IN DEFENCE OF NAÏVE REALISM, Naïve realism
Updated: Aug 13, 2022
IN DEFENCE OF NAÏVE REALISM
2. Naïve realism
The position that I am advocating as naïve realism consists of the following three claims:
(i) Perception is not psychologically mediated.
(ii) Perceptual experience is presentational in nature.
(iii) The objects that we perceive have an existence independent of our own, or anyone else’s, perception of them.
The first is a metaphysical claim about the nature of perception. The second is a claim about that which determines the way in which one is appeared to when perceiving. The third claim represents the basic realist commitment. I am of the opinion that realism should only be rejected if no realist account of perception can be made to work. As the issue of whether naïve realism should be rejected or not is prior to the assessment of all other realist positions the realist commitment will operate as a background assumption and I will not consider it here. I shall explain both of the other claims, and how they are related, before going on to justify why I combine them, together with the realist commitment, under the title of ‘naïve realism.’
2.1 The metaphysical claim
A distinction can be made between two different senses in which we can think of perception as being direct or not. A subject’s perception of some object might be indirect in the sense that they perceive that object in virtue of perceiving something else. But another way that we might think of a subject’s perception of some object as being indirect is in the sense that we might think that the subject’s sensory experience, their conscious visual experience, is only a component of his or her perception of something. Following Foster (2000), I shall refer to these two different ways of thinking of directness within the context of perception as perceptual and psychological mediation, respectively.
It is worth mentioning Austin at this point, for he doubted the very cogency of the idea of ‘directness’ that is applied during philosophical discussion of perception:
It is quite plain that the philosophers’ use of ‘directly perceive,’ whatever it may be, is not the ordinary, or any familiar, use; for in that use it is not only false but simply absurd to say that such objects as pens or cigarettes are never directly perceived. But we are given no explanation or definition of this new use – on the contrary, it is glibly trotted out as if we were all quite familiar with it already. (Austin 1962: 19)
Austin’s worries are directed against the direct/indirect distinction that is put to use by the sense-datum theorists he is arguing against, and he is insensitive to the distinction between perceptual and psychological mediation. Nevertheless we can view him as laying down a challenge to make good any concept of mediation that is put forward in arguments concerning the nature of experience and perception, where it does not have its ‘ordinary’ or ‘everyday’ usage. Can we arrive at a fairly robust and agreed upon technical conception of what both perceptual and psychological mediation consists in?
2.1.1 Perceptual mediation
As already mentioned the notion of perceptual mediation, roughly speaking, is that a subject, S, perceives an object, x, perceptually indirectly iff S perceives x in virtue of perceiving some object, y, and x is not identical with y. The philosophical point of this notion is that there is a question as to whether our perception of normal objects in the world around us is mediated in this way by our perception of objects of some other kind. And this question is important because dependent upon our answer to it is our account of our epistemic position in relation to the world of normal objects around us that we take ourselves to be in. We are not trying to merely capture in more precise terminology what we ordinarily mean when we say that we perceive such and such a thing indirectly. We are trying to come up with a concept that has some significant philosophical work to do. So an important point to bear in mind is that the worth of this concept should not be judged by its success at satisfying our intuitions about whether or not such and such a case of perception is direct or not.
The technical definition as it stands is only as good as the notion of ‘in virtue of’ which it employs. What we are trying to capture is the thought that there is a significant asymmetry in the relation that holds between the perception of something indirectly in virtue of the perception of something else. When S sees x in virtue of seeing y, S’s seeing x depends upon his seeing y, but not vice versa. In order to understand this dependency we must understand what grounds it. What is it about the perceptual situation that makes it the case that the two perceivings, the perceiving of x and the perceiving of y, are related in this way? In what way does the perception of x depend upon the perception of y? It is important to note that to talk about the two perceivings, of x and y, is not necessarily to talk of two acts of perception. There could be one act of perception in which two things are perceived, one in virtue of the other.
It seems to me that the best way to understand this asymmetry is in terms of the contribution that the different things seen make to the phenomenal character of the experience enjoyed by the subject in seeing them. The phenomenal character of an experience is the way in which a subject is appeared to. The idea, then, would be that something is immediately seen only if it fixes, or determines, the way in which a subject is appeared to. Something is indirectly seen when it does not play this role, but is related in the right kind of way to something that does play this role. So when S perceives x in virtue of perceiving y, y fixes how things appear to S, but x does not:
The immediate objects of sight will then be those that not only look some way, but which, in so looking, fix the way in which all objects we perceive look to us to be. Other objects will count as being seen (and hence as being seen mediately) through their relations to the objects immediately seen. It is through the immediate object of perception, and the way it looks, that these mediate objects come to look some way to us. (Martin 2005: 707) 
To illustrate this position, consider two examples in which, according to this conception of perceptual mediation, one thing is seen in virtue of seeing another thing.
