The Value of Perception

This paper develops a form of transcendental na € ı ve realism . According to na € ı ve realism, veridical perceptual experiences are essentially relational. According to transcendental na € ı ve realism, the na € ı ve realist theory of perception is not just one theory of perception amongst others, to be established as an inference to the best explanation and assessed on the basis of a cost-bene ﬁ t analysis that weighs performance along a number of different dimensions: for instance, ﬁ delity to appearances, simplicity, systematicity, ﬁ t with scienti ﬁ c theories, and so on. Rather, na € ı ve realism enjoys a special status in debates in the philosophy of perception because it represents part of the transcendental project of explaining how it is possible that perceptual experi- ence has the distinctive characteristics it does. One of the potentially most interesting prospects of adopting a transcendental attitude towards na € ı ve realism is that it promises to make the na € ı ve realist theory of perception, in some sense, immune to falsi ﬁ cation. This paper develops a modest form of transcendental na € ı ve realism modelled loosely on the account of the reactive attitudes provided by Strawson in ‘ Freedom and Resentment ’ , and suggests one way of understanding the claim that na € ı ve realism is immune to falsi ﬁ cation. complicated web of attitudes and feelings which form an essential part of moral life as we know it, and which are quite opposed to objectivity of attitude.


Transcendental Na€ ıve Realism
According to the version of the na€ ıve realist theory of perception that I will take as representative, veridical perceptual experiences are essentially relational: veridical perceptual experiences are constituted at least in part by the mind-independent objects and properties in our environment that they are experiences of. Because on this view veridical perceptual experiences are essentially relational, a particular perceptual experience could not have occurred if the subject had not been perceptually related to precisely those elements of their environment. 1 In this respect, na€ ıve realist theories of perception differ from common kind theories of perception, like sense-datum or representationalist theories, which allow that how things are with the subject is constitutively independent, at least on a particular occasion, of how things are in their environment. According to common kind theories, it is possible for the subject to have fundamentally the same kind of experience whether or not the environment This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. 1 For defences of na€ ıve realism, see, for example, Martin (2002), Campbell (2002Campbell ( , 2014, Stoneham (2008), Fish (2009), Brewer (2011, Logue (2012). Na€ ıve realists disagree about exactly how na€ ıve realism should be characterised: for instance, whether it should be understood as a theory about what perceptual experiences are essentially (Byrne and Logue 2008), or whether there are perceptual experiences, as distinct from perceivings (Stoneham 2008). However, I will set these disagreements aside here and work with a representative form of na€ ıve realism.

THE VALUE OF PERCEPTION 1
Philosophy and Phenomenological Research is as it is perceived to be. This could either be because when we perceive we are related to different kinds of objects, sense data, that are distinct from things in our environment; or it could be because perceptual experiences are not essentially relational at all, but rather represent things in our environment as being a certain way, and are veridical just in case these things are as they are represented as being. 2 The aim of this paper is to develop and defend a version of a position that I will call transcendental na€ ıve realism. Transcendental na€ ıve realism involves adopting a specific kind of meta-philosophical attitude towards na€ ıve realism: it involves regarding the na€ ıve realist theory of perception from a 'transcendental standpoint'. There are different forms that transcendental na€ ıve realism can take, corresponding to different ways of understanding na€ ıve realism, its motivations, and exactly what adopting the transcendental attitude involves. However, in very broad terms, the transcendental na€ ıve realist will not see na€ ıve realism simply as one philosophical theory amongst others, to be defended by inference to the best explanation and to be assessed on the basis of a cost-benefit analysis that weighs performance along a number of different dimensions: for instance, fidelity to appearances, simplicity, systematicity, fit with scientific theories, and so on. Viewed from the transcendental perspective, na€ ıve realism has a special status amongst philosophical theories; indeed, for the transcendental na€ ıve realist, na€ ıve realism will be, in some sense, immune to falsification. The transcendental na€ ıve realist takes seriously the perspective of consciousness and the nature and structure of conscious experience. From this perspective, na€ ıve realism is part of the transcendental project of explaining how it is possible that perceptual experience has the distinctive characteristics that it does.
By way of illustration, consider, the prominent line of argument for na€ ıve realism that na€ ıve realism articulates and explains the phenomenal character of experience, or 'what it is like' to be a subject of experience. 3 Perceptual experiences appear to put us into contact with the mind-independent world. They are not only 'transparent', in the sense that when we reflect on our experience we are aware of mind-independent objects and their properties, and not anything else. But moreover, perceptual experiences arguably appear to involve the unmediated presence of mind-independent objects and their properties, such that objects and properties in our environment themselves determine the phenomenal character of our experience. This, for instance, is one way of understanding C.D. Broad's (1952: 5) suggestive description of visual experiences as ostensively 'saltatory' and 'prehensive': visual experiences appear to 'leap the spatial gap' between us and the mindindependent environment, and put us in a touch-like contact with mind-independent objects and their properties. 4 For a transcendental na€ ıve realist, the claim that na€ ıve realism articulates and explains the phenomenological character of perceptual experience should not be understood as the claim that na€ ıve realism is (merely) the best explanation of 'what it is like' to be a subject of experience. Rather, the transcendental na€ ıve realist will see na€ ıve realism as providing an account of how it is possible that conscious perceptual experience puts us into contact with things in our environment, as it appears 2 For representative forms of sense-datum theory, see, for example, Price (1932) and Robinson (1994). For representative forms of representationalist theory, see, for example, Evans (1982), Burge (2010), and Siegel (2010). to: the unmediated presence of mind-independent objects and their properties is only possible because perceptual experience is essentially relational, and what we are related to in perception are elements of our mind-independent environment.
Similarly, consider the equally prominent epistemological argument for na€ ıve realism that na€ ıve realism explains how we can refer to, think about, and acquire knowledge of mind-independent objects. 5 A transcendental na€ ıve realist will not understand this as an inference to the best explanation: that na€ ıve realism provides (merely) a better explanation than competing theories of perception of how experience fulfils its distinctive epistemic roles. Rather, a transcendental na€ ıve realist motivated by this kind of consideration will see na€ ıve realism as a condition of the possibility of perceptual experience having these distinctive epistemic characteristics: it is only possible to refer to, think about, and acquire knowledge of mind-independent particulars because perceptual experience is essentially relational, and what we are related to in perception are elements of our mindindependent environment.
