1. The Two Views

By the Representational View of visual experience, Campbell seems to have in mind the following four-fold view: (a) what it is for an experience to be a visual experience is for it to have one of the members of a family of highly determinable phenomenal characters; (b) visual experience is common to visual perceptual experience and visual hallucinatory experience in that a visual perceptual experience and a visual hallucinatory experience could have exactly the same completely determinate phenomenal character; (c) whether a visual experience is perceptual or hallucinatory depends on how it came about; and (d) the phenomenal character of a visual experience, whether the experience is perceptual or hallucinatory, is, or is constituted by, its representational content.Footnote ³

Theses (a)–(c) comprise the traditional view of visual experience, a view that, it is fair to say, remains the dominant view in the United States. Given (b), the view is a ‘common factor’ view: it entails that a visual perceptual experience and a visual hallucinatory experience can share a common experiential factor. Common factor views contrast with disjunctive views, according to which a visual perceptual experience and a visual hallucination share no common experiential factor. As will emerge shortly, the Relational View is a disjunctive view.

Some theorists embrace (a)–(c), but deny (d). Thesis (d) is entailed by so-called representational theories of the phenomenal characters of visual experiences.Footnote ⁴ Proponents of such theories try to justify them, in part, by appeal to the well-known phenomenological thesis of the transparency (or diaphanousness) of visual experience.Footnote ⁵ That thesis has no canonical formulation. I'll use the following formulation: what it is like for one to have a visual experience is for it to be to one as if one is presented with a scene. This what-it-is-like aspect of a visual experience is its phenomenal character. According to the Representationalist, that phenomenal character is or is constituted by the representational content of the experience. Representational theories can differ markedly in their account of what it is for an experience to have a representational content. But they agree that representational contents are semantic contents, in that they are satisfied or instead fail to be satisfied. A visual experience is veridical if its representational content is satisfied; non-veridical if its representational content fails to be satisfied. And its representational content is satisfied just in case matters are as having the visual experience presents matters to the subject as being. A view is a Representational View if and only if it entails (a)–(d). It is not my concern here whether the Representational View is correct. My concern is only with whether Campbell has succeeded in making a case that the Relational View is superior to it.

In the remainder of this section, I'll examine Campbell's presentation of the Relational View in detail, using a scalpel. The aim of the examination is to pinpoint the fundamental, essential difference between the Representational View and the Relational View.

Discussing a case of having a visual perceptual experience as you look around a room, Campbell tells us:

A few pages later, he elaborates, saying:

It is Campbell's considered position that the egocentric layout of the scene a visual perceiver sees indeed constitutes the phenomenal (or qualitative) character of the perceiver's visual experience of it.

In the chapter following the one from which the above quotes are taken, Campbell says:

He writes in the passage of ‘the phenomenal content of the experience’, but he doesn't mean here a kind of content that is satisfied or that fails to be satisfied depending on whether matters are as the content characterizes them as being. He uses ‘phenomenal content’ here just as a synonym for ‘phenomenal character’. Campbell says nothing further, in addition to what he says in the passages above, by way of explicating the notion of a view or scene, or by way of explicating the notion of the egocentric or viewpoint relative layout of the objects in a scene. Although much can be said about these matters, I'll simply assume here that the basic idea of a scene is clear enough for present purposes. I will take it, then, that the Relational View that he takes to be superior to the Representational View is the view that the phenomenal character of a visual perceptual experience is the scene that the perceiver sees, and so consists of the objects that the perceiver sees, some of their properties, and how they are spatially arranged at the time in question relative to a spatial position suitably occupied by the perceiver at that time.

It is common ground between the Representational View and the Relational view that one has a visual perceptual experience if and only if one sees a scene. On the Relational View, the phenomenal character of a visual perceptual experience is the scene that the perceiver sees. The phenomenal character of a visual experience is what it is like for the subject to have the experience. Where the experience is a visual perceptual experience, the Relationalist maintains that what it is like for the subject to have the experience is what the scene experienced is like from the viewpoint occupied by the perceiver, its layout from that viewpoint. The Representationalist will agree, but will say that that is so because the perceptual experience is (completely) veridical. The phenomenal character of the experience is its representational content, and its representational content is satisfied by the scene.

