rather avoid having them. Sometimes, however, we accept undergoing a painful experience because we value the understanding we obtain through it. And we can learn to distance ourselves from experiences we find superficial although quite pleasurable. So experiences do not simply impose themselves on our life. With time and with the help of others, we get to know more about their nature and we learn to let them play a sensible role in our life. Moreover, we discover that experiences can lead us astray and we thus learn not to fall prey to them. We can come to understand a person's reaction to a certain situation by getting a grasp on the way she experiences it. We sometimes manage to predict a person's behaviour by putting ourselves in her shoes and thus experiencing her situation as if it were our own. Appreciation of the fact that experiences can lead us astray is crucial in our understanding of other persons and their behaviour.

When experiences matter to us, this is sometimes due to the fact that their simple presence commands a certain kind of response on our part. This is not a mere process, taking place in us, with us playing the role of a detached observer. When we are engaged in responding to an experience, the experience may indeed provide us with a reason for our response. In fact, the experience and our reaction are often part of a unity that may constitute a project we aim to realise. If one aims to understand better a friend's behaviour in a certain kind of circumstance, one may wish that one's own beliefs about that person and her motives are properly responsive to one's experiential, e.g. empathic, take of her situation. And if one wishes to learn something about the quality of wine, one wants one's judgement to be properly responsive to one's complex olfactory and taste sensations. Advances in such epistemic projects are not just a matter of collecting evidence for the corroboration of a hypothesis. It is rather a matter of getting one's own cognitive behaviour to be increasingly guided by the relevant experiences. The more one advances in this direction, the more one senses the cognitive obligation that stems from the presence of the experience. This is not a question of one having lost one's freedom to believe in front of the tribunal of experience. It is rather a manifestation of the fact that one comes to feel the responsibility one would have to carry if one were not to trust one's own experience. As such, this is the situation where believing one thing rather than another starts to become something we really care about. Beliefs one cares about are beliefs that impinge on one's web of experience.

Much of what is true of experiences in general is also true of perceptual experiences in particular.¹ When it comes to our knowledge concerning the external world, perceptual experiences play a central role. Typically, we come to know that the there is a red sphere in front of us by having an experience of something appearing red and spherical in front of us. Such a simple way of gaining knowledge presupposes that our experience is sensitive to shape, colour and orientation in egocentric space. There would not be much point in our experience possessing that kind of sensitivity if our beliefs were not responsive to the fact that they do possess it. Indeed, when one experiences a red sphere appearing in front of one, one is strongly inclined, in fact, one is motivated to judge that there is a red sphere in front of one.

Philosophers have not always agreed that this rational relation between perceptual experience and belief is constitutive of our knowledge about the external world. A subject may fulfil all the responsibilities that flow from her perceptual experiences and yet fail to be a reliable source of information about her environment. Establishing criteria for the satisfaction of the sort of rational requirement at issue, then, would fall short of providing genuine criteria for knowledge.

One may be tempted to react to such a line of argument by questioning the very notion of perceptual experience it appears to rely on. Indeed, it has been argued that once it is allowed that our experience does not put us into direct contact with the world, once it is allowed that the experience could occur without establishing such a contact to the world, the relation between perceptual experience and knowledge breaks down.² One should thus rather conceive of perceptual experiences in a way which makes it essential to them to put us in direct contact with the world. In this sense, perceptual experiences, as such, could not be faulty. Hallucinations and illusions would have to be experiences of a different kind.³

Perceptual experiences, it is generally admitted, are intentional; they are directed to something. But for an experience to be intentional, it needs to possess correctness conditions. This, at least, seems to be suggested by Brentano's famous idea that an intentional ‘relation' does not presuppose the existence of its object.⁴ Should we thus consider the claim that perceptual experiences put us into direct contact with the world as relying on the idea that perceptions are not intentional? Or should we abandon the idea of a connection between intentionality and correctness conditions and accept a modified notion of intentionality as referring simply to an asymmetrical relation between the experience and its object, such that the experience is directed onto its object while the object is not directed onto the experience?

One of the central problems with the relational view sketched above concerns the way perceptual experiences can guide us in the formation of beliefs. In order to rationally ground our beliefs, perceptions should provide reasons. Reasons for beliefs, it used to be said, need to be inferential. And so, it was concluded, not perceptual experiences as such, but the beliefs caused by them, provide reasons. In spite of its well-known weaknesses, this argument relies on a strong intuitive basis. It is the idea that a fact or event as such cannot provide any rational guidance for one's beliefs. For something to rationally guide one's cognitive activity there must be something normative about its very nature. Intentional content conceived as the possession of correctness conditions is one way of fulfilling this requirement. Let us explain.

3. Intentionality, correctness conditions, and the phenomenology of experience

Consider experiences such as a pain in one's foot and the visual experience of two parallel lines that appear to converge⁵. There is an ordinary sense in which there can be something misleading about experiences of this kind. A pain in the foot is misleading when there is no ailment in the foot.⁶ The perception of the two lines as converging is misleading when the lines are actually parallel. How should one account for the very possibility of experiences being faulty, or leading us astray, in the way described above?

One prominent line of thought suggests that the mistakes under consideration are related to the fact that those experiences are essentially intentional in the following sense: they possess correctness conditions, and the possession of those conditions is constitutive for their identity. The experiences would not be what they are if they did not have the correctness conditions they have. One's experience of pain would not be the pain it is if it did not present the foot as having an ailment, and one's visual experience of two lines appearing to join each other would not be the visual experience it is if it did not present the lines as joining each other.

Correctness conditions are conditions an experience typically imposes on something else, on something external to the experience, something that transcends the experience⁷. The ailment in the foot is external to the experience of pain, and the convergence of the lines is external to the visual experience of them appearing to converge. The externality under consideration can be captured with the notion of ontological independence: the experience does not depend for its existence on the satisfaction of its correctness conditions.⁸ One can experience a pain in the foot although there is no ailment there and the lines can appear to converge although they do not.