Transcendental na€ ıve realism is arguably not a new position. Versions of this approach may perhaps be found in Kant (see Allais 2015, Gomes 2017), Merleau-Ponty (see Allen forthcoming), and Putnam (1999). The position has similarities to the disjunctivist theory of McDowell (1982McDowell ( , 2008, who uses transcendental arguments to motivate a form of representationalism that rejects the common kind assumption. 6 Indeed, given that contemporary na€ ıve realism developed largely against the background of neo-Kantian Oxford Realism (see Kalderon and Travis 2013), prominent contemporary na€ ıve realists like Martin, Campbell, and Brewer might themselves be best understood as transcendental na€ ıve realists, at least to the extent that the phenomenological and epistemic arguments that they use to motivate na€ ıve realism can be understood as transcendental arguments. 7 This paper develops a concrete illustration of one form that transcendental na€ ıve realism can take, modelled loosely on Strawson's 'Freedom and Resentment'. One aim of approaching the issues in the philosophy of perception indirectly, via a comparison with a debate in a different area of philosophy, is to provide a way seeing a familiar debate in the philosophy of perception from a different perspective. Strawson's 'Freedom and Resentment' represents a promising comparator for a number of reasons. As well as being well-known, it promises to provide a way of bringing out connections between perception and value. It also has the advantage of representing a fairly modest form of transcendental approach; as such, it is more likely to appeal to-or at least, less likely to rankle with-the sensibilities of contemporary analytic philosophers of perception than a more full-blooded form of transcendentalism. It will hopefully be particularly appealing 5 For versions of this argument see, for instance, Campbell (2002Campbell ( , 2014 and Brewer (2011). 6 McDowell's disjunctivist theory is not a form of na€ ıve realism if it is taken to be an essential commitment of na€ ıve realism that perceptual experiences are most fundamentally non-representational: McDowell seeks to explain the relationality of perception in terms of its representational content. For further discussion see Sotoriou (2013: 107) and Allen (forthcoming). Note that not everyone takes this to be an essential commitment of na€ ıve realism, in which case McDowell's view can be understood as a form of na€ ıve realism (see e.g. Crane and French 2015), and thereby a form of transcendental na€ ıve realism. 7 This way of understanding Campbell's version of the epistemic argument for na€ ıve realism, for instance, is suggested by claims like 'I will argue that if we are to acknowledge the explanatory role of experience of objects, we have to appeal to what I call a Relation View of experience' (2002: 114, emphasis added). However, I will not consider here the exegetical question of whether contemporary na€ ıve realists are best understood as transcendental na€ ıve realists. to contemporary na€ ıve realists, given that contemporary na€ ıve realism is a theory that has its roots in the neo-Kantian Oxford Realist tradition, of which Strawson is a key figure.
Nevertheless, the comparison is not without its limitations. 'Freedom and Resentment' is in many ways a puzzling paper, whose argument is difficult to pin down, and which has spawned a large literature of its own. And although the outlook developed and defended in this paper is broadly Strawsonian, I will argue that at certain points the comparison, crucially, breaks down.
§2 starts by providing a brief outline of some of the key aspects of Strawson's argument in 'Freedom and Resentment'. §3 draws a parallel to the debate between na€ ıve realists and common kind theorists. §4 extends the comparison by emphasising the importance of acknowledging the value that we ascribe to the engaged, first person perspective. In §5 I argue that the parallel to 'Freedom and Resentment' breaks down insofar as the view that the comparison suggests-and indeed, the view of perception that Strawson himself defends-is not a na€ ıve realist theory of perception, but rather a form of common kind theory. However, I also argue that the official Strawsonian position fails to meet the desideratum that Strawson himself lays down: that of respecting the engaged perspective. §6 concludes by considering whether, and in what sense, a form of transcendental na€ ıve realism loosely modelled on Strawsonian lines can be said to be immune to falsification.

Freedom and Resentment
'Freedom and Resentment' concerns the apparent conflict between determinism, the thesis that all our actions are causally determined, and the reactive attitudes we experience when we engage in inter-personal relationships: for example, feeling resentment in response to acts of ill-will, or feeling gratitude in response to acts of kindness. Strawson's aim is to try to effect a reconciliation between the 'optimist' who thinks that free will and determinism are compatible, and the 'pessimist' who thinks that they are not.
The pessimist thinks that genuine, metaphysical, free will is-or would be-necessary to justify the reactive attitudes, as well as social practices like punishment and moral approval and condemnation more generally. The pessimist represents two different positions: the incompatibilist, who thinks that the reactive attitudes and the social practices of punishment, moral approval and condemnation are not justified; and the libertarian-typified, for Strawson, by the transcendental idealist-who thinks that they are justified, and so who thinks that human actions are not causally determined in the way that the truth of determinism requires. The optimist, meanwhile, represents a particular kind of compatibilist who thinks of, and attempts to justify, the reactive attitudes and social practices like punishment and moral reward in 'objective terms', as appropriate given their role in the effective and efficient regulation of social behaviour (1962/2008: 2).
As Strawson's optimist understands it, determinism is the view that all actions are causally determined in the way that we regard actions to be causally determined in certain kinds of 'bad case': for instance, cases involving very young children and the severely mentally ill. But as a matter of what he elsewhere calls 'descriptive metaphysics' (1959), Strawson argues that this understanding of determinism misrepresents 'the facts as we know them'.
Strawson distinguishes between two different kinds of case in which we ordinarily think of the reactive attitudes as inappropriate (1962/2008: 7-10). The first kind of case does not involve thinking of an agent as being such that reactive attitudes towards them are inappropriate in general; it is just that a certain action, in certain circumstances, does not merit a particular reactive attitude such as gratitude or resentment. These are situations in which may say of someone that they didn't realise what they were doing, or couldn't help doing what they did because (for example) they were under duress. The second kind of case-which might involve very young children or the severely mentally ill-is one in which we do not think of the person as the appropriate object of the reactive attitudes more generally (or at least, not for a certain period of time). Thinking of someone like this involves taking an objective attitude towards them.
Strawson argues that the truth of determinism, as understood by the optimist, requires thinking of all human actions on the model of the second kind of case; after all, determinism is not the view that whenever anyone acts considerations of the first kind (e.g. inattentiveness, duress) are always operative (1962/2008: 11). But ordinarily when we adopt the objective attitude towards someone in the second kind of case we do not do so because we assume that determinism is, in general, true. Rather, we adopt the objective attitude because we think that the person is incapacitated in some way: for instance, because they are too young, or they acted out a fantasy (1962/2008: 13). Hence Strawson's claim that the optimist's understanding of determinism misrepresents the facts as we know them.
Strawson concedes that there is some rational sense in which the optimist's position is an intelligible philosophical doctrine. It is intelligible to the extent that we are able to adopt towards everyone the same detached, objective attitude that we adopt in the second kind of 'bad case': we understand what it is for action to be causally determined, so we understand (in some sense) what it would be for all action to be causally determined. But whilst Strawson thinks that determinism so understood is not self-contradictory, and indeed there may even be some general truths that could be theoretical grounds for it, he nevertheless thinks that the thesis is not practically intelligible. As he puts it, the belief in this form of determinism: is, for us as we are, practically inconceivable. The human commitment to participation in ordinary inter-personal relationships is, I think, too thoroughgoing and deeply rooted for us to take seriously the thought that a general theoretical conviction might so change our world that, in it, there were no longer any such things as inter-personal relationships as we normally understand them. (1962/2008: 12) According to Strawson, the reactive attitudes are essential to our ordinary understanding of inter-personal relationships: 'The existence of the general framework of [reactive] attitudes itself is something we are given with the fact of human society'. Questions of the justification of the reactive attitudes can, and do, arise within this framework. But the framework as a whole 'neither calls for, nor permits, an external 'rational' justification' (1962/2008: 25).

Na€ ıve Realism and Resentment
The comparison that I want to develop and explore starts from the observation that the optimist's claim that all actions are determined in the way that they are in the 'bad case' is structurally similar to the common kind theorist's claim that all perceptual experiences are of fundamentally the same kind as they are in certain kinds of 'bad case' of hallucination and illusion. According to the na€ ıve realist, perceptual experiences are essentially, or most fundamentally, relational: they consist in the obtaining of a relation of conscious awareness between perceiving subjects and things in their environment. It follows that there is a fundamental psychological difference between 'good' cases and 'bad' cases. According to the common kind theorist, by contrast, the 'good' case in which we veridically perceive our environment, and the 'bad' case in which don't, are mental events of the same fundamental kind. This is either because perceptual experiences involve the awareness of sense-data that are distinct from ordinary material objects, and so which are directly present in both 'good' and 'bad cases'; or it is because perceptual experiences are essentially representational mental events, that are veridical just in case the perceiver's environment is as it is presented as being.