It is common ground between the Representational View and the Relational View that when one has a visual (completely) hallucinatory experience, there is no scene that one sees. But Representationalists maintain that visual hallucinatory experiences have phenomenal characters. When one has a visual hallucinatory experience, it is like something for one to have the experience. What it is like to have a (vivid) visual hallucinatory experience is for it to be as if there is a scene before one. According to the Representationalist, visual hallucinatory experiences have representational contents, and those contents are their phenomenal characters. It is open for a Relationalist to hold that as well. Indeed, on a charitable reading, Campbell, a proponent of the Relational View, holds that. Consider how Campbell compares and contrasts a case of visually perceiving a dagger with a case of visually hallucinating a dagger. He says:

I think Campbell writes loosely when he writes of ‘the hallucinatory dagger’. One can hallucinate a dagger, but there are no such things as hallucinatory daggers. Campbell would, I believe, agree with that claim. The Representationalist claims that what makes an hallucination an hallucination of a dagger is its representational content. Campbell uses the locution ‘having a visual experience as of something being F’ in characterizing visual hallucinatory experience.Footnote ¹⁰ When one visually hallucinates a dagger, one has a visual experience as of a dagger. Moreover, Campbell writes of an hallucinator ‘seeming to see objects’; thus, he says: ‘Even if I am hallucinating, the objects I seem to see…’.Footnote ¹¹ Someone visually hallucinating a dagger, ‘seems to see’ a dagger, which, I take it is intended as just another way of saying that the hallucinator has a visual experience as of a dagger. What Campbell seems to have in mind by saying that a seen dagger ‘is matched as closely as possible by the hallucinatory experience [of a dagger]’Footnote ¹² is that the dagger matches, or satisfies, the general representational content of the hallucination of a dagger. The Relational View, as Campbell presents it (and that is all that I have to go on), is perfectly compatible with the Representationalist account of the nature of hallucinatory experience.

What Relationalists and Representationalists disagree about is the nature of visual perceptual experience. Campbell often states the disagreement this way: on the Relational View, but not on the Representational View, if one has a visual perceptual experience of an object, then that object is a constituent of the experience. Thus, in discussing the difference between visually perceptually experiencing a dagger and visually hallucinating a dagger, he says, for instance:

And, as concerns the Relational View, he says: ‘[t]he idea is that visual experiences are relational: the object perceived is a constituent of the conscious experience itself’.Footnote ¹⁴ Given that the object is a constituent of the experience, it is not (no state of it is) a cause of the experience. Campbell notes, though, that the object will be a cause of a brain state of the subject.Footnote ¹⁵

This contrast, however, doesn't capture the essential difference between the Representational View and the Relational View. The reason is that it is open to a Representationalist to maintain that when one visually perceptually experiences an object, the object is a constituent of the experience. Visually perceptually experiencing an object is an experiential relation that a perceiver bears to the object – it is the relation of seeing the object. The object is a relatum of that relation. The Representationalist maintains that one bears that relation to an object by having a visual experience with a certain representational content and the object's bearing an appropriate causal connection to that visual experience. That leaves open, however, whether the visual perceiving of the object is identical with the visual experience with the representational content or is instead distinct from it, but generated by it.Footnote ¹⁶ The Representationalist can go either way on that issue. Which way a Representationalist goes will depend on where he or she stands on the issue of whether (to use a phrase of Jonathan BennettFootnote ¹⁷) identity holds across the by-relation. The issue is familiar from action theory. Suppose that S kills X by fatally shooting X (by shooting X in a way that appropriately results in X's death). On the view that identity holds across the by-relation, the killing is identical with the shooting. On the alternative view, the killing and the shooting are distinct actions, but the latter generates the former in the circumstances in question (circumstances in which the shooting appropriately brings about the death of X). The issue turns on general problems about the individuation of states and events. A Representationalist can go either way on the issue. Thus, it is open to a Representationalist to maintain that if one visually perceives an object, the object is a constituent of the relational experience of visually perceiving it. The object is, then, not a cause of the perceptual visual experience, though it is, on this view, a cause of a visual experience with a representational content.