The first kind of example is perceiving something by perceiving it on television. John sees Ronnie O’Sullivan in the process of winning the snooker world championship on television and he perceives O’Sullivan indirectly, because he perceives him in virtue of perceiving images on the television screen. And we will explain this claim in terms of the fact that it is the images on the screen that fix how John is appeared to, rather than O’Sullivan. John could, unbeknownst to himself, be watching a television screen whose images are generated, not by O’Sullivan, but by a computer program that has modelled the appearance of O’Sullivan perfectly. In both cases, whether watching the live televised event or the computer generated simulacrum, it is the images on the screen that fix how John is appeared to. In the case of watching the live televised event, John counts as having seen O’Sullivan in virtue of the relationship between the image on the screen and O’Sullivan himself, which is, quite plausibly, that of causal dependency of the former upon the latter.
The second example of indirect perception is that of perceiving a three dimensional object in virtue of perceiving its facing surface. When Sally sees an apple, she will see the apple in virtue of seeing its facing surface. This is because it is the facing surface of the apple, rather than the whole apple, that fixes the character of how Sally is appeared to. Were the rest of the apple to be removed, leaving only the facing surface, the character of Sally’s experience would remain the same. It is because the facing surface is a part of the apple that the apple is seen. So in this case the relation between facing surface and apple is quite different to that between image on the television screen and person. So there is not just one relation that must obtain between two things in order for us to be able to see one in virtue of seeing the other. Whether a relation can do the job of linking two things together in such a way that one counts as being perceived in virtue of perceiving another is probably best determined on a case by case basis.
The philosophically interesting question that this technical notion of perceptual mediation allows us to ask is whether our perceptions of normal things in the world are always mediately perceived through the perception of non-normal things.
Consider the television example. It is of no real philosophical interest that there are cases such as this in which one thing, Ronnie O’Sullivan, is perceived in virtue of perceiving another thing, the image on the television screen. What is of interest is whether, in John’s seeing of O’Sullivan, the most immediate thing that John perceives is a normal object in the world external to John, that exists independently of his perception of it, and is something that he would pre-reflectively take to be a normal object of sight. According to what we can refer to as the perceptually indirect realist, the most immediate thing that John perceives is not a normal object, but something quite unlike that which John would pre-reflectively take to be the objects of sight.
We can say, following Foster, that a subject S f terminally perceives a normal object, x, if and only if S perceives x and there is no other normal object, y, such that S’s perceiving of x is perceptually mediated by their perceiving of y. What is philosophically interesting, then, is whether f terminal perceiving is mediated by the perception of some non-normal object.
At this point we must pause to make mention of Austin again. Even if we accept that we have elucidated a clear technical notion of perceptual mediation, we can worry about whether the philosophical use to which we are trying to put it makes good sense. In particular, can we clearly articulate what is meant by something being a non-normal object? One of Austin’s criticisms of the arguments put forward by sense-datum theorists was that this notion of the most immediate object of perception being non-normal, or non-physical, or in some way not a part of the external world, is unsatisfactorily obscure.
There are two strands of thought to Austin’s criticism. The first is that an illegitimate distinction is being made by advocates of indirect theories of perception. They argue that we must accept that objects and properties that are not real are the immediate objects of perception. And so a distinction is made between the ‘real’ and the ‘unreal’ that, in fact, we are not licensed to make because, as Austin puts it, the word real is ‘substantive hungry’ (Austin 1962: 68-70). Whenever we say that something is real, or ask of something whether it is real, then we can always ask the question ‘a real what?’ But, when philosophers use the word ‘real’ in connection with perception, it is clear that there is no such question in the offing, and consequently, that the word as they use it has no real meaning.
The second, related thought is that advocates of the direct/indirect distinction try to understand what they take to be perceived directly in terms of what they take to be perceived indirectly and also try to understand what they take to be perceived indirectly in terms of what they take to be perceived directly. But no genuine understanding of what is being talked about can be achieved this way. Once again, or so the accusation goes, language is being used that is inherently mysterious, and that we cannot make good sense of.
Both of these objections to the direct/indirect distinction (as it is construed in its philosophically interesting form) are typically met by the claim that (a) the question over the nature of the immediate object of perception is a question about whether the object is mind-dependent or not, and so (b), what is meant by ‘real’ in this context is just ‘mind-independent,’ and considerations about illusions and the rest are supposed to encourage us to think that the objects and properties we are acquainted with in experience are not real in this sense. ‘Real,’ in this sense, is not substantive hungry at all.
While I accept that this is part of the story about what is at issue in discussion of direct versus indirect perception, I do not believe that it is the whole story. In particular this account of what it is to be a normal object excludes a possible conception of the nature of non-perceptual experience. Recall that non-perceptual experience is sensory experience that does not put the subject of such experience into cognitive contact with the normal objects of perception. That is to say, it does not put the subject of such experience in a position to think about and act towards normal objects like chairs, tables, mountains and the like. If we think of the normal objects of perception purely in terms of their being mind-independent, then this rules out the possibility that non-perceptual experience could involve mind-independent objects to which we are related. It rules out the possibility that there could be non-perceptual experience that puts us in cognitive contact with objects whose existence is independent of such experience. It excludes the possibility that there might be experience of objects that are not ‘real’ but which nevertheless have an existence that is independent of one’s awareness of them.