Adapting the line of reasoning that Strawson deploys against the optimist, the na€ ıve realist can argue that common kind theories misrepresent 'the facts as we know them'. Mirroring Strawson's account of the two kinds of case in which reactive attitudes are inappropriate, we can think of two kinds of situation in which subjects lack perceptual contact with the world. In the first, we don't think that the subject has lost perceptual contact with the world completely; it's just that particular experiences they are having, in particular situations, are misleading. So, for instance, suppose that someone sees a patch of light on a path as a stone, a tilted penny as elliptical, or the lines of a Muller-Lyer image as unequal in length. Assuming, at least for the sake of argument, that these experiences involve a local loss of perceptual contact with the environment, misperceiving one aspect of the scene is nevertheless consistent with veridically perceiving the rest of the scene: for instance, the path and trees alongside it, the colour and position of the coin, or the book in which Muller-Lyer illusion is printed. 8 Perceivers in these kinds of cases can themselves normally unmask the illusion by further perceptual investigation: for instance, by viewing the object from a different perspective or under different conditions. And even where illusory experiences within a subject are systematic and pervasive, such as the colour experiences of people who are colour-blind or the blurred experiences of people with myopia, it isn't plausible to suppose that these perceivers have lost perceptual contact with the world completely; it's just that certain aspects of the world are unavailable to them. These are cases of illusion, but there are similar cases of hallucination. Suppose, for example, you are in your bedroom at night about to fall asleep and 'insert' a dead relative into the otherwise veridically perceived scene. These kinds of case involve (at best) the local suspension of perceptual contact with the environment. The second general kind of case, by contrast, involves a wider incapacitation. So, for instance, in these kinds of case the subject may be suffering from severe mental illness, be subject to the whims of a neuroscientist meddling with their cortex, or be in the Matrix.
The truth of the common kind assumption is not normally taken to involve the claim that whenever we perceive, considerations of the first kind are always operative: that we 8 Indeed, whether these kinds of case involve any loss of perceptual contact with the environment is controversial. In some cases of illusion there may plausibly be aspects of the mind-independent environment that we remain in contact with, and which explain how things appear: for instance, patches of light, or in the case of constancy illusions like the tilted penny, 'apparent' properties that are relational mind-independent properties of things in our environment (e.g. Schellenberg 2008, Allen 2018. Alternatively, it might be suggested that although experiences themselves sustain complete contact with our environment, we are nevertheless disposed in certain circumstances to make erroneous judgments about how things are on the basis of experience (e.g. Fish 2009, Brewer 2011). However, I will set these questions aside here. always misperceive every aspect of a scene, or that every object we see is really an hallucination. The truth of the common kind assumption seemingly requires thinking of all perceptual experiences on the model of second kind of case. But nor do we ordinarily think that subjects in the second kind of case-those cases involving a wider incapacitation-are not perceptually related to their environment because we think, in general, that the common kind assumption is true. Rather, we ordinarily think that subjects are not in perceptual contact with their environment in these kinds of cases because we can tell some other story about why they are perceptually incapacitated. So like the optimist, the common kind theorist can therefore be said to misrepresent 'the facts as we know them'.
Of course, the na€ ıve realist may concede that there is some rational sense in which common kind theories are intelligible. We know that there are 'bad cases', in which experiences do not put us into perceptual contact with the world as the na€ ıve realist understands it, as involving the unmediated presence of objects and their properties. This is what happens in cases of hallucination and perhaps illusion. The falsity of na€ ıve realism is rationally intelligible to the extent that we can adopt towards all our experiences the same detached, objective attitude that we adopt to these 'bad cases': we understand what it is for experiences not to put us in contact with the world in the way that the na€ ıve realist conceives of it, so we understand (in some sense) what it would be for no experiences to put us into contact with the world in the way that the na€ ıve realist conceives of it. These views are not self-contradictory, and there may even be theoretical grounds for them-namely different versions of the Arguments from Illusion and Hallucination, which attempt to generalise from 'bad cases' to 'good cases' via either the subjective indiscriminability of veridical and non-veridical experiences, or their similar proximate causal aetiology in the brain. But the na€ ıve realist might nevertheless argue that these theories are not practically intelligible. To paraphrase Strawson, they might suggest that: "The human commitment to perceptual contact with the world is, I think, too thoroughgoing and deeply rooted for us to take seriously the thought that a general theoretical conviction might so change our world that, in it, there was no longer perceptual contact with the world as we normally understand it".
Just as Strawson argues that the reactive attitudes are integral to inter-personal relationships as we ordinarily understand them, the na€ ıve realist may suggest that 'The existence of the general framework of perceptual contact is something we are given with the fact of the world'. Legitimate questions about whether we are in perceptual contact with the world can, and do, arise within this framework. But the framework to which perceptual contact is integral neither calls for, nor permits, an external 'rational' justification.

Perceptual Utilitarianism and the Value of Perception
I will return to the claim that na€ ıve realism is beyond external 'rational' justification in §6. But first I want to explore in more detail the comparison with 'Freedom and Resentment', and in particular the claim that na€ ıve realism articulates a core element of our conceptual scheme. An initial response to this Strawsonian line of argument is that common kind theories of perception, like sense-datum and representationalist theories, can provide an understanding of 'perceptual contact' which-though it falls short of the na€ ıve realist'sis sufficient to explain our perceptual engagement with the world. These views can explain our perceptual engagement with the world insofar as they are able to explain the link between perception, belief, and action: the world causes in us experiences (understood either in terms of sense-data or representations), which in turn cause our beliefs, and these in turn cause us to act in various ways. Moreover, it is still possible for common kind theorists to draw a distinction between 'good' and 'bad' cases. Just as causal compatibilists in the free will debate argue that the truth of determinism is consistent with an understanding of 'free action' that makes experiencing reactive attitudes appropriate, so long as the action is caused in the right way, the common kind theorist can say that the essentially non-relational nature of perceptual experience is consistent with understanding veridical perceptual experiences as putting us into 'contact' with the world, so long as they are caused by the subject's environment in the appropriate way.
But arguably this doesn't get to the heart of the matter. One way of bringing this out is by comparing the common kind theorist to Strawson's compatibilist optimist, who attempts to justify the appropriateness of the reactive attitudes on utilitarian grounds, in terms of the 'efficacy of...regulating human behaviour in socially desirable ways ' (1962/ 2008: 2). Strawson describes this as a form of 'incomplete empiricism, a one-eyed utilitarianism' (1962/2008: 25), that 'is not a sufficient basis, it is not even the right sort of basis, for these practices as we understand them ' (1962/2008: 4). For Strawson, adopting this kind of objective attitude towards the reactive attitudes and their value seemingly leaves out something vital to our understanding of them: it overlooks the engaged perspective of someone participating in ordinary inter-personal relationships, who is prone to experiencing gratitude and hurt-feelings. The vital thing that they leave out: can be restored by attending to that complicated web of attitudes and feelings which form an essential part of moral life as we know it, and which are quite opposed to objectivity of attitude. (1962/2008: 24) For Strawson, the reactive attitudes are not merely instruments for regulating human behaviour; rather, on a personal level, they are expressions of moral attitudes (1962/ 2008: 27). Common kind theorists who, like Strawson's optimist, would seek to explain our perceptual contact with the world in terms of the beliefs and actions to which perceptual experiences give rise can be thought of as offering a similarly 'utilitarian' account of perception and its value. Common kind theories that treat experiences in this way arguably leave out something vital, namely the perspective of the engaged subject. What it is they leave out can be restored by attending to your experience of, and engagement with, the world. Just as the reactive attitudes are merely not instruments for regulating social interaction, so we may say that perceptual experiences are not merely instruments by which we navigate the world. Rather, by relating subjects to things in their environment, perceptual experiences are an expression or manifestation of our being-in-the-world, and the fact that we are bodily perceiving subjects embedded in a mind-independent world. 9 9 This way of putting the point is inspired by Merleau-Ponty; see Allen (forthcoming) for further discussion.