As concerns the Relational View, Campbell says:

It doesn't, however, get at the essential difference between the views to say that the Relationalist holds that ‘S perceives O’ is primitive, while the Representationalist maintains that it is analyzable in such terms as ‘O causes S to have an experiential content as of something's being G’. An analysis requires non-circular necessary and sufficient conditions. It is wide open to a Representationalist to maintain that we cannot state non-circular necessary and sufficient conditions for someone's perceiving an object. As Campbell notes, Representationalists standardly maintain that if S (visually) perceives O, then (i) O exists, (ii) S has an experience with a certain experiential content, and (iii) O (or a state of O) figures as a cause of that experience. But while that is true, (i)-(iii) do not jointly suffice for S's perceiving O. It is a familiar point that there are many objects whose states figure as causes of an experience yet are not objects of the experience. A Representationalist need not maintain that there are non-circular conditions that are individually necessary for, and jointly sufficient for, S's perceiving O. Think of how truly few concepts there are for which there are non-circular necessary and sufficient conditions. Representationalists reject disjunctivism. But they need not embrace conjunctivism, the view that S's perceiving O can be analyzed as the conjunction of certain conditions.Footnote ¹⁹

The fundamental, essential difference between the Representational View and the Relational View is, rather, this: on the Representational View, one visually perceptually experiences an object by having a visual experience with a representational content and the object's bearing an appropriate causal connection to that visual experience. On the Relational View, that is false. Indeed the Relationalist denies that having a visual perceptual experience of an object involves having a visual experience with a representational content. As Campbell puts it:

If the Relational View is right, then a visual perceptual experience could not involve a visual experience that has a phenomenal character in the sense in which an hallucinatory experience has a phenomenal character. We are left with only a disjunctive notion of the phenomenal character of a visual experience. The phenomenal character of a visual perceptual experience is a scene. The phenomenal character of a visual hallucinatory experience is a representational content.

Although Campbell is a Relationalist, he thinks that seeing an object involves psychological factors. These psychological factors include cognitive processing. Alluding to the vision science thesis that the visual system employs feature maps to solve binding problems, he says:

He further points out that the cognitive processing in question is information processing, involving cognitive states that have content. And he notes that this invites the view that:

But he holds that although the cognitive contents function as common factors, no experiential contents do. Indeed, he holds that there is no experiential factor that is common to perception and hallucination. To repeat: the Relational View is a disjunctive view.

Let's turn now to why Campbell thinks the Relational View is superior to the Representationalist View.Footnote ²³

2. An Outline of Campbell's Case and a Diagnosis

Campbell maintains that in adjudicating the dispute between the Relational and the Representational View,

The common core notion of the phenomenal character of an experience is this: the phenomenal character of an experience is what it is like for the subject to have the experience. Campbell is precisely right about the way to proceed. He takes the role the notion plays in our reflective thinking to be an explanatory role; and that is right too. We invoke the notion for various explanatory purposes; we invoke it in order to explain certain things. If the Relational View can reveal how the phenomenal character of experiences can serve the explanatory purposes in question, and the Representational View cannot do so, then that would be a compelling reason indeed in favor of the Relational View over the Representational View.

As may already be apparent, however, it is curious indeed that Campbell asks why we need the notion of the phenomenal character of experience, given that he is committed to the view that the notion covers two different sorts of things. The phenomenal character of a visual perceptual experience is a scene; the phenomenal character of a visual hallucinatory experience, a kind of representational content. What reason is there to think that there is an explanatory role that such a disjunctive notion plays?