Externality of experience as such does not yield mind-independence. It is required that the experience be independent of the satisfaction of its correctness conditions, not that whatever satisfies those conditions be mind-independent. Cases where the correctness conditions are obviously satisfied by mind-dependent states, properties or events are readily available. One's second order belief that one is in pain is correct if and only if one is in pain, a plain mental state. The intentional character of those second order states would however be questionable if it could be shown that one's thought that one is in pain depends for its very existence on one's being in pain.

We can now see how experiences can be misleading. Experiences, one may surmise, are misleading when their correctness conditions are not satisfied. This is fine, but one must be fully aware of what one is thereby putting into the notion of correctness conditions. The simple fact that an experience has correctness conditions that are not satisfied does not make it faulty. We further need the requirement that the experience's occurrence should be related to those conditions being satisfied. Since we saw that externality prevents the existence of the experience from depending on the satisfaction of those conditions, the relation cannot be one of existential dependence. We should rather say that in order to be correct the experience ought to occur only when its correctness conditions are satisfied, which of course does not mean that it can only occur under such circumstances. Correctness conditions of experiences thus generate a twofold constraint: an external constraint on whatever might satisfy those conditions, and an internal constraint on the very occurrence of the experience.⁹ The possibility for the experiences to be faulty depends on the intertwining of these two constraints. For an experience to be intentional - for it to have intentional content, as it is slightly misleadingly said - means, then, for it to be subject to the intertwining of those two constraints.¹⁰

We can now see how the intentionality of perceptual experience is related to its guiding role. It is because perceptual experiences ought to occur only if their correctness conditions are satisfied, that their occurrence provides one with a reason to believe that the world is as it perceptually appears to be. The mere occurrence of the perceptual state, void of the norm that applies to it, would not provide such a reason.¹¹

On one prominent view, a view we share, the claim that perceptual experiences are intentional is part of their descriptive phenomenology. By saying this one means that the property of having certain correctness conditions is constitutive of how the experience is consciously presented to the subject. It is part of what it is like for someone to have that experience. We shall not argue in favour of this view in the present paper, but some elements of the relation between intentionality and the phenomenology of experience will be set out in what follows.

A terminological clarification may help at this point. Suppose that in perception the lines appear to a subject S as converging. The perception is veridical only if the lines converge. Call this condition F.¹² A possible world w, or something in a possible world (a situation), can satisfy this condition. A world w may thus have the property of being a world where the lines converge. Condition F would be a property of w, w would be an F-world, w would be a member of the set of worlds where the lines converge. The phenomenological claim under consideration is that F determines a conscious feature of S's perceptual experience. Contrary to w, the experience does not possess the property F, the experience is not an element of the extension of F. Rather: having F as an intentional content is part of how the experience is given in consciousness. Since we said that an object perceptually appears to the subject as F, we might classify the subject's experience as an F-appearance. Obviously, being an F-world and being an F-appearance are properties of very different kinds.¹³

It is one thing to determine the intentional content of an experience. It is another thing to determine whether that content correlates de facto with conditions external to the experience. Take the example of colours. It is one thing to say that the visual experience of a red tomato presents the tomato as being red independently of one's experience of it. It is another thing to claim that redness is a property something can instantiate without the subject experiencing it. The first claim concerns the phenomenology of the experience; the second claim concerns the metaphysical constraints on the correctness conditions of the experience. The intentional content of an experience may be constituted by correctness conditions that cannot, or can only partially be satisfied. Even under such conditions, it may still be true that those correctness conditions stand in a significant relation to conditions that do possibly obtain in the world. There may not be, as some think, any colour properties in the world, but it may still be true that there is an interesting set of properties that are instantiated in the external world whenever an object looks red. Elaborating on such connections and constraints is no part of descriptive phenomenology. It is part of the metaphysics of colour and of the epistemology of colour perception.

4. Phenomenal properties and the metaphysics of experience

Experiences possess various properties. They are likely to possess properties that do not constitute the way it is like to have them. They may possess properties that are sometimes, or maybe even typically, unconscious (i.e. not conscious). For instance, perceptions of a certain kind may cause modifications in our body we are not aware of, modifications nobody knows about. Having that sort of causal power is a property of the experience that does not belong to its phenomenology. Properties that belong to its phenomenology are properties that characterise the way the experience presents itself in the stream of consciousness. They constitute what it is like to have those experiences. Let us call these properties phenomenal properties.

Among the phenomenal properties there are properties an experience has by virtue of things appearing to one in the experience. They are phenomenal properties constituted by the intentional content of the experience.

There is a philosophical position, pure intentionalism, as we should call it, which claims that all phenomenal properties are constituted by the intentional content of the experience. That means that every property which influences how an experience is given when it occurs in the stream of consciousness is determined by a property the object presented in the experience appears to have. The determination is the one we specified above: if the experience represents something as F, then the experience has the property of representing something to be F, in short of being an F-appearance.¹⁴

Let us now look for a moment at the identity conditions of experiences. Consider identity over time. Suppose there is a strong pain in the leg.¹⁵ As time goes by it decreases and eventually disappears. Being strong, then, was not an essential property of the pain. The very same pain could have been less strong. Or consider the power a certain pain has to cause one's heart to beat quicker. Again, the heart may beat slower and the pain remain just the same. A pain sometimes moves: it started in the knee, now it is in the thigh. So, location too is no essential property of the pain.

The point is not to establish these specific claims. We rather aim at understanding what is at issue when one inquires into them. If some properties are not essential to the experience, then this may be expressed by saying that some properties do not constitute the nature of the experience.¹⁶ Are there properties that constitute the nature of an experience? Are there essential properties of an experience? Many philosophers have submitted that at least some, if not all the phenomenal properties of an experience are essential to it. It would then be essential for a pain, for instance, to be painful, to hurt. An experience cannot be a pain, it is said, if it does not hurt. What does this mean?

