Naïve Realism Face to Face with the Time Lag Argument

Naïve realists traditionally reject the time lag argument by replying that we can be in a direct visual perceptual relation to temporally distant facts or objects. I first show that this answer entails that some visual perceptions—i.e., those that are direct relation between us and an external material object that has visually changed, or ceased to exist, during the time lag—should also count as illusions and hallucinations, respectively. I then examine the possible attempts by the naïve realist to tell such perceptions apart from illusions and hallucinations, and after showing the inadequacy of the answers relying on a mere counterfactual or causal criterion, I explain why the problem is solved by introducing a view of visual perception as temporally extended into the past of objects and, in particular, as consisting in the whole causal chain of events or states of affairs going from external material object x to subject S. But this solution is not immune from defects for the naïve realist. I show that this view of perception raises a number of significant concerns, hence leaving the issue of the time lag problem still open for naïve realism.


1.
Direct realism is the thesis that "in veridical cases we directly experience external material objects, without the mediation of either sense-data or adverbial contents" (Bonjour 2013). As Le Morvan (2004, p. 221) puts it: Direct Realists hold that perception is an immediate or direct awareness of mind-independent physical objects or events in the external world; in taking this awareness to be immediate or direct, Direct Realists deny that the perception of these physical objects or events requires a prior awareness of some Indeed, this is not entirely true. For, if when we are at t in a direct perceptual relation to an external object, we are at t in a relation to something that might have relevantly changed or even ceased to exist at t, then all our (direct and relational) perceptions are at risk to also count as illusions and hallucinations. And, indeed, those perceptions that are direct relations to things that have actually visually changed or ceased to exist, do count also as illusions or hallucinations, respectively.
Of course, the naïve realist can try a series of responses to this problem. Each of these responses can be seen as the proposal of a new view of what an illusion and a hallucination are aimed at excluding from illusions and hallucinations, respectively, the (direct and relational) perceptions of things that have simply visually changed or ceased to exist through the time lag. After explaining why it is not possible for the new view to rely on a mere counterfactual or causal criterion, I examine one way of arranging them that solves the problem. The basic idea consists in conceiving a (direct and relational) visual perception as temporally extended and, in particular, as consisting in the whole causal chain of events or states of affairs going from external material object x to subject S. By relying on this view of (direct and relational) visual perceptions, it is possible to solve the problem raised by the time lag argument.
However, this view of visual perceptions is not immune from defects and raises not insignificant concerns. In the second part of the paper, I critically examine those seeming to me the main of these concerns. If I am right, those supporting naïve realism or the disjunctivist position should not consider even the response to the time lag argument which is based upon this view of visual perceptions as completely satisfying. This means, in turn, that they should not take the threat coming from the time lag argument as totally thwarted yet.

2.
Consider first the traditional time lag argument against naïve realism. The argument was originally developed by Russell (1912Russell ( , 1948: […] It takes about eight minutes for the sun's light to reach us; thus, when we see the sun we are seeing the sun of eight minutes ago. So far as our sense-data afford evidence as to the physical sun they afford evidence as to the physical sun of eight minutes ago; if the physical sun had ceased to exist within the last eight minutes, that would make no difference to the sense-data which we call 'seeing the sun'. This affords a fresh illustration of the necessity of distinguishing between sense-data and physical objects. (Russell 1948, ch. 3) The argument generalises to all cases of visual perception, since the light from every object takes time to affect us in order for us to perceive the object. 2 We always visually perceive objects as they were in the past rather than as they are in the present time. But-the argument goes-whatever is in a direct perceptual relation with the subject perceiving it at t should be temporally located at t rather than in the past of t. 3 We should conclude that whatever it is that we are in a direct perception relation to while we perceive, if any, it is not the external objects, and naïve realism is false.
The time lag argument has seldom appeared particularly serious to direct realists. 4 For instance, Le Morvan (2004) says that the argument only points out how we directly perceive mind-independent material objects in the visual modality: given the laws of physics, we can only directly visually perceive them with a time lag, however minute. Many direct realists have similarly replied that we do directly perceive mind-independent material objects in the visual modality, provided we accept that we always perceive them in the past (see, e.g., Snowdon 1992). In short, the naïve realist can concede that: (1) Every direct and relational visual perception of x is a visual experience of x as it was in the past.
Nonetheless, change being ubiquitous, we must acknowledge that: (2) It is possible for a visual experience of x as it was in the past to be a visual experience of x as having a particular visible quality that x has ceased to have at the time at which the visual experience occurs.
(1) and (2) jointly entail that: (3) It is possible for a direct and relational visual perception of x occurring at t to be a visual experience of x as having a particular visible quality that x has ceased to have at t. 5 Similarly, we cannot ever exclude the possibility that x has ceased to exist within the time lag. In other words: (4) It is possible for a visual experience of x as it was in the past to occur at a time at which x has ceased to exist. 5 Of course, each visual experience of x is not instantaneous, but has a duration, at least because of the finite speed of sensory information-processes in our brains. I will continue to refer to the time at which a visual experience occurs as an instant t for simplicity's sake, but what I will always intend is that it has a duration, however short. This will not affect my line of thought. For example, what I mean here by (3) is that it is possible for a direct and relational visual perception of x occurring in a certain interval of time T to be a visual experience of x as having a particular visible quality that x no longer possesses at each instant t i that is included in T.
3 In Moran's (2019) terms, if you hold that perceiving x is being in a direct relation to x where x is a constituent part of the perception of x, then you seem to have no choice but adhering to the Existence Principle, according to which "It is not possible to see something that no longer exists". 4 Moran (2019) being a notable exception.