If this is right, then it suggests that it would be misguided to try to motivate na€ ıve realism on equally 'utilitarian' grounds: for example, as the theory that best explains why perceptual experiences are instrumentally valuable in virtue of their role in allowing us to navigate the world. This doesn't address the basic problem with common kind theories. Rather, the pull of na€ ıve realism is visceral, even moral; the na€ ıve realist should be liable to find themselves in a similar position to Strawson's libertarian pessimist, who experiences 'emotional shock', or even offense, at the optimist's claim that punishment and the reactive attitudes can be justified on utilitarian grounds (1962/2008: 22). At any rate, if the na€ ıve realist does think of na€ ıve realism as necessary to explain the epistemic role of perceptual experience-for instance in enabling us to refer to, think about, and acquire knowledge of mind-independent objects-the knowledge of the mind-independent world that perception provides should itself be thought of as an expression or manifestation of our being-in-the-world, and not merely instrumentally valuable. 10 This may explain the feeling it is possible to have-and also the impression that na€ ıve realists are sometimes liable to give-that there is something distasteful, perhaps even grotesque, about common kind views that deny our perceptual contact with a mind-independent world. This kind of sentiment is reflected in Hume's description of his sceptical doubts as inducing 'melancholy and delirium ' (1739-40: 1.7.7, 175), and is nicely expressed by Russell in his autobiography when he describes rejecting British Idealism. This, he says, felt like: a great liberation, as if I had escaped from a hothouse on to a wind-swept headland. I hated the stuffiness involved in supposing that space and time were only in my mind. I liked the starry heavens even better than the moral law, and could not bear Kant's view that the one I liked best was only a subjective figment. In the first exuberance of liberation, I became a naive realist and rejoiced in the thought that grass is really green, in spite of the adverse opinions of all philosophers from Locke onwards.
(1959: 48) 11 Indeed, the sense that our perceptual experiences-though, of course, they do enable us to navigate the world-are not merely instruments for navigating the world may be part of what lies behind the intuitive resistance that many people feel towards the prospect of plugging into an experience machine that would guarantee them a lifetime of pleasurable experiences. Nozick uses the intuitive resistance to this possibility to argue to there is more that matters to us than the kinds of pleasure and pain that our experiences give rise to. At least part of what is objectionable about permanently plugging into an experience machine, according to Nozick, is that it denies us perceptual contact with reality: plugging into an experience machine limits us to a man-made reality, to a world no deeper or more important than that which people can construct. There is no actual contact with any deeper reality, though the experience of it can be simulated. (1974: 43) 10 See §6 for further discussion of this theme.

11
Compare Stroud's claim that attempts to dispel the disappointment we are likely to feel in response to the claim that we are 'imprisoned' behind a veil of appearances provide only cold-comfort: 'It is natural and perhaps always advisable for a prisoner to try to make the best of the restricted life behind bars. But however much more bearable it makes the prospect of life-imprisonment, it should not lead him to deny the greater desirability, let alone the existence, of life outside ' (1984: 34).
Although Nozick's primary aim is to argue that we don't think of experiences as valuable because they are instruments for deriving pleasure, this is part of a more general, noninstrumental view of the value of experience: What we want and value is an actual connection with reality....To focus on external reality, with your beliefs, evaluations, and emotions, is valuable in itself not just as a means to more pleasure or happiness. And it is this connecting that is valuable, not simply having within ourselves true beliefs. (1989: 106) For Nozick, perceptual contact with a mind-independent world is not valuable because of what it allows us feel or do; it is intrinsically valuable. This is not to say that there can be no value in plugging into an experience machine, putting on a virtual reality headset, or taking mind-altering drugs. These experiences-like forms of imaginative engagement, such as daydreaming or engaging with fiction-can also be valuable in their own way. The thought is just that there is something distressing about the idea that these experiences might permanently exclude or replace our genuine contact with the world. Perhaps particularly telling in this respect-connecting up the discussion of perception with the discussion of the reactive attitudes-are the kinds of perceptual experiences involved in ordinary inter-personal relationships: when we look in the eyes of a loved one, touch a lover, or hear the resentment in someone else's voice. The resistance to plugging into the experience machine does not depend solely on the prospect that the experiences we will enjoy will be unique to us. If others were also plugged into the machine in such a way that our individual experiences were coordinated with each other, and so we occupied different perspectives in the same virtual world, then this would still be distressing (cf. Nozick 1989: 107). It is not just the lack of contact with the physical world that is distressing; part of what is distressing is the lack of contact that we would thereby have with other people.

The Common Kind Theory and the Engaged Perspective
So far I have emphasised the importance of recognising the engaged, first person perspective in giving an account of perceptual experience. A second line of objection at this point is that acknowledging the value of perception from the first person perspective is perfectly consistent with accepting a common kind theory; indeed, it is precisely this combination of views to which the comparison with 'Freedom and Resentment' points, and which Strawson himself defends in his writings on perception.
Strawson's aim in 'Freedom and Resentment' is to effect a reconciliation between the optimist and the pessimist. Strawson wants to encourage the optimist to recognise the lacuna that the pessimist finds morally objectionable in their defence of compatibilism, but without thereby rejecting determinism; his hope is that 'there might be a formal withdrawal on one side [the pessimist's] in return for a substantial concession on the other [the optimist's]' (1962/1998: 2). What Strawson thinks the optimist needs to recognise is the engaged perspective of a moral subject; but, contrary to the pessimist, he does not think that recognising this requires us to resort to the 'obscure and panicky metaphysics' of libertarianism. Rather, he thinks (or seems to think) that recognising the value of the engaged perspective is consistent with a commitment to determinism. 12 This points to a tension in the comparison that I've been drawing so far. The na€ ıve realist may seem more like the pessimistic libertarian than the Strawsonian compatibilist who seeks to combine a belief in determinism with a recognition of the value of the engaged perspective. Taking the analogy seriously, it might be suggested that the conclusion we should draw is not that common kind theories of perception are practically inconceivable, but rather that what is practically inconceivable are just versions of the common kind theory that think of perceptual experiences in objective terms: for instance, versions of 'belief theories' of perception according to which perceptual experiences are just the acquiring of beliefs (e.g. Armstrong 1968), forms of representationalism according to which perceptual experiences are themselves belief-like mental events (for example, because the content of experience is conceptual, e.g. McDowell 1994), or perhaps forms of reductive naturalism that attempt to explain perceptual experiences in purely physical terms (e.g. Tye 2000). The more authentic Strawsonian position might seem to be a version of the common kind theory that recognises the distinctively subjective point of view.