Should we say that the experience appears as hurting, that hurting is the way the experience appears? Although philosophers sometimes talk like that, this way of speaking can be seriously misleading. It is misleading when it suggests that a pain experience¹⁷ appears hurting in the sense in which an apple appears red in perception. For to be hurting would then be the property the pain experience appears to have, instead of being a property it simply possesses. For the former to be the case there should be something, some further experience, in which the pain experience appears as hurting, just as much as we need a perception for the apple to have the property of appearing red. There are many serious philosophical problems with this picture of introspection. But apart from those problems, the issue at stake here is simply that there is a difference between the claim that the pain possesses the property of hurting and the claim that the pain appears to possess that property. When we say that a pain experience has the essential feature of hurting, we are not saying that it has the essential feature of appearing to hurt.¹⁸

Remember the difference we made above. An experience in which something appears F is an F-appearance. But: an F-appearance does not appear F. Thus: a red-appearance is an experience of something appearing red. The experience itself, however, does not appear red. Being a red-appearance is one of the experience's phenomenal properties: the experience represents something as being red and this intentional content constitutes one of the features the experience presents itself as having when occurring in the stream of consciousness. Similar considerations apply in the case of pain. Hurting is one of the ways in which a certain region of the body appears in the experience of pain: the foot hurts. The experience has the property of presenting the foot as hurting, just as much as the visual perception presents the lines as converging. To be an appearance of something as hurting, then, is a phenomenal property of the experience of pain. Strictly speaking, we should not say that the pain hurts. We should rather say that the pain has the property of representing something as hurting and that this property constitutes part of what it is like to have a pain. The essentialist claim under consideration is that this phenomenal property is a necessary property of the pain: it constitutes the nature of the experience of pain.

Is this true for all phenomenal properties of an experience? Some of the examples mentioned above seem to suggest that it is not. The object of a pain appears to be located at a certain place in one's body, but it is not immediately clear that the location it appears to have could not change without the pain experience stopping to be the very same experience (i.e., that very same pain could 'move' or could 'be' somewhere else in the body). Does the essentialist claim apply only to phenomenal properties? Could it not be essential for a pain to cause a certain kind of behaviour? These are all genuine questions, but we shall not address them in what follows. As mentioned above, we do not intend to propose a theory about pain experiences; we rather want to suggest an understanding of the philosophical claims at issue.

5. The nature of experience

Suppose that there are phenomenal properties that are not determined by an intentional content and thus are not appearances. It is not easy to find clear examples of such phenomenal properties. Blurredness is sometimes given as an example in point.¹⁹ But maybe even painfulness is such a case. In fact, it may be submitted that there are two distinct phenomenal properties: the painfulness-appearance on the one side, and the pure painfulness on the other. To claim that there is a sense in which painfulness is a property an experience has by virtue of there being something that appears painful, is not to deny that a pain may also have a pure phenomenal property of painfulness. All we need, for our purpose, is to be clear about the fact that these are different properties.

Some would submit that only phenomenal properties, whether appearances or not, can count as determining the nature of a conscious experience. But our argument need not depend on such an assumption. Suppose thus that an experience has an unconscious property G that is essential to it. It belongs to the nature of the experience to be a G. G is not experienced in the flow of consciousness. Although being essentially a G, the experience does not possess any phenomenal G-ish property. In that case, the nature of the experience, or part of it, is not manifest in consciousness. This does not imply that one cannot come to judge, by some other means, that one's own experience is G.

Now, take a relational property of standing in relation R to an object of a certain kind.²⁰ Suppose that the relational property is essential to the experience: it corresponds to part of its nature. Suppose further that the property is in fact given in experience. When a subject has an experience that possesses that relational property, then the experience is an appearance of itself as standing in the relevant relation to an object of a certain kind. We may say that the experience is a reflexive R-appearance. This is a token-reflexive phenomenal property of the experience. It is not a case of the experience appearing in a certain way to another experience, it is instead the case of the experience appearing to itself in a certain way.²¹ In such a case, then, the experience possesses a phenomenal property that corresponds to an essential correctness condition satisfied by the experience itself. The very nature of the experience (or part of it) is consciously manifest through a phenomenal feature of the experience.

Consider at present another experience with the same intentional content and the same phenomenal properties as the experience above, with the notable difference, however, that it does not satisfy the condition set by its intentional content. If R is taken to be a casual relation, then the case under consideration would be one where the experience is not caused by an external object. Now, if being caused by an external object is an essential property of any experience that has it, then representing an experience as being caused by an external object when it is not is to represent the experience has having a nature it does not have. When the intentional content determines the phenomenal property of the experience, that is, when the experience has the phenomenal property of being a reflexive R-appearance, then the experience appears to itself as having a nature it does not have. One of the experience's phenomenal properties misleads the subject of the experience about the very nature of the experience.

This, of course, leaves the possibility open for the phenomenal property to be essential to the experience. Thus, an experience may possess an essential phenomenal property that misleads the subject about the nature of the experience. This happens precisely when standing in relation to an object of a certain kind is an essential condition on experiences, the experience is essentially a reflexive R-appearance, but the experience does not stand in relation to an object of the intended kind. In our example: the relational property of being caused by an external object and the phenomenal property of appearing itself to be caused by an external object would both be essential properties of experiences, they would determine together the nature of the experience and thus the fundamental kind to which it belongs. Under these conditions, an experience that reflexively appears to be caused by an external object when in fact it is not, misleads the subject about its own nature.

6. Experiential and Phenomenal Error

The case described above concerns one of the ways the experience can mislead the subject about its own nature. Let us try to better understand the kind of error that occurs in such a case and let us distinguish it from other, related errors.²²

Forget about phenomenal properties for a moment, and consider the sort of unconscious causal powers of an experience we mentioned above. The subject can clearly be wrong about that sort of property: the subject can believe that the pain doesn't cause the heart to beat faster although it does. Now, if that specific causal power is an essential property of the experience, one has a false belief about the nature of the experience. If this is possible with unconscious properties of an experience, why should it not be possible with respect to the phenomenal properties of an experience? Suppose we have a pain experience, it hurts, but for some reason we do not believe that it hurts. Maybe we should start to worry about ourselves, maybe this is the starting point of serious psychological problems, but it is hard to see what in principle could prevent that kind of situation from happening. Now, if it happens, and if hurting is an essential property of our experience, then it seems again right to say that we are wrong about the nature, in fact about the phenomenal nature of our experience.

