What’s so naïve about naïve realism?

Naïve Realism claims that veridical perceptual experiences essentially consist in genuine relations between perceivers and mind-independent objects and their features. The contemporary debate in the philosophy of perception has devoted little attention to assessing one of the main motivations to endorse Naïve Realism–namely, that it is the only view which articulates our ‘intuitive’ conception of perception. In this paper, I first clarify in which sense Naïve Realism is supposed to be ‘naïve’. In this respect, I argue that it is put forward as the only view which can take our introspective knowledge of perception at face value, and I identify the two (alleged) key features of such introspective knowledge. Second, I challenge the claim that one of these features-namely, that it seems as one could not be in the same perceptual state unless the putative objects of perception existed and were perceived–is introspectively evident. Consequently, I argue that a view of perceptual experience–such as Intentionalism–which denies that this feature is true of perception can still take introspection at face value. This undermines the claim that Naïve Realism is the only account which accommodates our intuitions on the nature of perception.

and that the greatest 'chasm' in the philosophy of perception is represented by the debate over perception being essentially relational or representational in nature (2005, p. 128). Preliminarily, we may characterise the two sides of the debate as follows: according to a relational account, perception essentially consists in a genuine relation between a perceiver and some perceived objects (independently from how they are conceived-i.e. as mind-dependent or mindindependent); according to a representational account, perception is the kind of state it is in virtue of representing things in a certain way-that is, in virtue of having a representational content.
The latter approach has constituted the dominant paradigm throughout the second half of the last century. 1 However, in the last couple of decades many philosophers have been attracted by a version of the relational account according to which not only perception is essentially relational, but it should be construed as a relation with mind-independent objects and their (equally mind-independent) features. The theories endorsing this fundamental claim have been labelled in different ways, such as the 'Object View' (Brewer 2011) or simply the 'relational approach' (Campbell 2002). In this paper, I will adopt the denomination of 'Naïve Realism' (Martin 1995 onwards;Nudds 2009;Fish 2009), which reflects the widely assumed presupposition that this approach best captures our 'naïve' intuitions about the nature of perception. While much of the recent debate focusses on some difficulties Naïve Realism encounters in accounting for cases of non-veridical perceptual experiences, 2 very little attention has been devoted to assessing this basic assumption, which arguably constitutes one of the main motivations for endorsing the view in the first place (Fish 2009, pp. 18-23). 3 The purpose of this paper is to clarify (Sects. 2 and 3.1) and, ultimately, to challenge (Sect. 3.2) the claim that Naïve Realism is the only account that can take at face value our intuitive conception of perception.
My analysis develops around two main questions: (1) what does it mean for a conception of perception to be 'naïve'? For instance, does it simply correspond to what a 'common' person would say about perception, or is it something more/ different? (2) What are the features of this conception Naïve Realism claims to be the only account to accommodate?
In the first section of this paper, I briefly present Naïve Realism by focussing on how it differs from two other theories, the Sense-Datum Theory and Intentionalism. In the second section, I focus on question (1) above; in a nutshell, my answer is that the advocates of Naïve Realism refer to it as 'naïve' in the sense that it best articulates our introspective knowledge of perception. In the third section, I focus on what is allegedly part of such introspective knowledge, and I challenge the claim that only Naïve Realism can account for it.
1 Naïve realism, sense-datum theory and intentionalism Naïve Realism endorses two main claims, the first ontological and the second phenomenological. The ontological claim is that perception is essentially relationalthat is, it fundamentally consists in a relation between a perceiver and one or more objects-and that the objects of perception are mind-independent, such as books, tables and so on. The phenomenological claim is that the phenomenal character of perception derives from this ontological structure.
The phenomenological claim is more difficult to specify, mainly because it relies on a technical notion-that of 'phenomenal character' (from now on 'PC')-that has proved to be particularly hard to handle in the recent literature on the philosophy of mind. 4 One common way to understand it involves the concept of 'what it is like' to be in an experiential state-that is, to how the experience strikes the subject to be 'from the inside'. As such, any conscious experiential state will have a property of 'being like' a specific way. The PC of the experience can be thus variously defined as what captures the 'what-it-is-like properties' of the experience (Martin 1997, p. 93), or to correspond to those properties of the experience that explain how the experience 'seems' to the subject (Nudds 2009, p. 335), or as 'the property of the experience that types it according to what it is like to have it' (Fish 2009, p. 9). 5 Therefore, the phenomenological claim amounts to the idea that the specific way perception 'seems' to the subject-what it is like to have it-is determined (Martin 1997, p. 94) by the fact that perception consists in a relation with mind-independent objects and their features. According to Naïve Realism, therefore, the PC of perception is a relational property, as the experience could not have it unless its relata-the perceiver and the perceived objects-existed. This is why the Naïve Realists often claim that the PC of perception is at least partially constituted by the perceived (mind-independent) objects, or that these objects 'shape the contours of the subject's conscious experience' (Martin 2004, p. 64).