3
(1) and (4) jointly entail that: (5) It is possible for a direct and relational visual perception of x occurring at t that x has ceased to exist at t.
So, the naïve realist must embrace (3) and (5). Consider, however, what we normally take as a sufficient condition for illusion and hallucination to occur (e.g., Smith 2002, p. 23;p 191): (6) A visual experience of x at t counts as an illusion if x visually appears at t as having a particular visible quality that x has not at t.
(7) A visual experience of x at t counts as a hallucination if x at t is not there where it visually appears to be at t. 6 For example, visually experiencing a Christmas decoration as a dagger hanging in the air counts as an illusion; and visually experiencing a dagger hanging in the air in case there is actually nothing in front of us counts as a hallucination.
Now it seems that, given what we have said so far, the same visual experience will count as both a direct and relational perception and an illusion. In fact, from (3) and (6) it follows that: (8) A particular visual experience of x at t which (i) counts as a direct and relational perception, and (ii) is such that x visually appears at t as having a particular visible quality that x has ceased to have at t, counts as both a direct and relational perception and an illusion.
Similarly, it seems that the same visual experience will count as both a direct and relational perception and a hallucination. In fact, from (5) and (7) it follows that: 6 Normally, necessary or sufficient conditions for illusion and hallucination to occur do not mention time: but I assume we should agree that they implicitly do, so that we can reformulate classic sufficient conditions such as those offered by Smith (2002) as (6) and (7). Note that Smith's conditions are at better only sufficient because they do not cover veridical hallucinations. A visual experience of x at t counts as a veridical hallucination if (a) x at t is there where it visually appears to be at t, and (b) x would visually appear to be there also in case x were not there. Veridical hallucinations are commonly thought as scoring a point for a causalist theory of (visual) perception and against perceptual disjunctivism. However, disjunctivists can account for veridical hallucinations by saying that for a subject S to (visually) perceive x requires for x to look some way to S; a case in which x visually appears to be in a specific location also if x were not there is not a case in which x can be said to look some way to S, because what actually goes on would go on whether the object were present or not; therefore, this case is not a case of (visual) perception. See Snowdon (1980-81, p. 42) on this point.
(9) A particular visual experience of x at t which (i) counts as a direct and relational perception, and (ii) is such that x has ceased to exist at t, counts as both a direct and relational perception and a hallucination. 7