In fact, Campbell goes on to ask and then to try to answer a question that is only relevant if one takes the phenomenal character of an experience to be a scene, and so to be something that a visual hallucinatory experience lacks. Although he says that we should proceed by asking and then answering the question why we need the notion of the phenomenal character of experience, he instead proceeds to ask and answer a question that the Representationalist regards as a very different question. He proceeds to pursue the question of why we need the notion of ‘an experience of an object’. By an experience of an object he means conscious attention to an object, in the sense of one's conscious attention being focused on an object.Footnote ²⁵ Thus, he says: ‘we use the notion of experience of an object – or more precisely, conscious attention to an object’.Footnote ²⁶ More precisely still, he means one's conscious visual attention being focused on a seen object – a visual perceptual experience of an object in which, as he puts it, the object is ‘highlighted’ in the experience by one's focus of attention.Footnote ²⁷ One has a visual perceptual experience of an object if and only if one sees the object. One can see an object without one's conscious visual attention being focused on it. But Campbell has in mind attentively seeing an object. Hereafter, I'll follow Campbell in using ‘experience of an object’ in this way, as short-hand for one's conscious visual attention being focused on an object that one sees. In the present context, I'll also follow him in sometimes using ‘conscious attention to an object’ as short-hand for that; and I'll sometimes use ‘visual attention to an object’ in that way too.

On the Relational View, the phenomenal character of a visual perceptual experience is the scene that the perceiver sees. Moreover, on the Relational View, the phenomenal character of an experience of an object is the object itself and certain of its properties, all of which are elements of the scene that the subject sees. Thus, on the Relational View, the question of why we need the notion of an experience of an object is one of the questions we must answer to answer the question why we need the notion of the phenomenal character of an experience. From the Representationalist's point of view, however, the question why we need the notion of an experience of an object is a very different question from the question of why we need the notion of the phenomenal character of experience. On the Representational View, one can have an experience with a phenomenal character, even when one fails to experience any object at all (as in a case of complete hallucination).

Campbell cites, by my count, basically four explanatory roles that he maintains the notion of an experience of an object is supposed to play. Here they are, presented mainly in his own words:

One issue is whether experience of objects is supposed to explain all or indeed any of the matters in question. A second issue is whether, even if it is, what that is supposed to have to do with the dispute between Relationalism and Representationalism. In section 3, I'll address the first issue at length. For the remainder of this section, I'll focus on the second issue.

Campbell thinks that the Representational View can show us how experience of an object can play explanatory roles 1–4 only if it can show us how an experience's having a certain representational content can play those roles. He thinks that because he thinks this:

Campbell offers no justification whatsoever for this claim, however. He just assumes it. In fact, the Representationalist has no such commitment. Indeed the Representationalist will flatly reject the claim that whenever an experience explains something, it does so in virtue of its representational content. The Representationalist will flatly reject that, even if the Representationalist holds the thesis of identity across the by-relation. Even such a Representationalist holds that the fact that an experience is an experience of a certain object explains certain matters, and that the fact that the experience has a certain representational content explains certain other matters. The Representationalist maintains that the former can explain things the latter cannot explain, because the latter does not entail the former. The fact that an experience has a certain representational content will not entail that the experience is an experience of an object. Thus, from the fact that an experience's being an experience of an object explains something, it does not follow that the experience's having a certain representational content explains that something. Moreover, the Representationalist maintains that to invoke the fact that an experience is an experience of an object in the explanation of matters properly explained by the fact that the experience has a certain representational content would be to invoke superfluous information.

In discussing the explanatory role of experiencing an object, Campbell says at one point: ‘experience of the object has to be enough to guarantee that the object exists’.Footnote ³³ But that is no problem whatsoever for Representationalism. If one experiences O (i.e. consciously visually attends to O), then O exists. For one can experience (i.e. consciously attend to) only what exists. Instances of the schema ‘S experiences O’ (‘S consciously visually attends to O’) entail corresponding instances of the schema ‘O exists’.