In principle, the two claims could come apart: one could coherently endorse the ontological claim while rejecting that the PC of perception is determined by its relational structure (for instance, by claiming that perception also involves a representational content and that it is the nature of this content that determines its PC). But since the phenomenological claim arguably constitutes the main motivation to endorse Naïve Realism, and since the main purpose of this paper is to analyse this motivation, I will ignore this possibility in what follows. 6 Given its commitment to these claims, Naïve Realism is usually conceived as a theory of veridical perceptual experience only. Take, for instance, a situation in which someone is hallucinating an oasis in a desert. This experience obtains in the absence of the suitable object of experience-the oasis. Therefore, the ontological claim-and, hence, also the phenomenological one-appears to be false for cases of hallucination. 7 The fact that Naïve Realism does not account for the nature of misperception is at the basis of the main reason why it has often been considered as a non-starter position in the debate on perception. Consider again the hallucination above. It seems conceivable for that situation to be indiscriminable from a corresponding veridical one-one where the oasis is present in the subject's environment. This has been taken by many philosophers to entail that 'what it is like' to be in the two situations is the same, and hence that they have the same PC. But since the phenomenological claim is false for the hallucinatory case, it seems to follow that also the PC of the veridical case cannot be determined by the experience being genuinely relational. In turn, as I have mentioned above, this seems to undermine the main motivation to endorse the ontological claim as well. 8 The Sense-Datum Theory and Intentionalism can be seen as alternative ways to exploit the intuition that 'sense-experience'-the term I will use to identify the class of experiences including both veridical and non-veridical cases-should be analysed in terms of a common 'factor'-or a common 'fundamental' kind of mental state 9that is allegedly present both in the 'good' and the 'bad' cases. The Naïve Realist's analysis of experience cannot be applied to this alleged common-factor because some of the 'bad' cases do not consist in a relation with the mind-independent objects we take them to present us with. The Sense-Datum Theory deals with this issue by claiming that Naïve Realism is correct when it comes to identifying how the objects of perception are made manifest to the subject-that is, it agrees with Naïve Realism that perception is an essentially relational mental state. But since, according to what we may call the 'common-factor' assumption, this also applies to illusions and hallucinations, it follows that the objects of sense-experience are minddependent entities which are usually called 'sense-data'. 10 On the other hand, Intentionalism agrees with Naïve Realism that the objects of perception are mind-independent, but it disagrees over how these objects are made manifest to the perceiver. 11 According to Intentionalism, sense-experiences are representational states-that is, mental states characterised by the possession of representational content. 7 For reasons of space, I will not discuss whether a similar line of reasoning can be extended to perceptual illusions. 8 More work is needed to turn this objection into a proper argument, but this goes beyond the purposes of this paper. For some detailed analysis on this matter, see Robinson (1994) and Martin (2004). 9 See Martin (2006, pp. 360-61) for a more careful articulation of this point. 10 The notion of 'sense-datum' was actually introduced by Moore (1910) to identify whatever 'the object of sensory awareness' was, while the problem of defining its nature was a matter of further philosophical enquiry. For a long time, Moore remained undecided on whether construing sense-data as minddependent objects or as the visible parts of mind-independent ones (Martin, 2015). It is precisely because of the role played by the arguments from illusion and hallucination, however, that the former view is usually marked as the 'traditional' version of the Sense-Datum Theory. 11 I am indebted to Martin (1995Martin ( , 1998 for this analysis of the relationship between the three views. This idea is usually elucidated by drawing an analogy with other mental states that are traditionally considered representational, such as belief. Consider a case in which I falsely believe that there is a table right in front of me. In this situation, it would not seem right to claim that the object of my belief is a mind-dependent, or 'mental', object: what I believe to be in front of me is a perfectly physical, mindindependent object. 12 I just happen to be wrong in believing that this table was present in my surrounding environment-that is, the belief was representing my environment inaccurately. Intentionalism construes sense-experience in the same way. Therefore, what individuates it as the kind of mental state it is (including its 'what-it-is-like' properties), is the fact that it possesses an intentional content-a condition that a perceptual state can meet regardless the existence of the represented objects. The difference between veridical and non-veridical cases amounts to whether the intentional content is an accurate representation of the subject's environment or not.
To help us with the transition to the next section, the situation can be outlined as follows. According to Naïve Realism, veridical perceptual experience presents two fundamental features: 1. Transparency (T) It makes manifest to us mind-independent objects and their features; 2. Actualism (A) It makes the perceived objects manifest in a way that it cannot be the case that they do not exist and are not perceived. 13 Due to the mentioned considerations linked with the possibility of misperception, Sense-Datum Theories reject the claim that perception possesses (T) while endorsing (A) for both veridical and non-veridical cases: when it comes to senseexperience, it is true that when the subject senses an object, that object must exist and be sensed, but false that this object is mind-independent. On the contrary, Intentionalism rejects (A) while accepting (T): the sensed objects are mindindependent, but they need not exist for the experience to occur.
The problem is that, according to some philosophers, 14 (A) and (T) are both part of our 'intuitive' conception of perception. It follows that only Naïve Realism 12 See Harman (1990) for a development of this line of reasoning. 13 These labels come from Martin (2001, pp. 217-218). The second feature is a slightly modified version of Martin's one, which focusses on sense-experience. Many Naïve Realists, however, would not accept (A) for sense-experience, as this would entail that the Sense-Datum Theory is the correct account of hallucinations and (arguably) illusions. This, in turn, generates an issue for Naïve Realism that Martin calls the 'screening-off problem' (Martin 2004). Also, notice that (T) is neutral towards the idea that mind-independent objects are all that we are aware of in perception-that is, it is a weaker claim than the so-called 'Transparency Thesis'. (T) should be read as a claim about what an acceptable answer would be if we were asked to describe our experience: according to (T), mentioning the mind-independent objects we see (and their features) would be a relevant way to address the question-even if, possibly (and this is what marks the difference with the Transparency Thesis), only a partial one. This is how we should read the claim that perception 'makes manifest' to us mind-independent objects and their features. For more on (T) see 3.1 of this paper. 14 Even among those who do not endorse Naïve Realism. See, for instance, Crane (2005). articulates such intuitive conception, while the other parties must reject at least one of its two fundamental features. To 'save' our intuitions on the nature of perception, in the last couple of decades some philosophers have denied that 'good' and 'bad' experiential cases should be explained in terms of a common fundamental state-thus embracing a 'disjunctive' account of sense-experience. 15 In this way, it is possible to claim that our intuitive conception of perception is correct at least when it comes to the veridical cases-that veridical perceptual experience possesses both (A) and (T).
Most of the recent debate focusses on whether a disjunctive account of senseexperience is defensible, and on the kind of account Disjunctivism could give of the non-veridical cases. This paper focusses on two more basic questions: what does it mean to say that (A) and (T) are part of our 'intuitive' view of perception? Is it true?
Section 2 focusses on what we should make of the idea that a theory of perception accommodates our 'intuitive' conception of perception, independently from what Naïve Realism takes to be part of such conception. Section 3.1 provides some considerations concerning the two main features of this (allegedly) 'intuitive' conception.