The sort of possibility that I am envisaging here is one in which a subject is plugged into some giant artificial reality device. I want to leave open the possibility that we can understand such a situation as genuinely involving objects that the subject of the experience comes to be aware of. This seems to be at least intuitively plausible. In such a case the computer in charge of the artificial reality creates a virtual environment, populated with virtual objects that one can become aware of and interact with through one’s virtual body. These objects would certainly seem to have some claim to be awareness-independent. They are dependent for their existence, not upon the subject’s awareness of them, but upon the operations of the computer. It is possible for the same object to be seen by more than one subject, whether at once or at different times. It is possible for one subject to see the same object at different times, and there is a distinction between how an object really is and how it appears to be.
In such a scenario a subject would be enjoying non-perceptual experience, but his sensory experience would put him into cognitive contact with objects whose existence transcends his awareness of them. The objects that the subject could become aware of are independent of their awareness of them, but are in some sense unreal, or non-normal.
I propose, then, to add to this conception of what we take ourselves to see, not only mind-independence (in the sense of being independent of our acts of awareness) but also the idea that they are in some sense fundamental. By this I mean that we take the things that we f terminally perceive to not depend for their existence upon anything other than themselves or their parts. This would rule out the objects that the subject in the virtual world is aware of as being normal, because they depend for their existence upon the computer, and the computer is not a part of those objects. Now of course, there are some things that we might claim to see that are not fundamental in this way, such as holes or shadows. But, we might say that we do not really see holes or shadows, but only the things whose existence the holes and shadows depend upon, or, that if we do see such things it is only in virtue of seeing things that are fundamental. That is, they are not f terminal objects of perception
The idea would be then, that the philosophically interesting question of perceptual mediation is as to whether we f terminally perceive normal objects in virtue of f terminally perceiving non-normal objects, where a normal object is an object whose existence is independent of the subject’s perception of it and whose existence is fundamental.
We are close, then, to coming to a clear understanding of the technical notion of perceptual mediation. There is, however, one further worry about it that I wish to clear up. This is that there is something objectionable about the thought that we perceive three dimensional objects in virtue of perceiving their facing surfaces, and so any account of perceptual mediation which leads to this conclusion must be rejected.
What is objectionable is that the technical definition of mediation seems to lead to the conclusion that the most immediate objects of perception could not be three-dimensional objects. At best they could only be the facing surfaces of such objects. And so, by mere definition, we have rendered the objective, three-dimensional world, beyond the scope of perceptual contact. And, in doing so, we have made it problematical as to how we have knowledge about such a world.
What is the source of our resistance to this idea? One explanation is that we are illegitimately incorporating aspects of the informal notion of perceptual directness into our assessment of the adequacy of the formal notion.
Normally, when we talk about seeing one thing indirectly by seeing another thing, the two things are entirely distinct. And we bring this intuition with us into our attempts to elucidate a philosophically interesting formal concept of perceptual mediation. And so we balk at the claim that our perception of things is mediated by our perception of their facing surfaces, because this makes it sound like we are creating two things that are seen, where we know there is only one thing to be seen. But, in fact, if we attend to the formal notion we can see that we are doing nothing of the sort. The uncomfortable feeling that we are doing something like this is the unfortunate consequence of our using expressions that have a common, everyday meaning. We are forced, by our technical notion, to say things like ‘we never immediately perceive three-dimensional objects,’ which, given that the reasons for supposing this are meant to be prior to a commitment to indirect theories of perception, as philosophy has traditionally conceived of them, sounds rather shocking. But it is not. We can rephrase it as ‘we never perceive three-dimensional objects except by, or in virtue of, perceiving parts of those objects,’ and this does not sound nearly so shocking when we bear in mind what we mean by ‘in virtue of’ and also ‘parts of those objects.’ It is the use, I think, of the word ‘immediately’ that puts us on our guard. It is not as if there is one thing, the object, and then there is another thing, a part of the object, that in some sense ‘gets in the way’ of the object.
Perception of something is immediate if it does not obtain in virtue of the perceiving of something else, just as the notion of perceptual mediation tells us. But in addition to this, we can consider whether our perception of something is such that it is available for immediate demonstrative identification. A lot of the time when we talk about perceiving something indirectly, we are talking about cases in which we are in a position to only demonstratively refer to that thing through being in a position to non-dependently demonstratively refer to what we directly perceive. So a sonar operator, we might say, can indirectly see the submarine by seeing the blip on his radar screen, and, what is more, he can demonstratively refer to the submarine, but he can only do so by demonstratively referring to the blip on his screen. But in the case of indirectly seeing some object in virtue of seeing its facing surface, what we are in a position to non-dependently demonstratively refer to is the object that is indirectly seen. At least, that is the intuition before considerations about illusions and the like set in. But it is important to note that this intuition is not challenged simply by our technical notion of perceptual mediation. We have not somehow put objects further away from us, epistemically speaking, by a mere technical definition. But I think it is precisely that worry that motivates suspicion of this notion of perceptual mediation.