This kind of doxastic error would have to be attributed not to the perceptual experience itself, but to a higher order introspective belief that is intentionally directed to it. The mistake concerns a belief the subject has about her own experience. The question may now be raised whether the way the experience is given in the stream of consciousness could be the very source of this kind of error. Consider the following analogy. When looking at an artificially produced banana one may mistake it for a real one. It is not that one just wrongly judges that the artificial banana is real; the way it looks provides a reason to believe that it is. The source of the false belief lies in the very way the artificial banana is given in perception. In perception the fake banana, one may say, conceals its true nature. The parallel question would then be whether a perceptual experience could conceal its nature by the way it is given in consciousness. We have seen that the nature of an experience may be determined by different kinds of properties. Suppose, as above, that among the properties that determine the nature of an experience, some are phenomenal and some are not. Being an F-appearance is a phenomenal property, being G is not, and both properties constitute the nature of the experience.²³ There is a fundamental kind of experience that is both an F-appearance and a G. Now, if it is possible that some F-appearance are G-experiences and some are not, then we may obtain a situation in which the subject would not only erroneously believe that his experience is G, the experience itself, by being an F-appearance, would provide him a reason to believe so.

This line of argument would not be uncontroversial. One may wonder how the way the artificial banana looks could provide a reason for the (false) belief that the banana is real, given that the artificial and the real banana are supposed to be indistinguishable on the basis of how they look. Could one not, by symmetry, consider the way the real banana looks to provide a reason for the (false) belief that the banana is artificial? Should one thus not conclude that the way a banana looks cannot provide any more reason for the belief that it is real than it does for the belief that it is artificial? Yet if such a result could be secured in the perceptual case, the analogy would be useless with respect to the introspective case.

Once again it ought to be stressed that we are not committed to defend the asymmetry under consideration. We simply intend to locate the possibility of error one would have to accommodate if one were to maintain the asymmetry claim. This being said, an argument in favour of the asymmetry in the perceptual case might be provided by a combination of common considerations concerning relevant alternatives and warrant transmission. One may thus submit that where the appearance of a real banana can provide a reason to believe that there is a real banana without having to provide a reason to believe that it is not an artificial one, any appearance that would provide a reason to believe that it is an artificial one would have to provide a reason to believe that it is not a real one. The asymmetry may be rooted in the fact that our perceptual system is tuned to represent real bananas, rather than fake ones, without having been tuned to distinguish the real from the fakes bananas by the way they look. This priority may, but need not, be explained in statistical or nomological terms. A better explanation may be provided by considering the evolutionary function of perception: the perceptual experience represents real bananas because real bananas, and not fake ones, lead to an evolutionary advantage and thus to the selection of the kind of perceptual experience under consideration.²⁴ Whatever the merits of this line of argument, its application to the introspective case would obviously require further instructions. If one intends to argue that an F-appearance equips the subject with a reason to believe that the experience is G rather than non-G, and that an F-appearance thus misleads about its true nature when in fact it is not G, one needs a specific argument for the priority of F-appearances that are G over those that are not. This will obviously depend on the specific instances taken into consideration, such as perception and hallucination. A point to which we shall return below.

The relevant result at this stage is that this kind of error, if it could occur, would not concern the phenomenal nature of the experience: although the experience conceals some of its nature, it does not conceal its phenomenal nature. This, at least, holds when G is not itself a phenomenal property of the experience. When the experience provides the subject with a reason to believe that it is G by being an F-appearance, then the way the experience is given in consciousness misleads the subject about one of its non-phenomenal properties. We suggest calling this kind of error, where the experience occurs in consciousness in a way that misleads about its non-phenomenal properties, experiential error. An experiential error can certainly be the source of a doxastic error, in this case of a false introspective belief about the non-phenomenal nature of one's own experience.

Suppose now that the F-appearance we have been considering has a further phenomenal property, the property H. The question would now be whether it is possible for an F-appearance to mislead the subject about its own H-ness. Several situations would have to be distinguished here, depending on whether H is an essential phenomenal property of the experience and on whether H is determined by the intentional content of the experience. Consider first the case where the F-appearance is also an J-appearance. An example would be a perceptual experience of a red square. The object appears red and square, the experience is both a red-appearance and a square-appearance.²⁵ Now, could an F-appearance mislead the subject with respect to its being an J-appearance? By definition, if the F-appearance is also a J-appearance, then the experience presents itself as a J-appearance in consciousness. And if being a J-appearance is an essential property of the experience, then in cases of the kind under consideration the phenomenal nature of the experience is manifest in consciousness.

Could an F-appearance not warrant the subject's belief that she is not having an J-appearance although she is? Consider circumstances where one and the same object appears to have incompatible properties: the object may look both F and J although no ordinary material object can be both F and J.²⁶ In such a case the fact that the object looks J provides a reason to believe that it is J but the fact that it looks F also provides a reason to believe that it is not J. One and the same experience grounds contradictory judgements about the perceived object. We may call this a case of phenomenal dissonance. A situation of this kind may eventually lead the subject to cast doubt on the reliability of her own introspective abilities.²⁷ She may conclude that she has an F-appearance but not really an J-appearance, thus entertaining a false belief about the nature of her own experience. The most obvious explanation for this kind of doxastic error, it appears, would appeal to the pressure general assumptions about material objects and their properties exercise on the particular case. The source of the false belief is not the way the experience is given in consciousness, but general background beliefs about the nature of objects given in perception.

Consider further the case of an F-appearance that has some pure phenomenal property. Imagine a subject with blurred vision who reasons about the origin of the blurred character of his visual experience. Two kinds of questions would have to be distinguished. The subject may wonder first whether the object's edges actually are fuzzy, or if it just appears to him as if they were. This is a question about the correctness of his perceptual experience. But the subject might also wonder whether the blurredness comes from the fact that the object actually appears fuzzy to him or rather from the experience itself. Consider an analogy. Suppose you are looking at a photograph representing an unidentified object on a uniformly coloured background. Suppose the image of the object is blurred. You may genuinely wonder whether the object really has fuzzy borders, like a cloud. Suppose instead that you identify the depicted object as being of a kind you are much more familiar with, like a knife, say. It would then be peculiar for you to wonder whether the knife has fuzzy borders. The default assessment would be that here is something wrong with the picture.²⁸ I^{n analogy, the subject might some to consider blurredness as a pure phenomenal property of his experience, instead of being linked to the presentation of an object as having a certain property.} More should obviously be said about this distinction. But suppose one accepts it. Could then the way the experience is given in consciousness mislead the subject in a way such that he judges, for instance, that the experience has the pure phenomenal property of blurredness while in fact the experience is rather a presentation of an object with fuzzy borders? It is important for this line of thought to admit that the intentional blurredness (B_i) be phenomenally distinct from the pure blurredness (B_p). If this were not the case, then the introspective judgement that one has one kind of experience rather than the other would in any case not be based on the way the experience is given in consciousness.²⁹ So, if the subject mistakenly comes to judge that her experience is B_p while in fact it is B_i, where could the source of the mistake lie? Could the way the experience is given in consciousness provide a reason for the subject to judge erroneously that her experience is B_p?