Common-sense and introspective evidence
In this section, I address question (1)-the question concerning what the advocates of Naïve Realism mean when they claim that it best captures our intuitions concerning the nature of perception, independently from the (alleged) specific features of such a conception.
When it comes to this question, there is a widespread oscillation between two at least prima facie different ideas. The first idea is that a naïve conception of perception corresponds to our 'pre-theoretical intuitions concerning perception or, in other words, to the view of the common man' (Fish 2009, p. 5). Martin, for instance, puts things in this way on different occasions: in (2002), he claims that subscribing to Naïve Realism allows us to avoid needing 'to posit some kind of error in our naïve or common sense judgments about perception ' (2002, p. 420); in (2006) that Naïve Realism 'correctly captures the common-sense conception of perception' (2006, p. 356) and that it is 'the best philosophical articulation of what we all pretheoretically accept concerning the nature of our sense-experience' (2006, p. 404). Crane expresses the same idea when he mentions that our 'pre-theoretical conception of perception' is that of 'a relation to a mind-independent object' (2006, p. 133). In fewer words, according to some commentators, Naïve Realism is the best articulation of our 'common-sensical' or 'pre-theoretical' judgments concerning the nature of perception.
The second idea is sometimes presented in association with the first, while some other times it just replaces it. It is the idea that Naïve Realism best articulates what is evident to us by reflecting on, or introspecting, a veridical perceptual experience. Martin endorses it multiple times, for instance in (2001), when he argues that 'reflection on experience gives equal support to both Transparency and Actualism' and that Intentionalism, therefore, denies 'introspective evidence' for (A) (2001, p. 219). He then presents it together with the first idea by claiming that Naïve Realism aims to take 'our introspection of experience at face value, thereby avoiding the need to posit some kind of error in our naïve or common sense judgments about perception ' (2002, p. 420). Finally, he insists that Naïve Realism 'best articulates how sensory experience seems to us to be just through reflection ' (2006, p. 354). Nudds also pursues this conception of 'naiveness' by arguing that 'when we introspect a visual experiential episode, it seems that we are related to some mind-independent object or feature that is present and is a part, or a constituent, of the experience' (2009, p. 334). Crane claims that evidence for the relational view can be found when we 'turn inwards' or 'reflect' on our experience (2005, p. 247).
What is the relationship between these two ideas? Is Naïve Realism 'naïve' in the sense that it articulates our 'common-sense judgments' about perception or our introspective knowledge of it? If we assume them to be different things, one possible interpretation is that Naïve Realism is taken to represent the best articulation of both. This is implausible for at least two reasons: (a) the two ideas are sometimes explicitly presented as connected claims (see, for instance, the passage mentioned above from Martin (2002)); (b) none of the authors above provide different sets of features Naïve Realism articulates, one for our pre-theoretical conception of experience and the other for our introspective knowledge of it: in both cases, the features Naïve Realism accommodates are (T) and (A).
It seems plausible, therefore, that by mentioning our pre-theoretical/commonsense judgments on perception and our reflective/introspective knowledge of it, these authors refer to the same thing. One way to accommodate this is to claim that whatever is 'common-sensical' or 'intuitive' to say about perception is exclusively a consequence of what we can know about it through introspection. This could make sense of the idea that Naïve Realism accommodates both, as one would consist in a collection of judgments grounded on our introspective access to perception.
This claim, however, is far from uncontentious. If we imagine asking someone (some 'common' person) about her beliefs concerning the nature of perception, we would probably get an answer along these lines: that perception is a way to obtain information about the external environment through our senses. It is not implausible that some of us could include some rough information about how the external world causes our experience: light being reflected by external objects and captured by the eyes, signals being transmitted by the nervous system to the brain and so on. 16 Could we get to this kind of answer exclusively by introspection? This seems very implausible, but I see no reason why such an answer should not count as 'intuitive' or 'common-sensical'. Of course, one can argue it is not a 'pre-theoretical' conception, as part of it depends on some (albeit vague and, possibly, even incorrect 17 ) notions concerning scientific studies on perception. This may be true, but how is it a problem? Common-sense, after all, is shaped by different sources, most importantly including the society in which we live-which, in turn, embeds a whole series of conditionings, also coming from the most diverse areas of knowledge. 18 It is not surprising if common-sense is partially shaped by a sort of vulgarization of highly 'theoretical' acquisitions. Yet, maybe Naïve Realism is supposed to capture precisely what we would say about perception if we were not influenced by this cluster of diverse conditionings. If this is the case, it would become difficult to understand in which sense such a conception would be 'common-sensical'. 19 Moreover, besides the obvious problem of identifying what precisely this 'naïve' conception would amount to, it would become unclear why accommodating it should work as a motivation for Naïve Realism: how does accommodating Tarzan's (alleged) conception of experience constitute a reason to endorse Naïve Realism?
We need to look for a different way of making sense of the connection between the two claims, besides them being simply distinct or one the direct consequence of the other. One option is to give a central role to the second claim rather than the first-that is, to the idea that Naïve Realism is (allegedly) the best way to account for our introspective knowledge of perception. If we focus on this claim, we can more easily connect it with the idea of this account being pre-theoretical: Naïve Realism is, allegedly, the best way to account for what introspection reveals to us about perception when it is not influenced by any further theoretical commitment. The underlying assumption, therefore, is that introspection on perception delivers for everyone, when performed correctly, the same results-so that via introspection we can reach a starting point everyone agrees on.