To say that that S perceives x indirectly in virtue of perceiving y does not have any epistemic implications. The epistemic significance of this fact of perceptual mediation will lie in the nature of the objects perceived directly and indirectly and their relation to one another.
2.1.2 Psychological mediation
Having dealt with the notion of perceptual mediation we can now turn to that of psychological mediation. It is far easier to provide a straightforward account of this concept, quite possibly because it has not attracted as much attention in the literature as perceptual mediation.
When someone sees something then that thing is available for them in thought and action. They can direct their thoughts at that object in entertaining, for example, beliefs about that object. They can also direct their actions towards that thing in, for example, walking around that object or picking it up. Whatever state of affairs that some subject’s seeing of some object consists in, it is what explains and makes possible this object-directedness of their thought and behaviour.
As well of the notion of seeing, we can also distinguish the notion of visual sensory experience. Experience is the conscious subjective episode that we enjoy when we see something. It is what we are referring to when we talk about the way in which things appear to us, as old or shiny, blue or sad, flat or voluminous.
The claim that perception is not psychologically mediated is a claim about the relation between perception and the experience one enjoys when perceiving (perceptual experience). It is the claim that perceptual experience does not fall short of perception. It is, in itself, that which puts us into perceptual contact with the world.
John Foster defines the notion of psychological mediation in the following way:
S’s perceiving of x at t is psychologically mediated by his being in Σ if and only if
(1) Σ is a psychological state;
(2) Σ is not, in itself, x-perceptive (i.e. being in Σ does not, on its own, logically suffice to put one in perceptual contact with x);
(3) S’s perceiving of x at t breaks down into his being in Σ at t and certain additional facts;
(4) These additional facts do not involve anything further about S’s psychological condition at t (anything over and above what is already covered by S’s being in Σ). In other words, in combining with the fact of S’s being in Σ, they do not add any further psychological facts, about S at t, to the constitutive base. (Foster 2000: 10)
Belief that perception is psychologically mediated in this way seems to be an expression of the belief that perception is essentially conjunctive in nature. The conjunctivist maintains that the psychological episode that is the subject’s conscious experience features as an element in perception. What does it mean for experience to feature as an element in perception? The having of experience, in conjunction with the obtaining of certain other conditions (such as a particular sort of causal relation between said experience and the world) is what perception of the world consists in. Here, for example, is Paul Coates’ understanding of perception, which is a typical conjunctive picture:
At a first approximation, a subject S sees a physical object X, if and only if:
(1) X exists
(2) S has an inner visual experience E, comprising:
(a) a visual phenomenal state
(b) a perceptual taking, an episode that involves the exercise of concepts
(3) The object X causes E, in the appropriate manner for seeing.
(Coates 2007: 57)
And we can find similar sorts of account in Vision (1997), Grice (1988) and Strawson (1988). The key feature of these accounts is not the particular way in which they break down the perceptual episode, but that the episode is broken down into a visual experience, which is not in itself perceptive of the world, plus a set of other factors. I say ‘in itself’ because there is a sense in which, on this kind of view, the experiential component of perception is the perception of something. But it is the perception of something not because of its intrinsic nature. Its counting as a perception is a relational property of the experience to be explained by some suitable account of the relation between the experience and the thing perceived. And this is commonly, as in the case of Coates’ view above, to be understood as a causal relation between experience and object.
2.1.3 Naïvety about mediation
We can, then, make a distinction between S’s perception of x being perceptually mediated, and its being psychologically mediated. We can say that S’s perception of x is perceptually mediated if S’s perception is constituted by, or obtains in virtue of, their perception of some non-normal object. We can say that S’s perception of x is psychologically mediated if their perception has, as a component, a psychological state of affairs which fully exhausts the psychological aspect of S’s perception of x but is not in itself perceptive of x.
This recognition of there being two senses in which we can speak of perception as being indirect gives us four different possibilities as to the metaphysical structure of normal object perception:
(i) Perception of normal objects is neither psychologically mediated nor perceptually mediated.
(ii) Perception of normal objects is psychologically mediated, but not perceptually mediated.
(iii) Perception of normal objects is perceptually mediated, but not psychologically mediated.
(iv) Perception of normal objects is psychologically mediated and perceptually mediated.
Position (i) represents the naïve realist conception of the structure of perception. To advocate position (ii) is to adopt some form of what is usually called direct realism but, on this approach, is best thought of as perceptually direct and psychologically indirect realism. f terminal perceiving is not perceptually mediated, but it does break down into an experiential component plus other factors. Because perception is not perceptually mediated on this account, the experiential component of perception is not, in itself, the awareness or perception of anything. Adverbial accounts and (most) intentional accounts of experience are versions of this.