If a subject has a B_i-experience, then the experience provides him with a reason to believe that he has an experience of something with fuzzy edges. If the subject nevertheless finds a reason to judge that his experience is a B_p-experience, then this will typically come from supplementary considerations that undercut the rational power stemming from the fact that it is a B_i-experience. Typically, such considerations will concern the nature of the perceived object. If one believes that the object is of a kind that can't have fuzzy borders, then one will normally resist the idea that it appears blurred on some particular occasion.

The result above would be compatible with the fact that the undercutting considerations might have an impact on the phenomenal properties of the experience. The object might in fact stop to appear blurred when one comes to believe that it cannot be fuzzy. This kind of phenomenal contamination may occur not only by virtue of pressure coming from background beliefs. It may also emerge from relations between phenomenal properties: an object that appears to have fuzzy edges when seen on its own, may appear not to have such edges, but simply to be more distant, when seen in the background of a focussed object.

There finally is a very specific kind of error one might need to consider in our context. Remember the case mentioned above of an experience that is essentially relational and that presents itself in consciousness as such. The experience is, as we said above, a reflexive R-appearance. Now, suppose that the relation R itself constitutes a further phenomenal property, R*.³⁰ This would be a pure phenomenal property of the experience, not an appearance property. An experience of this kind, let us assume, would provide the subject with a reason to judge that his experience is indeed relational and with a reason to believe that it has the pure phenomenal property R*.

Suppose now an experience that presents itself in consciousness as standing in relation to an external object while it does not. We said that the experience would then mislead the subject about its own non-phenomenal nature. We called this kind of error an experiential error. But if the relation in question constitutes a phenomenal property, would we then not have to say that the experience misleads the subject about one of its phenomenal properties, and thus, when the phenomenal property under consideration is an essential one, about the experience's phenomenal nature?

It is again important to distinguish the pure phenomenal property R*, from the experience's property of being a reflexive R-appearance. If the two properties where not phenomenally distinct,³¹ then the presence of the latter would not be compatible with the absence of the former.³² It is further important to recognise that even if the relation R constitutes the phenomenal property R*, the fact that the experience is a self-reflexive R-appearance does not imply that it is a self-reflexive R*-appearance. This being so, the fact that the property of being a reflexive R-appearance can mislead the subject about the relational nature of the experience does not imply as such that it can mislead the subject about the phenomenal property that is constituted by the relation. The fact that an experience is given in consciousness as a reflexive R-appearance provides a reason for the subject to believe that the experience is relational, but not as such a reason to believe that it has the phenomenal property R*. In order to obtain this further reason the subject would have to be credited with a reason to believe that the phenomenal property R* is constituted by, or at least strongly correlated with, R. Such a supplementary reason may of course be available to the subject (if needed through philosophical investigation)³³, but it would not be a reason he has simply by virtue of the way the experience is given in consciousness. If the subject comes to form the false belief that he is having and R*-experience, the origin of his mistake cannot be the simple fact he has a reflexive R-appearance. The origin of his mistake must rather lie in the assumption that when one has a reason for believing that one's experience is relational, then one thereby obtains a reason for believing that it is R* by virtue of the fact that he property R* is constituted by the relation under consideration. A warrant concerning that constitutive constraint cannot be obtained simply by virtue of having the experience.

Phenomenal error, in the way we suggest to understand it,³⁴ would be a case of one being mislead about the phenomenal properties of one's own experience by the way the experience is given in the stream of consciousness. When the phenomenal properties are essential to the experience, phenomenal error would lead to a situation where the phenomenal nature of the experience is not manifest in consciousness. The arguments presented above appear to show that, strictly speaking, phenomenal error is impossible. If a subject has a false belief about a phenomenal property of her own perceptual experience, then the source of her mistake cannot lie in the way the experience is given in consciousness. A perceptual experience manifests each of its phenomenal properties in consciousness. This, we submit, is the sense one ought to give to the common claim that, for an experience, there is no gulf between being and appearing. Thus Husserl writes³⁵:

In the psychic sphere there is … no difference between appearance and being, and if nature is a being that appears in appearances, then appearances … are not themselves beings that appear in further appearances” (Husserl 1911: 311-12, our translation).

What Husserl suggests here is that, while a non-mental entity may “appear” in the sense of being the intentional objects of an experience, the experience itself “appears” in a different sense. And this may be identified with the way the experience is given in the stream of consciousness. Furthermore, what is distinctive of this second of kind of “appearing” is that, according to Husserl, there is indeed no distinction between the way an experience “appears” and the way it is: a way an experience “appears” is a way for it to be.

A similar claim is effective in Kripke's view, when he contends that:

To be in the same epistemic situation that would obtain if one had a pain is to have a pain; to be in the same epistemic situation that would obtain in the absence of pain is not to have a pain … (Kripke 1980: 152)

If «to be in the same epistemic situation that would obtain if one had a pain» means to be in a situation where the way an experience is given in consciousness constitutes a reason for the belief that one is in pain, then Kripke's claim is indeed that an experience cannot offer by the way it is given in consciousness an epistemic ground that misleads the subject about the experience's phenomenal properties. And if «to be in the same epistemic situation that would obtain in the absence of pain» means to be in a situation where the way an experience is given in consciousness does not constitute a reason to believe that one is in pain, then Kripke's claim is indeed that an experience cannot offer by the way it occurs in the stream of consciousness the epistemic ground for attributing to the experience a phenomenal property it does not posses.