This would be consistent with the theory of perception that Strawson himself defends. Against the sense-datum theorist, Strawson insists that perceptual experience typically presents itself as an awareness of a mind-independent world. Hence, if you ask someone to describe their experience, it would be natural for them (at least if they were a rather sophisticated na€ ıve subject, of the kind that you might meet at Oxford high table) to say something of the following kind: 'I see the red light of the setting sun filtering through the black and thickly clustered branches of the elms; I see the dappled deer grazing in groups on the vivid green grass... ' (1979: 97). But Strawson is not a na€ ıve realist, for whom perceptual experiences are essentially relational; indeed, for Strawson, contemporary na€ ıve realism is symptomatic of a 'disorder' affecting much recent philosophical thought, which attempts to 'minimize or reduce or even, in extreme cases, to deny the reality of what I shall unashamedly refer to as inner or subjective experience ' (1998a: 292). Rather, Strawson thinks of veridical perceptual experiences and hallucinations and illusions as mental events of fundamentally the same kind, that are differentiated in terms of the causal relationship that they stand in to that which they are experiences of. Though he does not himself characterise his view this way, we can perhaps think of Strawson's view in terms of views suggested by two of his pupils: either, as Evans (1982) argues, a form of intentionalism according to which the content of perceptual experience is nonconceptual; or, as Cassam (2014) suggests, a non-reductive form of intentionalism according to which the representational content of experience is determined by conscious, sensational aspects of experience. 13 Strawson's inclination towards determinism, and his commitment to the common kind theory of perception, reflect his more general commitment to a form of naturalism. For 12 Modulo the qualification that Strawson confesses to be unsure exactly what the thesis of determinism is (1962/2008: 1). 13 It should be noted, however, that neither view is entirely unproblematic from a Strawsonian perspective. On the one hand, it isn't clear that Evans's view fits with Strawson's insistence that perceptual experiences are 'thoroughly permeated-saturated, one might say' with the concepts employed in judgments made on the basis of experience (e.g. 1992: 62). On the hand, the distinction that Cassam draws between different 'layers' of experience-the representational and the non-representational-may not fit neatly with Strawson's criticisms of sense-datum views which think of experiences in two stages (1992: 62). Strawson, the objective and the participant perspectives are really just two different perspectives that it is possible to take on one and the same thing. Even though these standpoints appear to be radically opposed, once the relativity of perspectives is acknowledged, he thinks that there is no genuine tension between them (e.g. 1979: 59).
However, Strawson's attempts to reconcile the engaged and objective standpoints, both in general and in particular cases, face a number of problems. In general, to effect an equitable reconciliation between the manifest and scientific descriptions of the world, there cannot be a reason to prefer one description as capturing the world 'as it is really is' at the expense of the other. Strawson's suggestion that what is 'real' is relative to our perspective can be understood as attempting to meet this desideratum by denying that there is a perspective-independent reality, in something like the way that you might think that there are no perspective-independent facts about whether or not cabbage is disgusting. On this view, aspects of the manifest image would be real relative to the perceptual perspective, not real relative to the scientific perspective, and there would be no further fact of the matter about the way the world is that would allow us say that either of these descriptions is a more accurate representation of reality than the other. Put in these terms, however, this is liable to sound like it involves a potentially unappealing form of antirealism. 14 Certainly Strawson himself seems to prefer the description of the world provided by science as the description that captures how the world 'as it is in itself'. This comes out in his response to Snowdon, another of his students, who defends a disjunctivist account of perceptual experience of the kind that na€ ıve realists often appeal to in order to explain cases of hallucination. According to Strawson, the view that veridical perceptual experiences are essentially relational requires us to suppose that there are logical relations between 'natural items', as opposed to logical relations between descriptions of natural items; and this, he thinks, is tantamount to being a 'category howler' (1998b: 314; see also 1979: 136-7 and 1998a: 289-90). The implication is that for Strawson it is ultimately only the scientific descriptions that adequately characterise the 'natural items', and so the description of the manifest image of the world is ultimately a mere description. 15 These problems with the general structure of Strawson's attempt to reconcile the participant and objective perspectives mirror concerns about the attempted reconciliation in particular cases. The Strawsonian complaint against the optimist is that they think of attitudes and experiences in objective terms, and so miss out what is vital to the participant standpoint. In the case of perception, however, this lacuna arguably cannot be filled simply by pointing out that perceptual experiences are mental events with distinctive qualitative characters-that there is 'something it is like' to be a subject of experienceand suggesting that this qualitative character can be explained by appealing to non-conceptual representational contents or conscious sensational properties of experience. The concern is that this still fails to take seriously the engaged perceptual perspective, and in particular the distinctive sense in which perceptual experience appears to put us in 14 Compare Stroud's (1984) objection to Carnap's structurally similar suggestion that existence questions are only intelligible internal to a linguistic framework-although unlike Carnap, Strawson does not think that we are able to choose which framework to employ (see §6 below). contact with the mind-independent world, where this is understood in the na€ ıve realist's sense as involving the unmediated presence of mind-independent objects.
According to the common kind theorist, the very same mental event could be occurring, at least on a particular occasion, whether or not the world is as it appears to be. Whether the experience is a veridical representation of the environment depends on whether it is causally related in the appropriate way to that which it is a representation of. But from the engaged perceptual perspective, our experiences arguably do not appear to be distinct causal effects of an independently existing world; they do not present themselves as the last link in a causal chain that starts with their objects. Rather, they appear to put us into contact with things in our environment in a way that is inconsistent with this kind of causal mediation. As A.D. Smith, for example, puts the point: when we perceive, we are not necessarily aware of a causal interaction between us and the world...Perceived objects do appear simply to be there. For consciousness they have an unmediated presence.
( 2002: 69) Because the very same mental event could be occurring whether or not the world is as it appears to be, the common kind theorist cannot explain why the phenomenal character of an experience appears to be determined by mind-independent objects and their properties by appealing to a relation of conscious awareness to those objects and properties. The best that they can do to explain this distinctive aspect of the engaged perspective is to appeal to a particular kind of representational content, or if they take seriously the claim that perceptual experience is relational, a relation of conscious awareness to something else: sense data that are distinct from mind-independent objects and their properties. 16 But neither fully explains our perceptual contact with the mind-independent environment. This is why Broad, for example, argues that although visual experience is ostensibly saltatory and prehensive of objects in our environment, this appearance ultimately cannot be taken at face value. Only na€ ıve realism explains how it is possible that perceptual experience puts us into contact with the mind-independent world: perceptual experience involves the unmediated presence of mind-independent objects because it is essentially relational, and what we are related to in perception are mind-independent objects and properties. 17 In defending the causal theory of perception, Strawson argues that the causal theory is implicit in our pre-theoretical view of ourselves and our relationship to the world 'from the very start ' (1992: 61). In support of this conceptual claim, Strawson argues that the assumption of the causal dependence of experiences on their objects is necessary to ground the assumption that our perceptual experiences are generally reliable; this in turn is necessary to ground the belief that perception is a way of coming to form true beliefs about the mind-independent world; and these beliefs, in turn, count as knowledge given their causal relationship to the world: 16 See, for instance, Price's insistence that although there is much we can doubt when we see a material object like a tomato, we cannot doubt the ''direct' presence' of something that is red and round (1950: 3). 17 An alternative line of response to the na€ ıve realist at this point would be to challenge the description of the engaged perspective as involving a kind of unmediated presence of mind-independent objects, such that the phenomenal character of experience is determined by those objects and their properties. I will not defend this description here, except to say that it is widely accepted that this is at least how experience seems, even by those (like Broad) who think that this is not how it is. For further discussion see Allen (forthcoming). the minimum that seems to be involved in the notion of sense perception generally yielding true judgements about an objective spatio-temporal world is that there should be some pretty regular relation of dependence of the experience enjoyed in sense perception on the way things objectively are...this is, in the broadest sense, a notion of the causal dependence of the experience on the objective features in question...It is not something we discover with the advance of science, or even by refined common observation. It is conceptually inherent in a gross and obvious way in the very notion of sense perception as yielding true judgements about an objective spatio-temporal world...Of course, what we do find out by refined or scientific observation and investigation is how this general relation of causal dependence is actually realized, what forms it takes, what mechanisms are involved in it.