While position (iii) is a logical possibility it seems quite implausible. If f terminal perception is mediated, then we only ever perceive normal objects by perceiving non-normal objects. But if perception is not psychologically mediated, then visual experience is not a component of perception. So, we perceive normal objects by perceiving non-normal objects, and this perception of non-normal objects does not constitute visual experience, for if it did, it would have to play a psychologically mediating role. But now it is starting to look as if visual experience and perception are coming apart. We want to say that to consciously perceive something is to have an experience of that thing, but this present account makes it entirely mysterious as to how they are connected. The implausibility of this picture allows us to conclude that if perception is perceptually mediated, then it must also be psychologically mediated
Position (iv) is that of what has been traditionally referred to as indirect realism, but under this taxonomy must be referred to as perceptually and psychologically indirect realism. This is to maintain that perceptual experience is in itself the awareness or perception of some object, but this is not a normal object. It is traditionally associated with sense-datum accounts of experience. f terminal perceiving is perceptually mediated, because it breaks down into awareness of some non-normal object, and it is psychologically mediated because it is this awareness plus other non-perceptual factors that is the experiential component of perception.
2.2 The presentational claim
The presentational claim represents a particular way of understanding the psychological immediacy of perception. It is the claim that the objects we perceive in an episode of perception constitutively contribute to the way in which those objects appear to us in such an episode. I shall understand this as the claim that perceptual experience is presentational in nature, in that something is literally present in experience and, furthermore, that this thing is the object of perception. To say that something is literally present in experience is to take a stance about the nature of the object of experience when one enjoys a perceptual experience of some object.
What is the ‘object of experience’? As well as a subject being appeared to in a certain way in perceptual experience, there is also that which is present in experience. When we enjoy a sensory experience something appears to us a certain way. This ‘something’ is what I shall refer to, following J. J. Valberg (1992), as the ‘object of experience.’ He describes it as follows:
By an ‘object of experience’ we shall mean something present in experience: something which is right there, available for us to pick out or focus on, and refer to demonstratively. (Valberg 1992: 7)
The object of experience is something that is present in experience in the sense that it is available to the subject of the experience to focus upon and demonstratively refer to. This expression should not be thought of as meaning that there is some object or entity that appears to us. To say that what appears to us is in an experience is an entity of some kind is to express a substantial commitment as to the nature of that experience. But that there are objects of experience in the sense that something appears to us when we consciously perceive is something that seems apparent. For we can understand by this merely that there is an answer to the question ‘what appears to you in your experience?’
It seems highly plausible that there are objects of experience in this sense. When one consciously perceives some normal object and it appears to one a certain way there is clearly something present to one, right there, that one can focus upon and which appears to one as being a certain way.
The presentational claim is a particular view as to the nature of this object of experience. It takes this object of experience to be, firstly, something actually present in experience, and secondly, to be the object of perception, that is, that which is perceived in virtue of the subject enjoying the experience. To say that the object of experience is something literally present in experience is to say that it is some entity that is present to one in an ‘ontologically immediate way,’ as a constituent of the experiential episode. In claiming that the object of experience in perceptual experience is the object of perception, the presentational nature of perceptual experience is taken to be such that the experience, in itself, provides for perceptual contact with the world. In this way the presentational claim embodies a particular conception of the way in which perception can be psychologically direct.
There will be various ways of fleshing out the presentational claim. A distinction can be made between two differing attitudes to understanding the role of the objects of perception in determining phenomenal character. It seems to me that on the one hand there is a popular way of understanding this according to which the ways in which we are appeared to in perceptual experience are the very ways that the objects we perceive are. There is an identity between the phenomenal properties of experience and the properties of the objects we perceive.
On the other hand, we could try and understand the determining role of the objects of perception in some other way. Whatever way this is, it would have to be such that the objects of perception play a constitutively determining role in the determination of phenomenal character. At the moment we should just acknowledge that the presentational claim that the naïve realist makes is merely that the objects of perception play a constitutive role in the determination of phenomenal character. That is, they are intrinsic elements of perceptual experience that determine the phenomenal character of such experience. It is not a part of this claim that they fulfil this role in virtue of an identity relationship between the qualities instantiated in experience and the qualities instantiated by those objects. This is just one way of working out the details of how the presentational claim may be true. Chapter five is devoted to laying out an alternative to this picture on behalf of the naïve realist.
2.3 Why is this ‘naïve’?
What are we intending to convey by describing someone’s conception of perception as ‘naïve realist’? The usual sense of this expression is to denote the pre-theoretical nature of one’s approach to perception. The naïve explanation is something the ‘common man’ would believe, who has not yet reflected in any great depth upon the position that he assents to. It is naïve because it is not born out of considered reflection.
Such usage of the term ‘naïve’ is quite often intended pejoratively, in which case to say that a position is naïve is to say that it is mistakenly simple, or mistakenly unreflective. Alternatively, we need not consider it to be something negative that an opinion is ‘common,’ ‘simple’ or ‘unreflective.’ Rather, we might use ‘naïve’ to indicate of some explanation that it is where enquiry begins, and as such not something that need necessarily be incorrect. This is how I intend to use the term ‘naïve’ – as expressing the beginning of reflective enquiry.