Here, then, is my diagnosis of where Campbell fundamentally goes wrong in making his case that the Relational View is superior to the Representational View. He makes the mistaken assumption that according to the Representationalist, whatever an experience explains in virtue of being an experience of an object (or of a certain object), the experience explains in virtue of its representational content. That, as I noted, is a claim the Representationalist flatly rejects, on the ground that the fact that an experience has a certain representational content does not entail that the experience is (or generates) an experience of an object. Consider an experience of a certain object, and let us suppose, as Representationalists typically do, that that experience itself has a representational content, rather than being generated by an experience with a representational content. According to the Representationalist, the experience of the object will explain certain things in virtue of being an experience of that object, and the experience will explain certain other things in virtue of having a certain representational content. Being an experience of a certain object and being an experience with a certain representational content play different explanatory roles; they are invoked for different explanatory purposes. That is directly relevant to evaluating Campbell's case against the Representational View. Even if the notion of an experience of an object plays explanatory roles 1–4, and the notion of an experience's having a representational content does not play any of those roles, that fact alone would be no reason to favor the Relational View over the Representational View.

Still, though, suppose that experience of an object is supposed to play one or more of the explanatory roles specified in 1–4. If experiencing an object plays the role on a Relationalist conception of it, but fails to play to it on a Representationalist conception of it, then that would be a reason in favor of Relationalism over Representationalism. So, in the next section, I'll examine 1–4 in detail. And I'll argue that Campbell fails to identify any such reason for favoring Relationalism over Representationalism.

3. Would-Be Explanatory Roles 1–4 Examined

Consider, again, explanatory role 1: ‘Experience of objects has to explain how it is that we can have the conception of objects as mind-independent’.Footnote ³⁴ It is a tall order indeed for experience of objects to explain that. It seems that many kinds of animals can focus their conscious visual attention on objects they see (think of an animal visually tracking its prey), yet are such that they neither have a conception of objects as mind-independent nor, for that matter, a conception of objects as mind-dependent; indeed such a conception seems cognitively closed to them. That is certainly so if having a conception of objects as mind-independent requires having a conception of mind. But it is uncertain whether it does. Campbell does not elaborate on what he means by the conception of objects as mind-independent.

Campbell tells us:

The Representational View, however, is a kind of common factor view, and the above remarks are a serious mischaracterization of it. On the Representational View, an experience of an object relates one to an object, not to an image of an object. Indeed, it is no part of the Representational View that experiences of objects even involve images. Also, as I noted, it is even open to a Representationalist to maintain that an experience of an object includes the object as a constituent (if the Representationalist rejects the thesis of identity across the by-relation). Further, let's say that one indirectly sees an object if and only if one sees the object by seeing something else; and that one directly sees an object if and only if one sees the object, but not indirectly. The Representationalist, like the Relationalist, maintains that we can directly see objects.Footnote ³⁶ To be sure, on the Representational View, one sees an object by having an experience with a certain representational content. But of course one doesn't see the experience; the experience doesn't look any way to one; moreover, the experience isn't an object of visual attention. That is so even in cases of complete hallucination. In such cases, the subject doesn't see anything; nothing looks anyway to the subject; there is nothing that is an object of visual attention. It is just that it is experientially for the hallucinator as if there is an object before him.

The Representationalist, you will recall, embraces the phenomenological thesis of the transparency of visual experience. What it is like to have a visual experience is what it is like to view a scene. As I mentioned earlier, Representationalists typically appeal to the transparency thesis to help justify the view that the phenomenal character of a visual experience is a representational content. Although Campbell never mentions the transparency thesis by name, it is clear from his discussion that he takes it to be at least relevant to explanatory role 1. Again calling a visual experience in the Representationalists sense an ‘image’, Campbell says:

I take it that he has in mind here that a common factor theorist such as a Representationalist can appeal to the transparency thesis. He readily acknowledges that the Representationalist can make the above reply. But he says:

It is true that, on the Representational View, experiences have intentionality or aboutness: a visual experience, whether perceptual or hallucinatory, will have a representational content, and a representational content is a kind of intentional content. But it is not part of the Representational View that experience is a way of grasping thoughts, or that to see an object is to grasp a demonstrative proposition. Indeed, the leading Representationalists deny both that experience is a way of grasping thoughts and that to see an object is to grasp a demonstrative proposition.Footnote ³⁹ Grasping thoughts and grasping demonstrative propositions are conceptual affairs. While we normally exercise concepts while having a visual experience, the leading Representationalists maintain that the representational content of an experience that is its phenomenal character is a non-conceptual content.Footnote ⁴⁰ To take a stock example from the literature, something can look red and at a right-angle to one, even when one lacks both the concept of redness and the concept of a right angle. The Representationalist can hold that in virtue of their non-conceptual contents, visual experiences, whether perceptual or hallucinatory, ‘display the world as objective’ (to use Campbell's phrase).