(1992: 60-1; see also 1979: 51) But there are two important slides in this line of argument. The first is from the claim that experiences depend on the world to the claim that experiences causally depend, in the broadest sense, on the world. The second is from the claim that experiences causally depend, in the broadest sense, on the world, to the claim that experiences causally depend on the world in the more specific Humean sense that they are logically distinct existences.
The first transition would be denied by those who think that the world (in effect) provides reasons (as distinct from causes) for experience (e.g. Hyman 1992), or by someone like Merleau-Pony (1945) who thinks that the environment 'motivates' experience, where motives are distinct from both reasons and causes. 18 The second transition would be denied by causal pluralists who allow that there can be causal relations that are not Humean causal relations between logically distinct existences. On this view, for instance, our experiences may causally depend the mind-independent objective world in the broad sense that the world 'makes a difference' to our experience. But this need not be say that perceptual experiences causally depend on the world in a more restrictive sense, such that they are logically distinct existences standing at the end of a 'process' that begins with the object. 19 Fully capturing the sense that perceptual experience puts us into contact with a mindindependent world arguably requires an understanding of the way that experience depends on the world than is different to that offered by Strawson-if not in terms of a non-causal dependence, then at least in terms of a different kind of causal dependence. 20 But although this way of making sense of the distinction between the engaged and objective perspectives requires a different metaphysical picture to that presented by Strawson, it arguably does not requires us to have recourse to 'obscure and panicky metaphysics'or at any rate, not the 'obscure and panicky metaphysics' of anything like a full-blooded transcendental idealism. One way of reconciling the two perspectives, for instance, is to 18 A slightly different way of resisting this step would be to argue that there are no perceptual experiences, just perceivings, and that perceivings are the wrong kinds of things to stand in causal relations. See Stoneham (2008), and compare Valberg's (1992: 120-8) 'horizonal' conception of experience. 19 For the distinction between 'difference-making' and 'process' conceptions of causation, see Woodward (2011). For a defence of this kind of causal pluralism, see Allen (2016: 91-110). For further discussion of Strawson's argument for the causal theory, see Snowdon (1998) and Roessler (2011). 20 It is possible to have similar reservations about Strawson's attempt to reconcile pessimists and optimists in the free will debate. Allais (2014), for example, argues that we cannot provide an adequate account of how we can elect to give up warranted resentment when we forgive someone if we understand their actions as caused by beliefs, desires, and other psychological states. Rather we need to understand others as choosing to act for reasons that make certain kinds of actions intelligible; and this sort of free action, Allais argues, cannot itself be understood in deterministic causal terms, but instead needs to be understood in terms of a different kind of causality, namely 'the capacity to initiate a new causal sequence in a way which is not a necessitated function of previous states of the universe and the laws of nature' (2014: 51). think of them as providing complementary descriptions of different aspects of one and the same world. Whereas the scientific image describes the causal mechanisms that bring about the occurrence of perceptual experiences, perceptual experiences themselves are essentially relational events that are causally enabled by the processes described by science. It will be natural, on this approach, to think of the different aspects of the world that the different perspectives disclose as closely related, for instance via some kind of supervenience thesis; but they are nevertheless distinct. 21

Beyond Justification?
Part of the promise of the transcendental approach is that it renders na€ ıve realism, in some sense, immune to falsification. In the final section I want to consider whether, and in what sense, this promise can be fulfilled.
Like Strawson's approach to the apparent conflict between determinism and the reactive attitudes on which it is modelled, the form of transcendental na€ ıve realism that I have developed is not an 'ambitious' form of transcendental na€ ıve realism. I have not tried to guarantee the truth of na€ ıve realism by embedding it within the framework of transcendental idealism, something which we might worry really would involve appealing to the 'obscure and panicky metaphysics' of the transcendental ego. 22 Nor has it been argued that the rejection of na€ ıve realism is incoherent, meaningless, or otherwise rationally unintelligible. In this respect, the Strawsonian-inspired form of transcendental na€ ıve realism differs from stronger forms of transcendental na€ ıve realism which insist, for example, that a condition of the possibility of distinguishing between 'good' and 'bad' cases on particular occasions is that we have previously had experience of the 'good' case (e.g. Merleau-Ponty 1945: lxxx, 310-1). This would represent an ambitious form of transcendental na€ ıve realism because it would rule out the challenge to the na€ ıve realist theory of perception as either false or meaningless: if a condition of the possibility of framing an intelligible sceptical challenge is that we have had perceptual contact with the world, then the challenge is false; otherwise, it is unintelligible. 23 The concern with ambitious forms of transcendental arguments like these is that it seems sufficient to draw the distinction between 'good' and 'bad' case that we believe that we have experienced reality, not that we actually have. 24 But although Strawson's approach to the conflict between determinism and the reactive attitudes represents a 'modest' transcendental outlook, he nevertheless argues that the general framework of the reactive attitudes 'neither calls for, nor permits, an external 'rational' justification' (1962/2008: 25). There are two strands to Strawson's defence of the reactive attitudes, and to basic beliefs, attitudes, and practices more generally: one 21 For further discussion see e.g. Allen (2016: 183-185), and compare Allais's (2014) suggestion in the case of free will. 22 Compare a recent strand of Kant interpretation according to which Kant is a na€ ıve realist about the phenomenal world, e.g. Allais (2015), Gomes (2017). 23 Compare Putnam's claim that 'Winning through to natural realism [a form of na€ ıve realism inspired by James] is seeing the needlessness and the unintelligibility of a picture that imposes an interface between ourselves and the world (1999: 41). Note that the claim that the rejection of na€ ıve realism is needless seems to be weaker than the claim that it is unintelligible; I return to this below. Humean, the other Kantian. The Humean strand emphasises the psychological impossibility of giving up the reactive attitudes, whereas the Kantian strand emphasises connections between the reactive attitudes and other aspects of our conceptual scheme. 25 In what remains I will consider whether either of these lines of defence can provide the na€ ıve realist with a way of arguing that na€ ıve realism is beyond external 'rational' justification, and I will conclude by tentatively suggesting a slightly different way of reaching the Strawsonian conclusion.
The Humean line of argument for the claim that basic features of our conceptual scheme are beyond external 'rational' justification, and so can neither be vindicated nor undermined, appeals to their psychologically inescapability. As Strawson argues in Skepticism and Naturalism, where basic features of our conceptual scheme like this are concerned, sceptical arguments and rational counter-arguments are: equally idle-not senseless, but idle-since what we have here are original, natural, inescapable commitments which we neither choose nor could give up.
( 1985: 28) This defence of basic elements of our conceptual scheme is modelled on Hume's response to scepticism about inductive reasoning and the existence of the external world, and provides one way of understanding Strawson's claim that determinism is 'practically inconceivable' (1962/2008: 12): determinism is practically inconceivable in the psychological sense that we simply cannot believe it. Adapting this Humean line of argument in defence of na€ ıve realism, the corresponding thought would be that rejecting na€ ıve realism is practically inconceivable because the belief in na€ ıve realism is an original, natural, inescapable commitment that we neither choose nor could give up. This suggestion raises two questions. First, is the psychological claim that we cannot give up the commitment to na€ ıve realism true? Second, if the psychological is true, is this an adequate defence of na€ ıve realism?