These results should not be confused with the common claim that introspective beliefs are infallible:

Not only people seem to have a special epistemic access to their pains, they seem to have a very special epistemic authority with respect to their pain: they seem to be incorrigible, or even infallible, about their pains and pain reports: necessarily, if I sincerely believe that I am in pain, then I am in pain. Conversely, if I feel pain, then I know that I am in pain. Again this conditional seems necessarily true. This is the self-intimating aspect of pain experiences. (Aydede 2008)

As we have seen above, one can perfectly form a sincere incorrect belief about one of one's own experience. This may happen, for instance, as a consequence of phenomenal dissonance. The claim cannot be that necessarily, if I sincerely believe to be in pain, then I am pain. The claim should rather be that necessarily, if my reason to believe that I am in pain lies in the way the experience is given to me in consciousness, then I am in pain.

7. Where disjunctivism can lead us astray

Mike Martin, an influential advocate of disjunctivism, writes:

The idea that introspection will lead us into error about how things seem to us is hardly an attractive one. Yet given the considerations about phenomenal transparency, it is difficult to avoid. In contrast to the kind of global errors in introspection posited by sense-datum theories and intentional accounts, the disjunctivist can claim that veridical perceptual experiences are exactly as they seem to us to be: states in which parts of how the world is are manifest to us. But even the disjunctivist is forced to concede that we are misled about the nature of some of our experiences by introspection: after all, it can hardly be denied that it is possible for one to have an illusion or hallucination which is indistinguishable for one from a veridical perception. Given the disjunctivist's account of veridical perception, he is required to deny that such experiences are as they seem to us to be. Such experience is misleading not only about the world, but about its own nature. So in the end, sense-datum theories, intentional theories and disjunctivist accounts all have to endorse some form of error-theory concerning perceptual appearances and the introspection of experience. (Martin 2002: 421)

In the current context, the suggestion is that in the light of introspection, our experience of the world seems to have a certain nature, i.e. that articulated by naïve realism, which it does not have. An account of perception then needs not only to tell us what the nature of appearances is, but also how states of being appeared to can come to seem to be different from how they really are. (Martin forthcoming, Ch. 1: 41)

What sort of errors is Martin appealing to in these passages? What does he mean, when he says that «introspection will lead us into error about how things seem to us»? What sort of seeming is he appealing to when he writes that a disjunctivist «is required to deny that such experiences [hallucinations] are as they seem to us to be». Is Martin suggesting that these are cases of phenomenal error? We shall see that this need not be Martin's view. Once this is made clear, however, some of Martin's disjunctivist conclusions can bee seen to be compatible with the sort of intentionalism presented at the beginning of this paper.³⁶

In order to understand Martin's point it may be useful to recall the sort of errors he thinks the traditional alternatives to disjunctivism have been obliged to accept.³⁷ These errors concern two distinct phenomenal properties of perceptual experiences: transparency and actuality. In the light of the distinctions made above, transparency may now be characterized as the property of being an appearance of an object as external and mind-independent. So, for instance, in perception an object appears to stand in front of one. The corresponding phenomenal property is the property of being an appearance of something standing in front of one. The phenomenal property is not identical to the appearing property, but no intrinsic feature of the experience would need to be added to the appearing property in order to characterise the phenomenal property. The experience simply has the property of being an appearance of something outside of one: it is an externality-appearance. Actuality, on the other hand, is the phenomenal property of being an appearance of an existent object. We may say that in the experience the object appears to exist and that the phenomenal property is the property of representing an object as existent. The experience is, we may say, an existence-appearance. On Martin's view sense-data theorists maintain that in perception we are always in contact with existing objects, but that those objects do not possess most of the properties we represent them as having. They are sense data, so they are not really red, not really located in the external world, etc.³⁸ The intentionalist, instead, is said to claim that the objects of perception may be correctly represented as red, in front of us, etc., but sometimes, i.e. in the case of hallucination, they simply do not exist.

What sort of errors are those theories then supposed to admit? As mentioned above, actuality and transparency are among the phenomenal properties of perceptual experiences. It appears, in fact, that all perceptual experiences possess those properties. An experience that lacks the phenomenal property of transparency is not an experience of something appearing to one as external and mind-independent. And an experience that lacks actuality is not an experience of there appearing to be something existent one stands in relation to. In the first case we might have something like a pure sensation,³⁹ in the second case we might have an imagination. So transparency and actuality are not only pervasive; they are more fundamental, too, since they contribute to the determination of the kind of experience under consideration (e.g. perception as opposed to imagination and sensation). But transparency and actuality are not pure phenomenal properties: they are both associated with correctness conditions that concern not only the external world, but the experiences themselves. Transparency concerns the nature of some (if not all) phenomenal properties of perceptual experiences, actuality concerns the relational character of perception. There is thus room for error concerning the satisfaction of those correctness conditions. The sense data theorist can now be seen as arguing that the objects of perception not only do not generally possess the properties they appear to possess, but that because of transparency perceptual experiences present those properties in a systematically misleading way, namely as properties of external objects. It is not just the fact that the objects that appear to be red, square, distant from us, etc., do not really possess those properties; it is much more the fact that the objects of perception are of a kind that could not possess them. The intentionalist, on the other side, can be seen as arguing that perceptual experiences misleadingly present themselves as standing in relation to existing things, as experiences whose instances depend on there actually being an object one stands in perceptual relation to. So the errors both theories appeal to concern the correctness conditions associated to specific phenomenal properties. In so far as those phenomenal properties are essential to perceptual experiences, we obtain cases where essential phenomenal properties misrepresent the experiences themselves.

Do these errors concern the nature of the experiences? Two important assumptions are at work at this stage. First, both theories maintain that perceptions and hallucinations belong to the same fundamental (phenomenal) kind: they share their essential phenomenal properties. And second, both theories assume that if perceptions and hallucinations have the same phenomenal nature, then they ought to have the same non-phenomenal nature, too. And here the two theories part company. Sense-data theorists maintain that perception is essentially relational; intentionalists argue that it is not. Intentionalists submit that (most) phenomenal properties of the perceptual experience are fully determined by the appearing properties of the object, sense-data theorists deny it. On both theories the essential phenomenal properties of perceptual experiences involve an error concerning their non-phenomenal nature. It is not a phenomenal error since the mistake concerns precisely the non-phenomenal nature of the experience. It is rather a case of an essential phenomenal property that misrepresents the non-phenomenal nature of the experience. This generates a mismatch between phenomenal and non-phenomenal nature – an experiential error.