If this is the case, we have avoided a source of potential confusion: despite being called 'naïve', Naïve Realism does not refer to the conception the 'common person' has of perception. To assess what is naïve about Naïve Realism, therefore, it will be irrelevant to identify what is our 'common-sensical' conception of perception: we rather need to reflect on what appears to us when perceiving and find out whether it is the case that, introspectively, perception seems to possess the two features discussed in Sect. 1. 17 For example, there is evidence that many adults believe that the process of vision includes some 'emanations' of the eye towards the world-an idea which is close to the so-called 'extramission' theory of perception, but that is completely at odds with the current scientific orthodoxy (see Winer et al. 2002). 18 There is some interesting recent empirical research on the topic of determining our 'common' conception of perception, in particular concerning the causal picture mentioned above. For instance, Roberts et al. (2016Roberts et al. ( , 2020 have found evidence that this picture is not a 'conceptual truth' about perception (2016, p. 729), but rather a 'widely accepted empirical, or more generally theoretical, belief about perception', much like the belief that 'whales are mammals ' (2016, p. 740). This might point in the same direction as my claim above: that a causal condition would be widely accepted mainly because of external conditionings-thus being something that we learn, just as we learn that whales are mammalsrather than because of some direct phenomenological evidence. Thanks to an anonymous referee for directing me to the relevant literature. 19 After all, the same word 'common' refers to the idea of something that is shared between two or more people.
At this point, one may be tempted to reassign a role to the conception of the 'common person' by claiming that Naïve Realism best articulates what she would answer not to a general question about the nature of perception, but to the corresponding phenomenological one-something on the line of 'what it is like' to perceive. However, there is a widespread acceptance of the idea that phenomenological questions are rather sophisticated ones; questions that are not easy to understand, let alone to answer. 20 One may then legitimately wonder whether there could ever be a 'naïve' response to such a sophisticated question. But even granting that this is not an issue, the main claim of this section would be preserved: Naïve Realism should not be taken to simply articulate our 'common-sense' conception of perception, but rather the conception of experience we gain when we 'pretheoretically' introspect a veridical perceptual experience. This proposal does not change the fact that the only relevant assessment of the idea that Naïve Realism is 'naïve' is to be gained via reflection or introspection.
Before moving to the next section, I want to point out how this result helps to answer a legitimate question about the motivations for endorsing Naïve Realism. In fact, one may wonder why the idea that only Naïve Realism can take our intuitive conception of perception at face value is supposed to be an advantage of the view. After all, it is even desirable for our advanced investigations on the nature of experience to overtake-or at least to refine-our previously 'naïve' opinions on the matter. Following this line of reasoning, Hawthorne and Kovakovich argue that 'it is still worth wondering why vulgarity is to be celebrated', and that there are reasons to doubt that 'vulgar common sense has seen in advance how to handle various challenges and commitments ' (2006, p. 180).
If my interpretation is correct, Naïve Realism has little to do with 'vulgar common sense'-as it is 'naïve' in the sense that it takes at face value the phenomenology of veridical perceptual experience. Accommodating the phenomenology of experience is not the only reason to endorse a theory of perception. Other considerations could enter in the picture: one could be attracted by a theory's ability to explain the kind of authority perception has on our empirical beliefs, or to the way it accounts for our knowledge of the external world. However, it is a desirable feature of any theory of perception, and one to which many theories of perception often appeal: as Martin underlines, both Sense-Datum Theory and Intentionalism have appealed to phenomenological evidence to defend their positions against their rivals (2001, pp. 216-217).
But if we accept that Naïve Realism is right about how veridical perceptual experience seems to us-that it seems to present both (A) and (T)-then Sense-Datum Theory and Intentionalism must hold that introspection is misleading about the PC of veridical perceptual experience. This is not an easy consequence to digest, because it entails that introspection is not even reliable when it comes to 20 Fish believes, for instance, that the likely answers would be a mixture of 'phenomenological, epistemological and metaphysical considerations […] lacking any specificity ' (2009, p. 19). To take a step out of the analytic tradition, consider Husserl's reflections on the relationship between the 'natural' and the 'phenomenological' attitude (i.e., 2012, pp. 52-62). Valberg expresses a similar idea when he describes the attitude of 'being open to experience' as an already philosophical one (1992, p. 34).
investigating how things seem us to be, which sounds almost paradoxical (Martin 2002, p. 421). Naïve Realism would then be the only account of perception which accommodates the fact that 'what it is like' to perceive involves both (A) and (T), and that this is what explains why it introspectively seems to present them both. If true, this would constitute a far better motivation for endorsing Naïve Realism than the appeal to some vague claims about the opinions of the 'common man '. 21 3 Transparency, actualism and phenomenal immediacy In this section, I complete the clarificatory task of the paper by giving some further details about how we should understand (T) and (A). In this way, it will be possible to provide a conclusive answer to questions (1) and (2)-thus identifying what is supposed to be 'naïve' about Naïve Realism. In the rest of the section, I argue that (A) is not part of our introspective knowledge of perception, and I identify a feature-Immediacy (I)-that is more suitable to capture the intuitions behind (A). I then show that Intentionalism is not committed to rejecting (I), thus challenging the claim that only Naïve Realism can take our intuitive conception of perception at face value.

Clarifying transparency and actualism
Let us first briefly focus on (T), which raises less interpretative issues than (A). Given the conclusion of Sect. 2, the claim that (T) is part of our 'intuitive' view of perception amounts to saying that, when we reflect on a veridical perceptual experience, we find that mind-independent objects and their features are made manifest to us.
As I have mentioned (note 15), the 'transparency thesis' is usually taken to claim that we can describe the PC of perception exclusively in terms of what it makes manifest to us (Harman 1990, p. 39). Under this formulation, 'transparency' has been used to argue against those theories committed to the idea that perception presents some 'intrinsic', non-intentional, features, as there would not be any phenomenological evidence for them.
True or not, this is not how Naïve Realism construes (T) in this context (Hellie 2007, p. 269). Thus formulated, in fact, not even the Sense-Datum Theory would be committed to denying (T), being compatible with the claim that all we are aware of in perception are its objects (construed as sense-data) and their features. (T) rather captures the idea that when we introspect our experience it is not as though the putative mind-independent objects of perception disappear to make way to some other kind of entities: 21 Notice that, by embracing disjunctivism, Naïve Realism must claim that introspection is misleading about some of the 'bad' cases, which would seem to possess (A) and (T) without possibly possessing them both. Therefore, as Martin is ready to acknowledge, there would be no real escape from the conclusion that introspection is systematically misleading about the nature of at least some class of experiences (2002,421). I come back to this in the conclusion of the paper.