First, the psychological claim isn't true in particular cases. That is, we are normally able to distinguish between episodes of perception, in which we are in perceptual contact with our environment, and episodes of illusion and hallucination, in which we are not. Often we are only able to make these distinctions retrospectively, when further experiences reveal the world to be other than we had taken it to be: for instance, when we realise that what to took to be a stone on the path is in fact just a patch of light, or when we try to touch a visual apparition. But sometimes it is possible to make the distinction between the 'good' and 'bad' cases as you are perceiving. In some cases it is possible to recognise the 'bad' case for what it is on the basis of the experience's phenomenal character, because the experience lacks the 'sense of reality' characteristic of veridical perceptual experiences: this is often true, for instance, of real-world hallucinations, in which objects are not presented at a determinate depth within a scene, or do not appear to vary as we move in relation to them. 26 In other cases, collateral information is needed to know that an experience is not veridical. If you are familiar with the Muller-Lyer illusion, for instance, you can know as you are perceiving that the lines are really of the same length. In a more extreme case, if you remember taking the red pill, 25 For helpful discussion of the two different strands-and the tensions between them-see e.g. Putnam (1998) andStern (2003). you can know as you are having them that your experiences within the Matrix are systematically illusory.
The ability to distinguish between the 'good' and 'bad cases' in particular instances plausibly presupposes, or is internal to, a more general framework or practice. As Strawson says in the case of the reactive attitudes: Inside the general structure or web of human attitudes and feelings of which I have been speaking, there is endless room for modification, redirection, criticism, and justification. But questions of justification are internal to the structure or relate to modifications internal to it. The existence of the general framework of attitudes itself is something we are given with the fact of human society. (1962/2008: 25) Similarly in the case of perception, the ability to distinguish between veridical and nonveridical perceptual experiences arguably presupposes a general framework for thinking about perception, such as that articulated by the na€ ıve realist theory of perception. Is it psychologically possible to call into question the entire framework within which we distinguish between the 'good' and 'bad cases' on particular occasions? To the extent that we are able to do this, it might seem that the belief that no experiences put us into contact with the world, as the na€ ıve realist understands it, is likely to be unstable. Na€ ıve realism forces itself upon us at every waking moment; even in the study, when we are doubting the truth of na€ ıve realism, we are thrown into the world through perception. As Hume says, even though he thinks there are reasons for rejecting na€ ıve realism, he nevertheless finds himself 'absolutely and necessarily determined to live, and talk, and act like other people in the common affairs of life (1739-40, 1.4.7). 27 Still, to say that the belief that na€ ıve realism is false is likely to be unstable is not yet to say that it is impossible. And certainly there are many, including Strawson himself, who seem from what they say to believe that na€ ıve realism is false. It might be that those who deny the truth of na€ ıve realism hold different beliefs at different times, different beliefs from different perspectives, or perhaps just simultaneously hold contradictory beliefs-both that perceptual experience puts us into contact with the world in the way that na€ ıve realism articulates, and that na€ ıve realism is false. But this is not yet to say that they don't believe that na€ ıve realism is false.
Besides, even if the psychological claim is true and the commitment to na€ ıve realism is inescapable, the second and more important question is whether this would thereby render na€ ıve realism immune to falsification; and it is by no means obvious that it would. Strawson, for instance, argues that rational arguments on either side are idle where 'original, natural, inescapable commitments which we neither choose nor could give up (1985: 28) are concerned. But it is unclear why the fact that we neither choose nor could give up these commitments puts them beyond external 'rational' justification. 28 One way of developing this point is by noting that, in effect, Strawson's argument for this claim appeals to an epistemic version of the principle of alternative possibilities. 27 Compare Strawson's claim that we can adopt objective attitudes towards others when faced with abnormal cases or in normal cases as a means of emotional self-defence, but 'I do not think it is a point of view or position which we can hold, or rest in, for very long. The price of doing so would be higher than we are willing, or able, to pay; it would be the loss of all human involvement in personal relationships, of all fully participant social engagement ' (1985: 34). According to the principle of alternative possibilities, you can only be held accountable for an action-either positively or negatively-if you were able to do otherwise. The epistemic version of this principle that Strawson appeals to is that you can only be held accountable for holding a belief-either positively or negatively-if you were able to believe otherwise.
The principle of alternative possibilities is normally associated with Strawson's pessimists (incompatibilists and libertarians), and its rejection with Strawson's optimists (compatibilists). Consistent with his preference for a modified form of optimism, Strawson is himself sceptical of the pessimist's appeal to the principle of alternative possibilities. According to Strawson, to say of someone that 'He could have acted otherwise' is not equivalent to saying that 'There was no sufficient natural impediment or bar, of any kind whatsoever, however complex, to his acting otherwise ' (1992: 136). When we say of someone that they could have acted otherwise, it is normally only with certain kinds of potential impediment in mind-leaving open the possibility of obstacles of other kinds. Moreover, if we deny that someone could have done otherwise because 'It simply wasn't in his nature to do so', this does not normally lead us to the regard the person's actions as beyond moral judgment-indeed, as Strawson notes, this kind of response might actually lead us to reinforce our moral judgment of someone (1992: 137).
By the same token, it is plausible to suppose that whether or not a belief is justified does not depend upon the absence of any kind of natural impediment to believing otherwise. And just as in the moral case, saying that someone could not have believed otherwise because 'It simply wasn't in his nature to do so' does not obviously put that belief beyond justification. Yet the fact that certain commitments are inescapable-that it is in our nature to believe them-is precisely what the Humean strand of Strawson's solution to sceptical challenges to our core common sense beliefs consists in.
Setting aside the Humean strand, Strawson's work also suggests a more optimistic, broadly Kantian line of reasoning in defence of the claim that basic beliefs, attitudes, and practices are beyond external 'rational' justification. In saying that the reactive attitudes are beyond external 'rational' justification, Strawson is aiming to rule out the possibility of either undermining or justifying the reactive attitudes from a perspective distinct from that of the engaged subject, such as the objective scientific perspective. But this leaves open the possibility of justifying the reactive attitudes from within the perspective of the engaged subject. Strawson's discussion suggests two forms that this kind of internal justification can take.
First, we can consider whether experiencing particular reactive attitudes in particular circumstances is appropriate. So, for instance, we can consider whether it is appropriate to feel gratitude towards someone who has done something that benefits us, even though they did not do it out of good-will towards us but for some self-serving reason. Second, and more generally, we can draw connections between different aspects of our conceptual scheme, and thereby provide a kind of 'connective', 'holistic' or 'horizontal' justification of one part of our conceptual scheme in terms of another, without thereby transcending our perspective. Strawson suggests the possibility of this more expansive form of internal justification when he states that 'The existence of the general framework of attitudes itself is something we are given with the fact of human society ' (1962/2008: 25). One way of understanding this claim is as the claim that the reactive attitudes are a transcendental condition of the possibility of ordinary interpersonal relationships (cf. Coates 2017). From the practical perspective, as agents engaged in ordinary interpersonal relationships, we cannot call into question the reactive attitudes that are a transcendental condition of those relationships. The reactive attitudes are essential to our conception of ourselves as humans engaged in inter-personal relationships. As such, these different elements of our conceptual scheme provide a 'mutually supportive natural metaphysics' (Strawson 1985: 29).