The fact that Representationalists can (and typically do) maintain that the relevant representational contents are non-conceptual contents bears on what Campbell takes to be the difficulty that Representationalism has in explaining how experience of an object can achieve explanatory role 2. I'll turn now to that matter. Explanatory role 2 is, you will recall, this:

Campbell says:

As I noted, however, although Representationalists maintain that experiences are intentional, the leading Representationalists deny that they are ways of grasping thoughts. To repeat: they maintain that the representational contents of experiences that are their phenomenal characters are non-conceptual contents.

Campbell is aware that a Representationalist can maintain that the representational contents in question are non-conceptual contents. He mentions that possibility in connection with a specific proposal of Tyler Burge.Footnote ⁴³ I won't pause here to discuss Burge's proposal. For present purposes, the first point to note is just that Burge posits a demonstrative element in visual experience, and Campbell notes that ‘a “demonstrative element” can be as it is, whether or not it refers to an object, and independently of which object it refers to’.Footnote ⁴⁴ That point is certainly correct. The second point to note is that he acknowledges that it might be maintained that

He immediately goes on to say, however:

It is certainly true that the demonstrative element (if such there be) cannot of itself distinguish the presentation of one object from the presentation of another, and that it cannot of itself even guarantee that it has any referent at all. What, if any, referent it has will be determined by the relevant contextual factors. But a point that I have been belaboring is that a Representationalist can appeal, in an explanation, directly to the fact that we experience objects – that is, directly to the fact that our visual attention can be focused on an object that we see.

In spelling out explanatory aim 2, Campbell tells us:

Here he makes the mistake that I have repeatedly underscored. He mistakenly assumes that the Representationalist is committed to the view that whatever an experience of an object explains, it explains in virtue of its representational content. To repeat: not only is the Representationalist not committed to that, the Representationalist flatly denies it.

Our ability to experience an object – that is, our ability to consciously visually attend to a seen object – will not by itself explain our ability to think demonstratively about an object we see. The reason is that the latter ability involves the ability to exercise demonstrative concepts in conscious thought, and the former doesn't. The fact that we can consciously attend to seen objects thus won't by itself explain how we can form demonstrative thoughts about seen objects. It will, however, at least figure in the explanation. When one experiences an object – when one's conscious visual attention is focused on a object that one sees – and one uses a demonstrative in conscious thought to try to demonstrate the object that one's attention is focused on, the demonstrative demonstrates that object. I take that to be common ground between the Relationalist and the Representationalist.

Let us turn to explanatory role 3, which is, you will recall:

I find this would-be explanatory role difficult to assess, because I find it obscure what Campbell means by ‘knowledge of the reference of a demonstrative’. One thing he sometimes seems to mean by it is knowing which object is demonstrated by a demonstrative. Knowing which object is demonstrated is a kind of knowing-that. One knows which object is demonstrated when one knows of some object that it is the object that is demonstrated. One's conscious attention being focused on an object that is the referent of a demonstrative could, of course, causally contribute to one's coming to know which object is demonstrated. (So, for that matter, could turning one's head in a certain direction.) But the fact that one's conscious visual attention is focused on an object that is the reference of a demonstrative could not possibly, by itself, explain how one has such de re knowledge-that. Still it can figure in the explanation of such a case of de re knowledge-that. The main point to note for present purposes is just that whatever explanatory role experience of an object plays as concerns such de re knowledge-that, Campbell gives us no reason to think there is any conflict between its playing that role and the Representational View being correct.