When my attention is directed out at the world, the lavender bush and its features occupy the centre of the stage […] when my attention is turned inwards instead to my experience, the bush is not replaced by some other entity belonging to the inner realm of the mind in contrast to the dilapidated street in which I live (Martin 2002, p. 380).
(T) is the claim that our introspective knowledge of perception presents it as directed towards the mind-independent world without intermediaries-that 'the character of one's experience involves in some sense, or is directed on, or of, the mind-independent objects and their features which we take to be around us in our environment' (Martin 2001, p. 217). If this phenomenological remark is correct, as I believe we should concede, Sense-Datum Theories must indeed deny that perception has (T), despite introspective evidence of the contrary. 22 We can now move to (A). According to (A), perception makes its objects manifest to us in such a way that it cannot be the case that they do not exist and are not perceived. As anticipated, this feature is more difficult to handle than (T), partly because the recent literature oscillates between three different formulations of it. 23 The first formulation is what I call the modal claim, according to which it introspectively seems to us that we could not be in the same experiential situation in the absence of the suitable objects of perception. This formulation can be found, for instance, in Nudds-who claims that 'your experience seems such that were the bowl of fruit not actually there, you could not be having this particular experience' (2013, p. 271)-and Martin-who believes that 'what was distinctive of my situation is that things could not be this way with me and no object be there at all' (2001, p. 220).
The second formulation is the constitutive claim, which says that it introspectively seems to one that the perceived mind-independent objects (and their features) are constituents of the experiential state. Nudds, for instance, argues that 'visual experiences seem to have the NR [Naïve Realist] property' (2009, p. 335), which he defines as 'the property of having some mind-independent object or feature as a constituent ' (2009, p. 334), and, more explicitly, that 'our experience […] seems to have mind-independent objects and features as constituents' (2013, p. 271). Martin claims that 'when one introspects one's veridical perception one recognises that this is a situation in which some mind-independent object is present and is a constituent of the experiential episode' (2004, p. 65).
Finally, the relational claim is that perception presents itself as a relation to mind-independent objects and their features. Nudds, for instance, claims that 'when 22 This does not mean there are no substantive issues in giving a complete account of (T). The most interesting one is that of specifying the relationship between perceiving mind-independent objects and their visible parts. While some philosophers (i.e. Bermudez (2000) and Campbell (2004)) seem satisfied with the idea that we simply perceive mind-independent objects in virtue of perceiving some of their parts, I believe this strategy eventually ends up being incompatible with (T). However, this issue does not concern the present paper. 23 That they try to capture the same feature of experience is evident from the fact that they are taken to play the same role-that is, to differentiate Naïve Realism from Intentionalism just as (T) differentiates it from Sense-Datum Theories.
we introspect a visual experiential episode, it seems we are related to some mindindependent object or features ' (2009, p. 334) and that 'our experience seems in this way to involve the mind-independent objects or features that the experience presents, and so our experience seems to be relational ' (2013, p. 271). Crane also believes that our 'pre-theoretical conception of perceptual experience is that of 'a relation to mind-independent objects ' (2006, p. 133). Bower claims that when perceiving 'we seem to encounter' (2019, p. 215) mind-independent objects, and that 'encountering' should be taken as a metaphor for being related with them (2019, p. 215). Finally, Langsam argues that 'reflection on the phenomenal character of experience […] suggests that an experience is a relation between a subject and an external object' (2017, p. 111). 24 We need to clarify the relationship between these formulations. Although they are obviously interconnected-Nudds, for instance, presents them as alternative ways of pointing out the same essential fact about Naïve Realism (2013, p. 271) and Martin does the same at least for the modal and the constitutive claim (2004, p. 65)it is not clear how they are interconnected. Hence, we need to understand under which circumstances the three formulations coincide, and which one best captures what the Naïve Realists believe to be evident by introspection on perception.
To answer this question, my strategy consists in considering the role (A) plays into the debate on perception. As we have seen, accepting that (A) is true for veridical perceptual experience is what marks the difference between Naïve Realism and Intentionalism. Therefore, whatever is (allegedly) evident by introspection must be something that Intentionalism is committed to denying. Although some specific considerations should be devoted to the constitutive claim, if we follow this line of reasoning it seems that the best option for Naïve Realism is to construe our introspective evidence of perception in terms of the modal claim, with the other formulations being different ways of making it intelligible.
First, consider the relational claim. According to the relational claim, perception introspectively seems to be a relation to mind-independent objects and their features. We need to ask under which circumstances this claim would be incompatible with Intentionalism-that is, in which circumstances Intentionalism is committed to denying that perception is relational as it (allegedly) introspectively seems. Intentionalism depicts perceptual experiences (veridical or not) as situations in which some mind-independent objects and their features are represented to a perceiver. This seems to make at least veridical perceptual experience a relational state: something is represented as being in a certain way to someone. Intentionalism is incompatible with a stronger claim-that is, that perception is fundamentally a kind of relation which entails the existence of its relata (Bower 2019, p. 212). 25 This is what some authors try to capture when they rely on the idea that the perceptual relation must be a 'real' (Bower 2019, p. 215) or 'genuine' (Crane 2006, p. 140;24 See also Hellie (2007, pp. 268-269). 25 Maybe every relation entails the existence of its relata, as this is part of the definition of the concept of 'relation'. This paragraph does not challenge this claim, but it makes explicit that this is the reason why Intentionalism is incompatible with the relational claim. See Genone (2016, p. 6) for a brief discussion of these matters. Langsam 2017, p. 111) one. Intentionalism must deny that perception is a relational state in this sense because it claims that the same kind of mental state can occur in the absence of the putative objects of perception, which-in some cases of misperception-can have a mere 'intentional' existence. 26 Therefore, if it is introspectively evident that perception is a genuine relation, then it is true that Intentionalism cannot take such evidence at face value. But claiming that perception is a genuine relation amounts to saying that it entails the existence of its relata-that is, that it is a state that could not occur in the absence of the relevant mindindependent objects. Hence, to take seriously the Naïve Realist's contention that Intentionalism cannot take our introspective knowledge of perception at face value, we need to construe (A) in terms of the modal claim.