Among the different reasons one may have to be unsatisfied with intentionalism and sense-data theory there is the fact that they appear not to be in a position to articulate a basic intuitive asymmetry between perception and hallucination. There is a sense in which there is nothing wrong with perception, but much wrong with hallucination. On the conjunctive solution there would be something wrong with both, perception and hallucination. At this point disjunctivism can be seen as a view that aims to restore and articulate the intuitive asymmetry. Indeed, along with the naïve realist, the disjunctivist will typically argue that in perception we do enter into a relation with an external object, and that most (if not all) phenomenal properties of the perceptual experience are determined by this relation to the external object.⁴⁰ Accordingly, transparency and actuality correctly represent the non-phenomenal properties in the case of perceptions. The asymmetry would thus be explained by the fact that the same is not true of hallucinations.

In this line of thought the disjunctivist would assume that hallucinations share their phenomenal properties with perceptions, notably their transparency and actuality. But he would then have to maintain that perceptions and hallucinations differ in their non-phenomenal nature. This requires treating their difference with respect to their relation to the external object both as non-phenomenal and as essential. Transparency and actuality would thus correctly represent the non-phenomenal nature of perceptual experiences; while they would misrepresent the non-phenomenal nature of hallucinations.

But if hallucinations and perceptions are experiences of a different fundamental kind by virtue of their different non-phenomenal natures, how can they share their phenomenal properties? If one accepts the two assumptions mentioned above, the assumptions that are common to intentionalism and sense-data theory, then the version of disjunctivism under consideration finds itself in an uncomfortable position. If perception and hallucination possess different non-phenomenal natures then, by virtue of the second assumption, they should also have different phenomenal natures. This, however, is precisely what the first assumption denies.⁴¹

The disjunctivist might be willing to challenge this assumption. Martin has indeed insisted on the fact that the indiscriminability of experiences does not imply their phenomenal identity.⁴² Whatever the merits of that argument, it is important to see that on pain of admitting phenomenal error, the disjunctivist needs to accept that the mistaken judgement cannot be based on the way the experience is given in consciousness. If one cannot discriminate a hallucination from a perception, although they are supposed to be phenomenally different, then this cannot have its source in the way the hallucination is given in consciousness.⁴³ The hallucination cannot be said to present itself in consciousness as a perception. So the disjunctivist would need to show that the error has an independent source.

Considerations related to acquaintance might be relevant at this stage. Remember the case of the pure phenomenal relational property R* mentioned above. Acquaintance might be such a property. The argument could then be that hallucinations lack indeed that essential phenomenal property and thus possess a different phenomenal nature than perceptions. We are mislead about the phenomenal nature of hallucinations, not because of the way they are given in consciousness, but because of background assumptions concerning the relation between phenomenal properties hallucinations share with perceptions and acquaintance. We typically associate, for instance, the phenomenal properties of actuality and transparency with acquaintance. This is the source of our false introspective belief about the phenomenal nature of hallucination.

But this proposal would have to be confronted with the possibility of removing the background assumptions under consideration. If the reason I have to believe that my present hallucination has the phenomenal property of being an acquaintance, although it is not given as such in my consciousness, is the presence of some background assumption, then discovering that those assumptions are misleading should open the experience to my cognitive scrutiny. One would then expect that the phenomenal difference between perception and hallucination should become accessible to my introspective scrutiny. But this does not seem right. Hallucinations do not show their supposedly true phenomenal nature once one ceases to fall prey to the cognitive illusion generated by some background assumptions. The disjunctivist must thus postulate a brute and thoroughly inaccessible phenomenal difference that would be responsible for the distinction in phenomenal nature of the experiences.

This is the point where the intentionalist may be in a position to offer an attractive alternative. Instead of challenging the first assumption common to traditional sense-data theory and intentionalism, he could suggest to forgo the second. He would thus argue that although perception and hallucination do in fact have the same phenomenal nature, they have different non-phenomenal natures. Perceptions are indeed states in which we enter in contact with external objects in such a way that this relation determines how the objects appear. In contrast, there are no external objects to which one stands in relation when hallucinating, but hallucinations present themselves as if they were relations to such objects. Hallucinations misrepresent their non-phenomenal nature. Again, this would not be a case of a phenomenal error, but a case of an experiential error.

The brand of intentionalism under consideration can thus readily admit that there is an asymmetry between perception and hallucination: perception manifests its true (non-phenomenal) nature in consciousness while hallucination does not. Indeed, a hallucination misleadingly provides one with a reason to believe that it is a perception, whereas a perception does not provide one with a reason to believe that it is a hallucination. The intentionalist can explain this asymmetry with the fact that the phenomenal properties of actuality and transparency are themselves intentional, they have correctness conditions that are satisfied by perceptions, and not by hallucinations.

We can now understand why phenomenologists such as Husserl seem to have been rather agnostic about disjunctivism concerning the non-phenomenal nature of experiences. Husserl was quite clear about the actuality and transparency of perception.⁴⁴ In the Logical Investigations he also states that in perception, in opposition to hallucination, the object itself is given to us. Yet this fact, the fact that in perception we stand in relation to an object, does not constitute the phenomenal character of the perceptual experience. In the light of what we saw so far we know why.⁴⁵

8. Bibliography

Aydede, Murat 2008: 'Pain'. In The Stanford Encyclopedia of Philosophy. Zalta, Edward N. (ed.).
Boghossian, Paul A. and Vellemen, J. David 1998: 'Colour as a Secondary Quality', Mind, 98 pp. 81-103.

Brentano, Franz 1995: Psychology from an Empircal Standpoint. London: Routledge. First published: Psychologie vom empirischen Standpunkt, Leipzig: Duncker & Humbolt, 1974.

Burge, Tyler 1991: 'Vision and Intentional Content'. In Lepore, Ernest & Van Gulick, Robert (eds.) 1991: John Searle and His Critics. Oxford: Backwell, pp. 195-213.