Sometimes, however, Campbell seems to mean something different by ‘knowlege of the reference of a demonstrative’. He says, for instance, ‘Experience of a perceived object is what provides you with knowledge of the reference of a demonstrative referring to it’.Footnote ⁴⁹ Indeed the claim of experience of an object providing knowledge of reference is repeated often in the book. There is no clear sense in which experience of an object that is the referent of a demonstrative ‘provides one’ with knowledge that it is the reference of the demonstrative. Perhaps, then, Campbell has something different in mind from de re knowledge-that. There is a distinction between knowing which thing is demonstrated by a demonstrative and knowing the thing which is demonstrated by a demonstrative. The former is de re knowledge-that, but the latter is knowledge by acquaintance. Some of Campbell's remarks suggest he might have in mind knowledge by acquaintance. He says, for instance:

By knowledge of reference, he means here knowledge of the reference of a demonstrative. The suggestion I'm now floating is that he doesn't mean knowing which thing is demonstrated, but rather knowing the thing which is demonstrated. He may be claiming that experience of an object provides knowledge by acquaintance of the object. But there are serious reasons to doubt that. If the demonstrated object is, say, the 4,000^th cup produced in the year 2,000 in a certain factory in New Jersey, then to have knowledge by acquaintance of the object that is demonstrated is to have knowledge by acquaintance of the 4,000^th cup produced in the year 2,000 in a certain factory in New Jersey. One can have knowledge by acquaintance of the cup that is the 4,000^th cup produced in the year 2,000 in a certain factory in New Jersey, without having any idea that the cup is the 4,000^th cup produced in the year 2,000 in a certain factory in New Jersey, indeed without ever having heard of New Jersey. Similarly, one can have knowledge by acquaintance of an object that is the referent of a demonstrative without having any idea that it is the referent of a demonstrative; indeed one can have such knowledge by acquaintance without even having the concept of the referent of a demonstrative. Given that, I'm inclined to think that he doesn't mean knowledge by acquaintance of the thing that is the referent of a demonstrative. But, as I noted above, it is also the case that there is no clear sense to the claim that experience of an object provides de re knowledge-that. I confess that I don't really know what Campbell means by ‘knowledge of reference’. So, I won't discuss role 3 any further.Footnote ⁵¹

Let us turn now to explanatory role 4:

In his discussion of 4, Campbell notes that we can visually track an object over time. And he calls our attention to the following pattern of inference, where the demonstrative ‘that’ demonstrates a woman that one's conscious visual attention is focused on:

As concerns this pattern of inference, he tells us:

Again, the common factor view conception he has in mind is the Representationalist conception. Campbell makes here the same sort of mistake that I have repeatedly underscored. He mistakenly assumes that if a Representationalist is to explain how we can recognize the validity of such an inference, the explanation must be in terms of the representational content of visual experiences.

On the Representational View, it is indeed the case that two experiences can have the same representational content even if they are experiences of different women. One might be a visual experience of Judy, the other a visual experience of her identical twin sister, Trudy. Indeed a visual experience of Judy could have the same representational content as an hallucinatory experience. But although that is the case, it is not the case that on the Representational View your experience of the woman would have been exactly the same whether there was one woman there throughout, or many, or none. Suppose, again, that the woman that you are visually perceptually experiencing is Judy. Even on a common factor view, a visual experience of Judy is distinct from a visual experience of Trudy, and both are distinct from any hallucinatory experience. You would not visually experience Judy if the woman before your eyes was Trudy, not Judy. And were you completely hallucinating, there would be no object that you are visually experiencing. The key point to note is that we type experiences in many ways. We type them by their phenomenal characters. But we also type them by their objects. An experience of Judy is different from an experience of Trudy, even when the experiences have exactly the same phenomenal character. The reason is that they have different objects.

Of course, one cannot always tell whether one is experiencing Judy, rather than Trudy, or rather than visually hallucinating, and so not experiencing anything at all. But the Relationalist agrees that one cannot always tell. Moreover, the Relationalist and Representationalist can agree that you don't have to always be able to tell such things in order to recognize the validity of the pattern of demonstrative reasoning in the sort of perceptual circumstance that Campbell describes; for the skeptical scenarios can fail to be relevant alternatives.