One may be tempted to apply the same line of reasoning to the constitutive claim. The idea would be roughly the following: according to the constitutive claim, it introspectively seems that the perceived mind-independent objects and their features are constituents of the experiential episode. Intentionalism, however, does accept that the presence of the suitable objects of perception in the subject's environment marks the difference between a veridical perceptual experience and a corresponding hallucination, as this determines the accuracy of the representational content. Therefore, without further specifications on what it means to be a constituent of the perceptual state, it may seem that Intentionalism is not committed to denying that the perceived mind-independent objects are constituents of veridical perceptual experiences. Even in this case, however, Intentionalism cannot claim that these mind-independent objects are essential constituents of the experience-as one could be in the same fundamental kind of mental state regardless of the veridicality of the experience. But if what is introspectively evident of perception is that the perceived objects are essential constituents of the experience, then it is introspectively evident that one could not be in the same mental state unless these objects were presentwhich is once again the modal claim.
The problem with this line of reasoning is that Intentionalism typically characterizes the connection between perception (taken as a representative state) and the perceived mind-independent objects as a merely causal one. 27 But if the connection is merely causal, then it seems natural to take the suitable mindindependent objects to be distinct from the experience itself and, therefore, not literally constituents of it. This seems to make Intentionalism incompatible with the (alleged) introspective evidence of the constitutive claim-even when it is not understood in terms of the perceived objects being essential constituents of the experience. 28 In this situation, I believe there is something to grant both in favour of the argument and the objection. On the one hand, Naïve Realists do typically seem to have in mind the idea that mind-independent objects seem to be essential constituents, and not merely constituents, of veridical perceptual experiences. 29 Moreover, construing (A) in terms of the modal claim appears to be the best way to capture the most fundamental line of disagreement between the Naïve Realist's and the Intentionalist's view of the metaphysics of perception-as if the modal claim were introspectively evident, there would be no way for the Intentionalist to take such introspective evidence at face value. This might not be true if the perceived objects were not taken to be essential constituents of perception, as Intentionalism is arguably not committed to the causal picture above. 30 On the other hand, it seems correct to argue both that this causal picture is not compatible with the constitutive claim and that it is the most natural way to characterize Intentionalism-indeed it seems hard to identify a credible alternative. 31 Therefore, I will proceed as follows: I will develop my argument in Sect. 3.2 against the introspective evidence of (A) by focussing on the modal claim rather than on the constitutive (or the relational) one. I will then show how the same argument can be adapted to argue against the introspective evidence of the constitutive claim as well. This means that, if the argument is successful, it does not matter whether we construe (A) in terms of the constitutive claim rather than (or in addition to) the modal one, as Intentionalism could deny both while still taking our introspective knowledge of perception at face value. 32 We finally have a satisfactory answer to the question giving the title to this paper: what is so naïve about Naïve Realism? Naïve Realism can take our introspective knowledge of perception at face value. Introspection tells us that the objects of perception are mind-independent, and that perception is such that one could not be in the same experiential situation in the absence of these objects. Naïve Realism claims that perception is exactly how it seems introspectively. Moreover, since Naïve Realism is the only view according to which both (T) and (A) are true about perception, it follows that it is the only account that can take introspective evidence at face value.
In the next section, I will challenge the latter claim by arguing that (A) is a plausible explanation of what is introspectively evident of perception, but that it is not itself introspectively evident. 29 See, for instance, Nudds (2013, p. 271) and Martin (2002, p. 398). Pautz also reconstructs the Naïve Realist 'slogan' as the claim that 'objects are essential constituents of nonhallucinatory experiences' (Pautz 2010, p. 295). 30 That is, an Intentionalist could, at least in principle, rely on a different account of the connection between the experience and the perceived objects. For a defence of this claim see Pautz, who argues that such a picture is not entailed by 'common-factor' Intentionalism (Pautz 2010, p. 284). 31 For instance, see again Pautz, who believes that this picture, although not compulsory, is 'extremely natural' for Intentionalism (2010, pp. 284-285). 32 The same consideration applies to the relational claim. Therefore, this discussion on the best characterization of (A) should be taken as an attempt of clarifying the Naïve Realist's position, rather than a way to reduce (A) to a formulation that can be challenged as I will do in Sect. 3.2.

Immediacy
My strategy to challenge the claim that (A) is part of our introspective knowledge of perception follows two steps. First, I reconstruct how most of the philosophical literature on perception introduces (A)-that is, how the 'intuitive basis' (Crane 2005, p. 247) for (A) is usually identified. I then provide some reasons to take (A) as a possible explanation for this intuitive basis, rather than as being itself part of our introspective knowledge of perception.
Before entering into the details of this line of reasoning, I want to express a general worry connected with the claim that introspection recommends (A) for perception: can modal claims ever be part of our introspective awareness of experience? Introspection does tell one something about how one's experience isalthough it is debatable which kind of authority we should grant to such evidence. But is it plausible to claim that introspection also tells us something about how experience could not be? This is what Naïve Realism asks for: that it must be introspectively evident that a certain mental state could not occur in the absence of the relevant mind-independent objects.
A claim that I take to be less contentious is that introspection can play a role in motivating modal claims about experience. To make an example, consider Husserl's method of 'eidetic variation ' (2012, pp. 157-160)-that is, the method through which one is supposed to identify the essential features of some phenomenon. Roughly put, eidetic variation consists in using imagination to determine what is possible to change in a certain phenomenon, without it ceasing to being the kind of phenomenon it is-thus determining its essential features (and, therefore, the conditions under which such phenomenon could not occur). Such a conclusion is reached by making extensive use of the deliverances of introspection, but it is not reached solely on this basis: other assumptions are at work (first of all, the role played by imagination) and claims about the essential nature of phenomena are the result of all of them. Therefore, although I am not providing here an argument in support of the idea that modal claims about experience are never introspectable, 33 I believe we should be very cautious to move from the claim that introspection is relevant to formulating modal claims about experience to the stronger one that modal claims about experience can be themselves evident by introspection.