When we speak of someone as being a naïve realist, we could mean to imply that there is something naïve about the realism that they adopt. They are ‘naïve’ in so much as they are a particular kind of realist. Alternatively, we could think of ‘naïve,’ not as something that offers a description of the kind of realism that a theorist employs, but as something separate and additional to their realist commitments
To be a realist within the domain of perception is to believe that the objects of perception have an existence independent of their being perceived. It is a claim about the nature of that which we perceive. Naïvety in regard to this claim might be construed as being naïve about the extent to which the objects of perception, and the properties that they appear to have, exist independently of our perception of them. This would be to think of ‘naïve’ as describing the kind of realism that someone adopts with regards to the objects of perception.
This is certainly one of the ways in which ‘naïve realism’ has been used. Jonathan Dancy, for example, uses ‘naïve realism’ in this sense when he writes:
The naïve direct realist holds that unperceived objects are able to retain properties of all the types we perceive them as having. By this he means that an unperceived object may still not only have a shape and size but also be hot or cold, have a colour, a taste and a smell, be rough or smooth and make a noise or keep silent. The naivete of this position lies in the word ‘all.’ The position becomes less naïve as ‘all’ retreats to ‘nearly all’ and then to ‘most’ and so on. (Dancy 1985: 147)
The other sense in which ‘naïve realist’ is used is that which attributes an opinion about something in addition to the scope of one’s realism about the objects of perception. On this reading, naïvety applies, not to our conception of the nature of the objects of perception, but to our conception of the nature of the subjective episodes that we enjoy when we perceive.
An illustrative example of this is a recent presentation of the naïve realist position by A. D. Smith:
What […] is essential to Naïve Realism as I am construing it is the claim that that which gives sensory character to perceptual consciousness is a public quality of some physical object. (2002: 43-44)
For Smith, then, naïvety is expressed in how one tries to explain the sensory character of perceptual consciousness. One is naïve if one thinks that some public quality ‘gives’ perceptual consciousness its sensory character, as opposed to, say, the qualities of an inner object of awareness. It is this sense of naïvety that is expressed in the phenomenal claim that I regard as one of the commitments of the position I refer to as ‘naïve realism.’ But note that, as mentioned in the previous section, the sense in which the object perceived ‘gives’ experience its phenomenal character is not yet explained if we only restrict ourselves to the presentational claim. And the naïve realist should restrict themselves in this way, as it is too early in the stage of enquiry to commit oneself to anything stronger about the nature of this presentation – that perceptual objects explain, or give, phenomenal character to experience through an identity between the qualities of such character and the qualities of the objects.
When characterising naïve realism in terms of naïvety with respect to the nature of perceptual experience, it is normally expressed in terms of a commitment to the relation between experience and the mind-independent world. For example:
Visual and tactual sense-data are parts of the surfaces of physical objects. (Price 1950: 55)
If an experience E is a genuine perception by a subject S of object O then the occurrence of E places S in such a relation to O that were S able to entertain demonstrative thoughts (and was equipped with the necessary concepts) then S could entertain the true demonstrative thought ‘that is O.’ (Snowdon 2005: 138)
But the kinds of commitments illustrated here as to the relation between experience and the mind-independent world are consequent to some serious thought about the nature of experience. Really, what these philosophers are saying is that naïve realism is committed to a certain claim about the relation of experience to the world, given certain facts about the nature of experience that have been argued for. Let us take each example in turn.
To the sense-datum philosophers of the early twentieth century, such as Price, naïve realism was a certain position with regards to the nature of that which is immediately given in sense experience: sense-data. Given that there are sense-data, to be naïve about them is to identify them with external things, or, to be more specific, surfaces of physical objects. Incorporated into the conception of naïve realism is the notion of sense-data. Once we accept that experience is a relation to something that is ‘given’ in experience, to be naïve about experience is to identify what is given with tracts of the physical world. But to accept the basic claim about experience one must already have travelled some distance down the road of enquiry.
For Snowdon, naïve realism is to be characterised in terms of our capacity to entertain demonstrative thoughts about objects in the world. If we think about experience from this perspective, then naïvety is expressed through taking immediate non-dependent reference to objects of experience to be referring to physical objects in the world.
What philosophers like Price and Snowdon are offering us are ways of elaborating any pre-theoretical stance we have towards perception and experience in terms of what reflection upon perception and experience reveals to us. All these characterisations are reconstructions of what the naïve realist is committed to in terms of a more worked out theoretical vocabulary. There is nothing incorrect about any of this, and indeed, it is only proper that naïve realism should be articulated in a way that fits in with how a philosopher talking about it views the philosophical landscape. My worry is rather that the heart of the naïvety is being overlooked. The claims given above are claims that the naïve realist should accept because of his naïvety, and so leave the essence of his pre-reflective intuition untouched.