Let us turn, then, to the question of whether the Representationalist can explain how, in a visual perceptual circumstance, we could recognize the validity of ‘That woman is running; that woman is jumping; that woman is running and jumping’. Representationalists can explain it as follows. We can recognize the validity of the pattern of demonstrative reasoning, because we can see that it continues to be the same woman involved in the activities. That, I take it, is what the Relationalist would say too. Indeed, the Representationalist and the Relationalist can offer the same explanation.

We would be able to continue to see that it is the same woman, in part, because we would be able to continue to visually attend to the woman through the interval in question. And we would be able to do so, because our visual system can track an object over time and change of location, as well as over other changes. As Campbell notes our visual system can so track objects as a result of the sub-personal level informational processes by which it solves various binding problems. The Representationalist can appeal to that same explanation. On the Representationalist View, an object is an object one sees in virtue of there being an appropriate causal process initiating with a state of the object and terminating with a visual experience with a representational content. Vision science studies the internal to the visual perceiver stage of such a causal process.

4. Two Well-Known Issues Facing the Relational View

The Representationalist holds that there is explanatory work for the phenomenal character of an experience to do that it won't do on the Relational View. The work concerns illusion and hallucination.

Campbell tells us, you will recall:

In cases of visual illusion, however, one is seeing an object, but the object looks some way that it isn't. The Representationalist maintains that this involves the visual experience representing the object as being some way that it isn't. There are both optical illusions, such as a straight stick's looking bent at the water-line, and psychological visual illusions, such as the Müller-Lyer arrows looking the same length. Despite the fact that he argues that the Relational View is superior to the Representational View, Campbell offers no account of how the Relational View can explain visual illusions of either sort.

Optical illusions such as a straight-stick's looking bent at the water-line, and psychological visual illusions such as the Müller-Lyer arrows looking the same length are common to all normal human visual perceivers. They are what Representationalists sometimes call ‘normal misperceptions’.Footnote ⁵⁶ Campbell does not discuss such cases of illusion. But he makes a few brief remarks about illusions resulting from idiosyncrasies of a perceiver. Immediately following the last quoted remark above, Campbell says:

He doesn't elaborate on the italicized remark. And there are only a few, brief, related remarks in the book. Campbell says at one point:

There could indeed be a case in which you see the scene before your eyes, but everything in the scene seems to have a yellowish cast, even though nothing in the scene actually has a yellowish cast. If so, you would be suffering an illusion, rather than hallucinating. As concerns this case of illusion, Campbell tells us it is ‘entirely compatible with the Relational View’.Footnote ⁵⁹ And he states:

But the point is that one may be seeing a scene, and yet the scene looks some way that it isn't. To accommodate this, it seems that we have to posit that the visual perceptual experience has a representational content. I can find in Campbell no alternative explanation of illusions involving idiosyncrasies of the perceiver.

Let's turn to hallucination. The Representationalist purports to be able to explain why a visual hallucination of a scene could seem to a subject just like a visual perception of a scene. The Representationalist maintains that a visual hallucination and a visual perception could have the same representational content, and so the same phenomenal character. It is standard for disjunctivists to point out that from the fact that a visual hallucination of a scene seems to a subject just like a visual perception of a scene, it does not follow that the subject is aware of something that is in common between them (other than their both seeming the same). Disjunctivists are indeed right that that does not follow. But the Representationalist maintains that the best explanation of why it seems that way to a subject is that the subject is aware of something that is in common, namely a representational content. The Relationalist, of course, rejects that explanation. But it is a standard challenge to the Relational View to offer a better explanation. Perhaps the Relationalist can do so. But no explanation is even ventured in Consciousness and Reference.

These well-known problems for the Relational View concerning illusion and hallucination are left unresolved in the book; indeed they go largely undiscussed. Given that Campbell tries to argue that the Relational View is superior to the Representational View, his failure to properly address these problems is a serious omission.Footnote ⁶¹