But let us forget about this general worry and take a closer look at how some philosophers defend the claim that (A) is part of our introspective knowledge of perception. Some authors move directly from the idea that perception seems to 'involve' mind-independent objects to the claim that they seem to present (A) (Nudds 2009, p. 334). Based on what we have seen above, this move is inadequate: Intentionalism accepts that perception 'involves' mind-independent objects-as it accepts (T)-but disagrees with Naïve Realism over how it involves them. Some other authors, however, make a further effort to explain what is intuitive about (A). Their strategy invariably consists in pointing out the difference between perception and other mental states which also seems to present (T), such as imagination or thought.
Consider a situation in which one is imagining the front door of one's house. The object of the act seems to be mind-independent (and, indeed, it is). However, it would not normally seem to me as though I was perceiving it. The idea, therefore, is that the difference between these mental states constitutes the intuitive basis for (A). Martin explicitly appeals to this line of reasoning, when he states that the difference between perceiving and 'imagining or remembering' is what 'recommends Actualism' to one (2001, p. 219); but we also find similar ideas, for instance, in Crane (2005, p. 247) and Bower (2019, p. 215).
How should we describe the phenomenological difference between perceiving an object and merely imagining or thinking about it? The philosophical literature on the matter is rich. Sturgeon, for instance, calls 'scene-immediacy' the feature that differentiates perception from other mental states, and he characterises it as follows: 'your visual experience will place a moving rock before the mind in a unique ''vivid'' way. The phenomenology will be as if the scene is made manifest to you ' (2000, p. 9). Valberg claims that when perceiving an object 'we are inclined to suppose it exists at the time when it is available ' (1992, pp. 21-22), which is not true for mere thought (1992, p. 20). Crane argues that 'when an object is experienced in perception, it is ''there '' or ''given'' or ''present to the mind'' in a way which is not in other mental states ' (2005, p. 247), and then defines this notion of 'perceptual presence' as follows: 'how the actual object of thought is at the moment I am thinking of it does not in any way constrain my thinking of it; but in the case of perception it does ' (2005, p. 247). Martin argues that 'sensory states involve a certain immediacy or apparent presence of an object which is simply not required in the case of pure thought ' (2002, pp. 387-388), while Millar refers to the same feature with the name of 'object-immediacy ' (2014, p. 240). Finally, to take again a step outside the analytic tradition, Husserl attributes to perception the distinctive feature of presenting its objects in 'flesh and blood ' (1907, pp. 12-13).
Let us call this feature of the phenomenology of perception Immediacy-from now on, (I). It should be already clear how difficult it is to specify (I) in comparison with (T): Sturgeon acknowledges that this feature is 'the most difficult to handle ' (2000, p. 12); Brewer that it is an idea that is 'extremely difficult' to make precise (2011, p. 2). According to Naïve Realism, we should simply take (I) to correspond to (A)that is, saying that (I) is introspectively evident is to say that it is introspectively evident that one could not be in the same situation unless the perceived objects were present. This is the claim I am challenging in this section. Now, (A) and (I) are not equivalent, but since the dispute concerns what is introspectively evident to one, this is not a reason to deny that what we mean when we assert that (I) is introspectively evident is just that (A) is evident. This makes it harder to offer proper arguments in favour or against the equivalence, but I will nonetheless try to specify why I find it suspicious.
My line of reasoning is based on identifying an asymmetry in the debate over what counts as taking introspective evidence at face value respectively for (T) and (A)/(I). As we have seen, Naïve Realism and Intentionalism both accept that (T) not only seems true of perception, but that it actually is true of it: the objects of perception really are mind-independent. But this does not mean that Sense-Datum Theories have no potential explanation for the fact that (T) seems true of perception. One possible explanation could be that we are not able to isolate the contribution given by the senses from how it is later 'interpreted'-so that we do not realise we perceive mind-dependent entities which are interpreted as mind-independent at a later stage. The problem is not that Sense-Datum Theories cannot explain why (T) seems to be true about perception, but that the explanations they give do not depend on (T) actually being true of perception.
Let us consider the parallel situation concerning (A). According to Naïve Realism, (A) is equivalent to (I), and it is true about perception because its metaphysical structure is such that one could not be in the same experiential situation unless the perceived objects were also present. If (A) is equivalent to (I), Intentionalism could not accept that it is true for perception, as it construes its metaphysical structure in such a way that one could be in the same situation in the absence of the putative perceived objects.
But let us bracket for a moment our judgment over whether (I) is equivalent to (A). We have said that the objects of perception seem to be 'present' to us in a uniquely 'vivid' way, as if they were 'there' in front of us-differently from what happens for imagination or mere thought. As some philosophers have pointed out (i.e. Martin 2002, pp. 388-389), Intentionalism can provide some plausible explanations for this feature. One possible strategy is to appeal to the idea that perceptual representations are unique kinds of representational relations, different from those that characterize other propositional attitudes. 34 An alternative could be to rely on the presence of a particular kind of content which is (allegedly) only present in perception-such as a non-conceptual content. Another option is to argue that the content of perception includes the representation of a causal connection between the perceived objects and the perceiving subject. 35 To draw the parallel with how Sense-Datum Theories account for the introspective evidence for (T), we need to say that these explanations are such that they explain why it seems that perception has (I) in a way that makes (I) actually false for perception. But this is not the case: the latter explanation, for instance, does intuitively accommodate the distinctive way in which the perceived objects are given, even though it considers both perceptual and imagistic states as representational. Hence, it seems that Intentionalism can account for introspective evidence in a way which would make such evidence actually true for perception, something that does not happen if we just assume that (A) and (I) are equivalent (as the same would not be true if (A) was introspectively evident). This looks to me to be a reason to be sceptical towards the equivalence between (I) and (A): if we test them separately, we find that Intentionalism seems compatible with the former, but not with the latter. 34 Notice that, in this case, an Intentionalist should provide a way of characterising the uniqueness of perceptual representations without recurring to (I)-as this would make the explanation of (I) circular. An alternative strategy would be to argue that not even Naïve Realism can provide a non-circular explanation of (I) (see Millar, 2014, for an argument in favour of this conclusion). 35 For this kind of proposal, see Searle (1983) and Millar (2014).