In my opinion naïve realism should be expressed as a commitment to the relationship between perceiving and experiencing and the presentational claim, rather than the relationship between experience and mind-independent objects. It is the former that makes the position naïve, while it is the latter that makes it realist. It has a naïve view of the relation between perceptual contact and sensory experience because it takes sensory experience to be, at least sometimes, itself perceptual. Sometimes, when a subject is having an experience, the experience simply is their perceiving. It has a naïve view of the nature of sensory experience because it takes it to be presentational. The realist thought is that the objects of perception do not depend upon the subject for their existence
Naïve realism should, then, be considered a combination of two thoughts – naïvety with regards to the relation between perception and experience, and the presentational nature of sensory experience, and realism about the objects of perception.
Naïve realism, as I understand it, is the claim that when a subject, S, perceives a normal object, x, the sensory experience in virtue of which S stands in this relation to x is, in its nature, such that this perceptual contact is explained. There is, on the naïve realist picture, nothing more to be added to such an account in order to provide an account of perception.
The purpose of this discussion is not to get to the bottom of what we actually, or should, mean by ‘naïve realism.’ It is rather to explain what I mean by ‘naïve realism,’ and to offer some justification for according it this title. I call it ‘naïve’ because I think that it is where we start – it is the intuitively correct answer to the first question that occurs to us when considering the nature of perception, namely, ‘what is the relation between experience and perception?’
3. Varieties of non-naïve realism
If one rejects the idea that perceptual experience is presentational of the normal objects of perception, then one must adopt some form of non-naïve realism. Forms of non-naïve realism can be distinguished according to their differing attitudes towards psychological and perceptual mediation. There can, then, be psychologically direct non-naïve realism, psychologically indirect and perceptually direct non-naïve realism, and psychologically indirect and perceptually indirect non-naïve realism. There is only one non-naïve form of psychologically direct realism because, as it may be recalled, conceptions of perception according to which it is both psychologically direct and perceptually indirect are not plausible.
3.1 Psychologically indirect and perceptually direct realism
This form of realism maintains that perceptual experience has a nature such that in itself it falls short of being perceptual of the world. It does, however, claim that we do not f terminally perceive things in the world in virtue of perceiving anything else. There are two main varieties of this approach, intentionalism and adverbialism.
The intentionalism that falls under this category is a view of the nature of perceptual experience according to which perceptual experience is an intentional state, but in which the content of this state, that specifies the way that the world is, is independent of the object of perception. The core phenomenal character (and indeed, usually the whole phenomenal character) of experience is to be explained in terms of the representational or intentional content of this state.
To understand phenomenal appearances as intentional in nature is, typically, to understand them as representing the world as being a certain way. Perceptual experience will be a kind of intentional state, where this is thought of as a state of affairs in which a subject is related to some content that specifies how the world must be in order for the experience to be veridical. As Siegel puts it, the sensory experience will have ‘accuracy conditions’ (Siegel 2006: 485). This content is typically thought of as a proposition to the effect that the world is such and such a way. What will make the intentional state a visual experience will be the peculiar way in which this content is entertained, or in the special nature of the attitude that the subject takes towards this content. Intentionalism, or representationalism, is the claim that the fundamental nature of perceptual experience, the core phenomenal character of a subject’s experience when they perceive something, is determined by a representational content.  Here, for example, is Dretske:
The way things phenomenally seem to be (when, for instance, one sees or hallucinates an orange pumpkin), are – all of them – properties the experience represents things as having. (Dretske 2003: 67)
Now, according to the intentionalist account the object of experience can still be thought of as a normal object, just like for the naïve realist, but for the intentionalist this normal object is not actually present in experience – it is present as an intentional object of the experience. So we can think of the intentionalist as maintaining that the object of experience is an intentional object. In this way the intentionalist can claim that perception is psychologically direct, for they can claim that, when a subject perceives, the object of experience is the object of perception. But, because they understand the object of experience intentionally this does not mean, as it does for the naïve realist, that it is literally present to the mind of the subject. The perceptual experience is not itself a relation between subject and object perceived. Rather, the object perceived is represented to the subject.
Adverbialism, on the other hand, is a picture of phenomenal appearance according to which it is mode of the perceiving subject. To be appeared to in a certain phenomenal manner is to have one’s consciousness modified in a particular way. It therefore eschews a picture of phenomenal appearance in which there is something that appears to the subject. Phenomenal appearance does not, in itself, consist of anything appearing to a subject. Naive realism, intentionalism of any form and, as we shall see, the sense-datum theory all conceive of phenomenal appearance in terms of something appearing to a subject. In the case of naïve realism and the sense-datum theory this object of experience is an actual object that the subject stands in a genuine relation to, while in the case of intentionalism, this object of experience is an intentional object. It rejects the claim that phenomenal appearance is presentational in the sense explained earlier.
3.2 Psychologically direct non-naïve realism
This account of the nature of perceptual experience and the structure of perception agrees with the naïve realist that perception is not composed of sensory experience plus a set of other factors. That is to say, perceptual experience is, in its very nature as experience, perceptual of the world around us. Where this account differs from the naïve realist’s is as to its conception of the nature of such experience. It does not believe that such experience is, in its nature, a presentation of the object of perception. It does not believe that phenomenal appearances are to be understood in such a way.