The na€ ıve realist could again adopt a similar line of argument. First, we can consider in particular cases whether our experience puts us into contact with the world, and here there is 'endless room for modification, redirection, criticism, and justification'. We are able to explore our environments, look at things from different angles, under different conditions, and perceive them via different sensory modalities. The different perspectives that we are able to adopt on things allow us to improve, confirm, or correct our experiences without stepping outside of the engaged, perceptual perspective. In this respect, there is a sense in which perceptual experience is self-justifying: doubts about the veracity of experience can be answered by experience itself. Second, we can seek to justify na€ ıve realism holistically, on the grounds that it is intimately connected to, and perhaps a transcendental condition of, our broader understanding of ourselves and our relationship to the world that we inhabit. Just as Strawson claims that the framework of the reactive attitudes is given with the fact of human society, the na€ ıve realist can argue that 'The existence of the general framework of perceptual contact is something we are given with the fact of the world'. Perceptual contact with the world is a condition of the possibility of our engagement with the world. From the practical perspective, as agents engaged with the world, we cannot call into the question the perceptual contact that is a condition of that engagement; perceptual contact is essential to our understanding of ourselves as humans engaged with a mind-independent reality. It is central to our understanding of the world, others, the past, and the future. 29 Of course, these kinds of internal justification will only take the na€ ıve realist so far. On the one hand, to the extent that perceptual experience is self-justifying, it does not provide the kind of independent justification that someone who is sceptical of the truth of na€ ıve realism would want. If what is in question is the truth of the perceptual framework that the na€ ıve realist theory articulates, then perceptual experience cannot provide an independent means of assuaging those doubts; the attempt to justify the perceptual framework by perceptual experience would be circular. On the other hand, it might be granted that na€ ıve realism is enmeshed within a wider web of beliefs, but nevertheless claimed that the whole web of beliefs in which it is enmeshed can, and should, be called into question. This is not merely an abstract threat, either. After all, the very same kinds of patterns of objective thinking that would lead someone to accept a common kind theory of perception are likely to lead them accept corresponding views elsewhere: for instance, that knowledge of the world is really just belief that meets some further conditions which distinguish it from 'bad cases' of mere belief; or that action on things in the world is a certain kind of bodily movement (or perhaps even just an attempt to move your body) that meets some further conditions which distinguish it from 'bad cases' in which you merely move you body (or merely try to move your body). The kind of objective thinking that threatens to undermine the na€ ıve realist theory of perception 29 Compare Strawson's claim that 'it would be hard to separate the conception of objects which we have and our acceptance of inductively formed beliefs from that conception of the past. All form part of our mutually supportive metaphysics ' (1985: 29). The same is arguably true of perception. simultaneously threatens to undermine the general view of ourselves as beings-in-theworld, which perception, as the na€ ıve realist understands it, is an expression or manifestation of.
It may be possible for the na€ ıve realist to go slightly further at this point. Without providing a positive external justification of the perceptual framework, they can at least seek to show that the perceptual framework is not undermined by challenges from an external standpoint. This kind of defence of basic elements of our conceptual scheme is something that Strawson himself seeks to provide when he attempts to diffuse the 'hard' naturalist's challenge, that not only is there is an incompatibility between the engaged and the objective standpoints, but that the engaged perspective gives way to the objective standpoint (1985: 36, 48-50). Strawson's appeal to the relativity of 'really' goes at least some way towards offering a rational defence of the engaged perspective: it is to say that it is not undermined, in the way it might seem to be, by the availability of an alternative, objective perspective. The variation on Strawson's approach (suggested in §5) which attempts to guarantee the compatibility of the two standpoints by seeing them as describing different aspects of one and the same thing fulfils a similar function. This is not a purely internal justification of the engaged perspective, since it involves transcending that perspective in order to resist an external challenge to it. But it is still in keeping with the spirit of a broadly transcendental approach insofar as it seeks to provide an explanation of how it is possible that, for example, the reactive attitudes are appropriate or that perceptual experience is relational.
However, this invites the following objection: to say that na€ ıve realism is not impossible is not yet to provide a reason for accepting it. Indeed, you can imagine an opponent of na€ ıve realism saying something similar to Salmon in his criticism of Strawson's (1952) defence of inductive reasoning as reasoning that just is justified: reasoning which is not justified because it is reliable, and not justified because we have independent grounds for thinking it is reasonable. Salmon complains that this defence of inductive reasoning elevates 'inductive method to the place of an intrinsic good', and 'The resulting justification of induction amounts to this: If you use inductive procedure you can call yourself 'reasonable'-and isn't that nice! ' (1957: 42). 30 Similarly, the opponent of na€ ıve realism might complain that the Kantian defence of the na€ ıve realist's view that perceptual experience is essentially relational has elevated perception as the na€ ıve realist conceives of it to the level of an intrinsic good. As such, the justification of na€ ıve realism amounts to this: We take perception to put us into contact with the mind-independent world-and isn't that nice! I want to conclude at this point by tentatively suggesting a slightly different way of reaching the Strawsonian conclusion that na€ ıve realism is beyond external 'rational' justification. Thinking of the visceral, even moral reaction, that the denial of na€ ıve realism is liable to provoke ( §4), perhaps the na€ ıve realist's response to this challenge should be 'Well, yes, it is nice-and it would be awful if the common kind assumption were true'. It is at least in the spirit of Strawson's claim that basic features of our conceptual scheme neither require, nor permit, external 'rational' justification to insist that the na€ ıve realist theory of perception is immune to falsification because the reason for accepting na€ ıve realism is not primarily rational, but emotional or even moral. To deny naive realism is to deny part of what it is to be human.
To say that the motivation for this kind of modest transcendental na€ ıve realism is ultimately non-rational might seem to put it at a disadvantage to other non-transcendental philosophical theories of perception: both to common kind theories of perception that are defended on the basis of the Arguments from Illusion and Hallucination, but also to forms of na€ ıve realism that are themselves defended on the basis of rational considerations, as (for example) that which best explains the phenomenological character of perceptual experience or the epistemic role of experience.
However, on the one hand, na€ ıve realist theories of perception are liable to seem vulnerable if they are treated simply as one philosophical theory of perception amongst others, to be assessed on the basis of a cost-benefit analysis that weighs performance along a number of different dimensions. Suppose, for instance, that na€ ıve realism is treated as a philosophical theory of perception of the same order as sense-datum and representationalist theories, and in arguing for it the na€ ıve realist seeks to accord particular importance to fidelity to the appearances in deciding between then competing philosophical theories of perception. This then raises the question of why fidelity to the appearances should be accorded a particular, or indeed any, privilege in assessing philosophical theories of perception? As Hawthorne and Kovakovich put this challenge, for example, an opponent of na€ ıve realism may well ask 'why vulgarity is to be celebrated ' (2006: 180)? At best, it might be suggested that respecting the appearances is just one consideration amongst many, one that provides a defeasible reason for accepting a philosophical theory of perception, but which can be outweighed by costs or benefits elsewhere. At worst it might be suggested that respecting the appearances is simply not a relevant consideration, any more than it would be if we were trying to explain why people believe in ghosts.
On the other hand, although there may be theoretical considerations (such as the Arguments from Hallucination or Illusion) in favour of common kind theories of perception, nevertheless the empirical facts do not obviously mandate these philosophical theories of perception; we can at least make sense of the how na€ ıve realism is possible, consistent with the facts as we know them. Given that we can tell a story about how na€ ıve realism is possible, we are not compelled to reject na€ ıve realism, or to accept that we are systematically mistaken about the nature of our experience. Why then should we reject it? In these circumstances, rejecting na€ ıve realism may look like an instance of what Strawson elsewhere describes as 'an extreme form of self-mortifying intellectual Puritanism ' (1985: 50