But if (A) and (I) represents different claims about what is introspectively evident
about perception, then it seems that the best way to depict their relationship is between an explanandum and its explanans. Crane claims that the 'immediacy' or 'vividness' of perception 'derives from the fact that the perceived objects and their properties are actually given to the perceiver when being perceived, and determine the nature of the experience' (Crane 2005, p. 247. Italics mine). In this case, it seems that the metaphysical structure of perception-construed as making (A) true for perception-is used to explain why experience seems to present (I). But if this is the case, then the claim that only Naïve Realism can take our introspective knowledge of perception at face value is not sustainable. Intentionalism, unlike Sense-Datum Theories, would not be committed to rejecting any 'intuitive' feature of perception, but just one possible explanation of it. 36 In turn, this entails that the real tiebreaker between Naïve Realism and Intentionalism is not whether they are compatible with introspective evidence-because they both are-but on which view can offer the best explanation for (I), which is an issue that remains very much open in the contemporary debate. 37

Conclusion
It is widely assumed that only Naïve Realism can depict perception in line with our 'intuitive' conception of it. This paper had two aims. The first was to clarify in which sense Naïve Realism is supposed to be 'naïve'. In Sect. 2, I argued that it is because Naïve Realism is supposed to be the only view which takes our introspective knowledge of perception at face value. In Sect. 3.1, I pointed out the two key features of such (alleged) introspective knowledge: that the objects of perception are mind-independent and that one could not be in a perceptual state unless the putative objects of perception existed. The second aim of the paper was to challenge that this latter claim is introspectively evident and to argue that also an Intentional account of perception can take introspection at face value. 36 In Sect. 3.1, I have raised the worry that the most typical versions of Intentionalism may not be compatible with (A) even when understood as the claim that the perceived mind-independent objects are merely constituents, and not essential constituents, of perception. It is easy to show how the argument above does not depend on the idea that (A) is best construed in terms of the modal claim, and that it can be adapted to challenge the alleged introspective evidence of the constitutive claim as well. The worry was that Intentionalism construes the metaphysical structure of perception in a way that makes (A) (taken as the constitutive claim) false for perception. Let suppose this is true. If we acknowledge that, on the contrary, Intentionalism is compatible with (I) actually being true of perception, we have a reason to question whether (I) and (A) (taken as the constitutive claim) are equivalent. Even in this case, it seems that taking the metaphysical structure of perception as satisfying the constitutive claim should be considered as a way to explain what is introspectively evident of perception-that is, (I)-rather than as being introspectively evident itself. 37 On the one hand, see  for some objections to Searle's (and Millar's) account of (I). On the other hand, Millar (2014) argues that the best account Naïve Realism can offer for (I) would be circular (and easily matched by a corresponding Intentionalist one), while Sturgeon argues that it has 'little more than a label for it ' (2000, p. 14). This conclusion has relevant consequences for the debate on perception, but also for the philosophy of mind more generally. As a theory of veridical perceptual experience only, Naïve Realism needs to embrace a form of disjunctivism about senseexperience-that is, it must claim that veridical and non-veridical cases are metaphysically different kinds of mental states. Naïve Realism, however, also typically accepts that some kinds of 'bad cases' (such as hallucinations) can be introspectively indistinguishable from corresponding veridical perceptual experiences (Martin 2002, p. 421). Hence, these experiences will also seem to present the distinctive features of perception-(T) and (A)-despite not possibly having them both. 38 In turn, this poses some severe restrictions to our knowledge of our own mental states.
The issue is that if both (A) and (T) are introspectively evident features of perception, Intentionalism and the Sense-Datum Theory are not in a better positionsince they respectively reject that (A) and (T) are true for perception. Therefore, both views would have to admit that introspection is systematically misleading about the nature of one's own experience-leaving them in a very weak position to attack Naïve Realism on the limitation it poses to the knowledge of our own mental states (Martin 2004, pp. 84-85). Two conclusions would follow from this: 1) that there is no way to avoid the claim that introspection is systematically misleading about the nature of (at least some kind of) sense-experiences; 2) that Naïve Realism is the only account which can accept introspection gets the nature of some of them right.
If my considerations in Sect. 3.2 are tenable, none of these conclusions are correct. Since perception does not introspectively seem to present (A), Intentionalism does not need to accept that introspection is systematically misleading about the nature of veridical cases-thus being, in this respect, at least at the same level as Naïve Realism. Naïve Realism tries to read too much into our introspective knowledge of perceptionwhich does not speak in favour of any specific account of perception which can accommodate (T) and (I)-misunderstanding evidence for (I) as evidence for a specific way (I) should be accounted for. But since Intentionalism is compatible with a common-factor reading of sense-experience, it also follows it can avoid the conclusion that introspection is systematically misleading about the nature of some 'bad-cases', such as hallucinations. This consequence is, I believe, particularly relevant for phenomenology, which traditionally embraces reflection on our own experience as its core method of investigation (Husserl 2012, p. 174).
This paper does not settle the debate on which is the best metaphysical account of perception: on the one hand, it does not engage with the problem of assessing which one can provide the best account of (I); on the other, it only focusses on the phenomenology of perception, while there are other relevant considerations to be made-such as, for instance, epistemological ones. I hope, however, that it can contribute to reviving attention to some aspects of the debate that, despite their centrality, have been overlooked in recent discussions of perception.