Generally, the way in which this account is fleshed out is in terms of the nature of phenomenal appearances being understood in intentional terms. Someone could adopt an intentionalist understanding of the nature of phenomenal appearances that gives a psychologically direct view of perception by taking the content of the mental state that is the phenomenal appearance to be object-dependent. The way in which the world appears to be, the way the world is represented as being, includes the object of perception as a part.
The intentional content of such an experience would be something like ‘that is red’ or ‘this is square,’ demonstrative propositions whose content is object-dependent.
On this account experience is a relation to the object of perception only in so much as it consists in the subject being related in a certain way to a certain content that contains that object as a constituent. The object of experience, that which is present in experience, is the object of perception, but it is present intentionally. When a subject enjoys a perceptual experience the object of perception is not present to the subject in an ontologically immediate way, but rather present in the content that represents the world as being a certain way. So experience is in itself a relation to the object of perception, is in itself perceptual, as the naïve realist maintains, but this intrinsic relationality is achieved via a representational content.
3.3 Psychologically indirect and perceptually indirect realism
This position is occupied by what I shall refer to as sense-datum theories of perception. It is in agreement with naïve realism to the extent that it conceives of phenomenal appearances as genuinely presentational in nature. The object of experience when one perceives some object is a genuine entity that is a constituent of the experience in an ontologically immediate way. That is to say, it is not just represented in experience. But unlike naïve realism it does not take the object of experience to be the object of perception. Instead it is a mind-dependent object, or sense-datum, which is typically taken to determine the phenomenal character of perceptual experience by instantiating the properties that characterise the way in which a subject is appeared to when enjoying such an experience.
In conceiving of the object of experience in perceptual experience as a mind-dependent entity the sense-datum theory takes perception to be both perceptually and psychologically mediated. f terminal perception is mediated by awareness of a sense-datum, and the perceptual experience is not in itself perceptual as it has its fundamental constituted by the awareness of this entity.
 As will become evident, much of what follows and in particular the metaphysical categories that are open to the perceptual theorist (who is a realist) are owed to Foster’s conception of the philosophical landscape in his The Nature of Perception (2000).  See Foster (2000).  Bermúdez (2000) advances a similar idea.  A minimal constraint is plausibly that the thing indirectly seen is available for the subject in thought and action.  See Foster (2000: 6).  It should be mentioned that some indirect realists do not like referring to the relation that obtains between subject and non-normal object as ‘perceiving.’ They are worried that if they conceive of the relation as being the same relation that obtains between subject and normal object when perceiving, then an infinite regress will be generated. For perceiving a normal object would imply perceiving a non-normal object, but then perceiving a non-normal object would imply perceiving another non-normal object, and so on. This fear is unfounded. To perceive a normal object is to perceive a non-normal object appropriately related to it, and this no more generates a regress than the fact that to touch something one must touch a part of that thing. See Harrison (1993).  See Locke (1967).  We could adopt the strategy of saying that in this situation the subject does perceive normal objects in virtue of the sensory experience that he enjoys. From his perspective, the objects of his sensory experience are the real, normal objects that exist in the world. To this I can only say that were such a subject to be removed from the matrix/influence of the evil demon he would surely come to the conclusion that what he was now perceiving were normal objects and that what he used to perceive were non-normal objects.  But some people do seem to think this. See, for example, Broad (1929: 149-151).  Most conjunctive accounts are causal theories of perception, in that they take perception to break down into experience plus a causal relation of the right kind to the object perceived.  McDowell (1982; 1986) is arguably an intentionalist whose view of perception is such that it is neither psychologically nor perceptually mediated.  See Foster (2000: 50).  This view seems prevalent amongst both critics of naïve realism and its proponents. See, for example, Smith (2002) and Brewer (2008).  An intentionalist does not have to think that it is the world that is represented in phenomenal appearance. They could (but invariably do not) maintain that non-normal objects of some kind are represented in this way.  Proponents of this view include Tye (2002), Dretske (2003) and Harman (1999). It can come in a strong or a weak variety according to whether it is just the claim that phenomenal character supervenes upon intentional content or that the very nature of such character is to be understood in terms of this content.  For the adverbialist position see Ducasse (1942; 1951).  This is only one way that someone could be a non-naïve realist and maintain that perception is nevertheless psychologically direct, and so there is, of course, space for someone to adopt a different account that falls into the same category. But it is, as far as I can tell, the only way that this has been attempted.  For articulations of this kind of view see Foster (2000), Campbell (2002) and Brewer (2006).  It should be noted that while the term ‘sense-datum’ is generally taken to have the meaning given in the text within contemporary philosophy, it once had a usage in which it was a term that was neutral between naïve and non-naïve theories of perception, and was taken to refer to whatever entity was the object of experience. In the course of my discussion I shall maintain the current usage of the term as referring to a mind-dependent object of experience. For a classic sense-datum account see Price (1950). For an example of the more modern approach see Jackson (1977).