Crane, Tim 2001: Elements of Mind. Oxford: Oxford University Press.

Crane, Tim 2006: 'Brentano's concept of intentional inexistence'. In Textor, Mark (ed.) 2006: The Austrian Contribution to Philosophy. London: Routledge, pp. 20-35.

Dorsch, Fabian forthcoming: The Phenomenal Error in Hallucinating, In: Fiona Macpherson (ed.), Hallucinations, MIT Press, 2010.

Dretske, Fred 1971: 'Conclusive reasons', Australasian Journal of Philosophy, 49 pp. 1-22. Reprinted: Dretske, Fred 2000: Perception, Kowledge and Belief. Selected Essays. Cambridge: Cambridge University Press, 3-29.

Dretske, Fred 1988: Explaining Behavior. Reasons in a World of Causes. Cambridge (MA)/ London: Bradford/MIT Press.

Dretske, Fred 2005: 'Is Knowledge Closed under Known Entitlement? The Case against Closure'. In Steup, Matthias & Sosa, Ernest (eds.) 2005: Contemporary Debates in Epistemology. Oxford: Blackwell, pp. 13-26.

Hellie, Benj 2006: 'Beyond Phenomenal Naiveté', Philosophers' Imprint, 6 (2),

Hinton, J M 1973: Experiences. Oxford: Oxford University Press.

Humberstone, Lloyd 1992: 'Direction of Fit', Mind, 101 pp. 59–83.

Husserl, Edmund 1911: 'Philosophie als strenge Wissenschaft', Logos, 1 pp. 289-341. Reprinted: Husserl, Edmund 1965: Philosophie als strenge Wissenschaft. Frankfurt: Klostermann.

Husserl, Edmund 1970: Logical Investigations Volume 2. London: Routledge & Keagan.

Husserl, Edmund 1970: Logical Investigations Volume 2. London: Routledge & Keagan.

Husserl, Edmund 1984: Logische Untersuchungen. Zweiter Band. Zweiter Teil. Untersuchungen zur Phänomenologie und Theorie der Erkenntnis. Den Haag: Martinus Nijhoff. In: Edmund Husserl: Gesammelte Werke (Husserliana), Vol. XIX/2. Editor: Panzer, Ursula. First published: Husserl, Edmund 1901: Logische Untersuchungen I-VI. Halle: Niemayer.

Husserl, Edmund 1984: Logische Untersuchungen. Zweiter Band. Zweiter Teil. Untersuchungen zur Phänomenologie und Theorie der Erkenntnis. Den Haag: Martinus Nijhoff. In: Edmund Husserl: Gesammelte Werke (Husserliana), Vol. XIX/2. Editor: Panzer, Ursula.First published: Husserl, Edmund 1901: Logische Untersuchungen I-VI. Halle: Niemayer.

Kripke, Saul A. 1980: Naming and Necessity. Oxford: Blackwell.

Martin, Michael 2002: 'The Transparency of Experience', Mind and Language, 17 (4), pp. 276-425.

Martin, Michael 2002: 'The Transparency of Experience', Mind and Language, 17 (4), pp. 276-425.

Martin, Michael 2004: 'The Limits of Self-Awareness', Philosophical Studies, 120 pp. 37-89.

Martin, Michael 2004: 'The Limits of Self-Awareness', Philosophical Studies, 120 pp. 37-89.

Martin, Michael forthcoming: Uncovering Appearance.

Martin, Michael G. F. 1998: 'Setting things before the mind',

McDowell, John 1998: Mind, Knowledge and Reality. Cambridge, MA: Harvard University Press.

Moore, George E. 1993: 'Sense-Data'. In Moore, George E. (ed.) 1993: Selected Writings. London: Routledge, pp. 54-58. First published: Moore, George E. 1953: Some Main Problems of Philosophy. London, 28-40.

Pace, Michael 2007: 'Blurred Vision and the Transparency of Experience', Pacific Philosophical Quarterly, 88 (3), pp. 328-254.

Putnam, Hilary 1994: 'Sense, Nonsense, and the Senses', Journal of Philosophy, 91 pp. 445-517.

Royal Institute of Philosophy supplement, 43 pp. 157-79.

Searle, John 1983: Intentionality. An Essay in the Philosophy of Mind. Cambridge: Cambridge University Press.

Siegel, Susanna 2004: 'Indiscriminability and the Phenomenal', Philosophical Studies, 120 pp. 90-112.

Smith, Michael 1994: The Moral Problem. Oxford: Blackwell.

Snowdon, Peter 2002: 'Perception, Vision and Causation'. In Noe, Alva & Thomson, Evan (eds.) 2002: Vision and Mind. Cambridge, MA: MIT Press, pp. 151-66. First published: Snowdown, Peter 1980-81: 'Perception, Vision and Causation', Proceedings of the Aristolelian Society, 81 pp. 175-92.

Soldati, Gianfranco 1996: 'Philosophische Probleme der Repräsentation'. In Braitenberg, Valentin & Hosp, Inga (eds.) 1996: Die Natur ist unser Modell von Ihr. Forschung und Philosophie. Hamburg: Rowohlt, pp. 131-54.

Soldati, Gianfranco 2008: 'Die Objektivität der Bedeutung (I. Logische Untersuchung, §§ 24–35) '. In Mayer, Verena (ed.) 2008: Edmund Husserl. Logische Untersuchungen. Berlin: Akademie Verlag, pp. 61-76.

Soldati, Gianfranco 2009: 'Transparenz der Gefühle'. In Merker, Barbara (ed.) 2009: Emotionen interdisziplinär. Paderborn: Mentis,

Tenenbaum, Sergio 2006: 'Direction of Fit and Motivational Cognitivism'. In Shafer-Landau, Russ (ed.) 2006: Oxford Studies in Metaethics. Oxford: Oxford University Press, pp. 235-64.

Tye, Michael 2003: 'Blurry Images, Double Vision, and Other Oddities: New Problems for Representationalism?'. In Smith, Q. & Jokic, A. (eds.) 2003: Consciousness: New Philosophical Perspectives. Oxford: Clarendon Press, pp. 69-97.