3.
No doubt that admitting (8) and (9) constitutes a serious problem for the naïve realist. For one thing, postulating sensory experiences that at the same time both fully are, and fully are not, direct perceptual relations to external material objects seems to be an inadmissible contradiction. Moreover, if we consider the disjunctivist position, it seems to be equally fatal to it that the ultimately disjunctive characterisation of a visual experience be inclusive rather than exclusive. This is so because disjunctivists are committed to denying that whatever fundamental kind of mental event occurs when one is veridically perceiving the world, that kind of event can occur also when one is having a hallucination (or, an illusion) (Martin 2006). Yet (9) entails that a particular occurrence of a visual experience can both be a hallucination and-in virtue of its also being a perception-fall under the fundamental kind which perceptions fall under. 8 Note that this does not mean denying that "there is nothing inconsistent in the idea that a single event (a veridical perceptual episode) can be an instance of two different kinds-veridical and hallucinatory" (Nudds 2013, p. 275). The latter claim can be true because, according to Martin, "whatever the most specific kind of mental event that is produced when having a causally matching hallucination, that same kind of event occurs when having a veridical perception" (Martin 2006, p. 369), while of course "no instance of the most specific kind of mental event that occurs when having a veridical perception occurs when having a (causally matching) hallucination" (Ivi, p. 361). Therefore, it is admissible for a disjunctivist to hold that a veridical perception also falls under the most specific kind of mental event that is produced when having a hallucination, provided that only for hallucinations is this their most fundamental kind. What a disjunctivist cannot hold is that a single event falling under both the most specific kind of mental event that occurs when having a veridical perception and the most specific kind of mental event that occurs when having a hallucination, has (also) the latter as its most fundamental kind. If an event counts as a perception, it cannot also count as a hallucination-and vice versa. On a stronger conception of disjunctivism that Byrne & Logue (2008) name "metaphysical disjunctivism," there is no common mental element to perceptions and hallucinations. Yet (9) clearly entails that a common element exists, and that metaphysical disjunctivism is false.
The naïve realist may try to organise a defense by rejecting (6) and (7) as proper sufficient conditions for an illusion and a hallucination to occur. But what is wrong in (6) and (7), and how should we modify them? One possibility is this: (10) A visual experience of x at t counts as an illusion if x visually appears at t as having a particular visible quality that x has not at t, unless x has had that quality in the past of t.
(11) A visual experience of x at t counts as a hallucination if x at t is not there where it visually appears to be at t, unless x has been there in the past of t.
Nonetheless this strategy seems doomed from the very beginning. Accepting (11) would require accepting it to be sufficient for a visual experience normally classified as a hallucination to occur in the right place for it to count as a perception. Imagine, for example, that a person living in New York happens to frequently hallucinate Julius Caesar. (11) entails that, if this person flies to Rome and waits for having her usual visual experience in Campus Martius, this Roman occurrence will count as a perception of Julius Caesar. This seems absurd.
We may wish to further amend (10) and (11) by introducing a counterfactual requirement: (12) A visual experience of x at t counts as an illusion if x visually appears at t as having a particular visible quality that x has not at t, unless (i) x has had that quality in the past of t, and (ii) x would not have visually appeared at t as having that quality, had x had not that quality in the past of t.
(13) A visual experience of x at t counts as a hallucination if x at t is not there where it visually appears to be at t, unless (i) x has been there in the past of t, and (ii) x would not have visually appeared as being there at t, had x not been there in the past of t.
Such reformulations accommodate the idea that directly and relationally perceiving x entails x (as well as that directly and relationally perceiving x as having certain properties entails x instantiating those properties), while the same is not true about hallucination and illusion respectively. Yet (12) and (13) are not completely satisfactory at a deeper analysis. In fact, it is possible to imagine the case of a person who only hallucinates Julius Caesar in Campus Martius in Rome, because she knows from history books that Julius Caesar lived in ancient Rome and frequented Campus Martius, and these (true) beliefs psychologically affect her so as to make her hallucinate Julius Caesar no other than in Campus Martius. What we need is not simple counterfactual dependence between the visual experience and the object, but the right kind of grounding for it.
So, we may decide to change (10) and (11) by directing attention to proper causal connection. The idea is that a visual experience of Julius Caesar in 2020 counts as a perception of Julius Caesar only if there is an adequate causal pathway connecting the real mind-independent Julius Caesar to the visual experience of Julius Caesar-just as a visual experience of a star in 2020 counts as a perception of that star only if there is an adequate causal pathway connecting the real mind-independent star to the visual experience of the star, no matter if that star is gone at the time at which the visual experience of it occurs. After all, if I could travel faster than light through the cosmic space and, say, get in just one year to a spot located 2065 light-years away from Earth, I could be directly and relationally visually aware of Julius Caesar in Campus Martius from that spot, no matter that Julius Caesar is dead long ago-as acknowledged by (1).
Here is an attempt: (14) A visual experience of x at t counts as an illusion if x visually appears at t as having a particular visible quality that x has not at t, unless (i) x has had that quality in the past of t, and (ii) x's having had that quality in the past of t is adequately causally responsible for the occurring of the visual experience at t as well as for x's visually appearing at t as having that quality.
(15) A visual experience of x at t counts as a hallucination if x at t is not there where it visually appears to be at t, unless (i) x has been there in the past of t, and (ii) x's being there in the past of t is adequately causally responsible for the occurring of the visual experience at t as well as for x's visually appearing at t as being there. Now, (15) has the merit of allowing a visual experience of a dead star to count only as a perception rather than also as a hallucination-as required by common sense. The argument seems no more a threat to naïve realism (provided that we are confident that we can easily distinguish among "adequate" and "non-adequate" ways for an object x to be causally responsible for the occurring of a visual experience as well as for that visual experience to have specific qualities).
But a different kind of problem arises for naïve realism if we accept (14) and (15). As a version of direct realism, naïve realism is, among other things, the rejection of indirect realism. Indirect realism is the thesis that in veridical cases we indirectly experience external material objects with the mediation of some directly perceived mind-dependent tertium quid like, for instance, sense-data. According to indirect realism, what makes the difference between veridical and delusive visual awarenesses is that veridical visual awarenesses are appropriately caused by some external mind-independent object located in the scene before the eyes. It is thanks to this causal connection that the visual experience can only be classified as a perception, and the external mind-independent object counts as an indirect object of experience. Johnston (2004) refers to this conception as "the Conjunctive Analysis" of seeing and opposes it to the Disjunctive Analysis. On the Conjunctive Analysis of seeing, there is a common mental element among perceptions and hallucinations, perceptions and hallucinations fall under the same fundamental psychological kind, and the difference among the two states is due to an indirect causal connection to some external mind-independent object: [On the Conjunctive Analysis of seeing] the act of awareness involved in seeing must simply be the common act of awareness "augmented" in a certain way, namely by being causally connected to external particulars. […] In seeing there is a single act of awareness whose direct objects are exhausted by what one could be aware of even if one were hallucinating. But the act of awareness has external particulars as its "indirect" objects in just this sense: it is appropriately caused by those external particulars. (Johnston 2004, p. 211) It seems that (14) and (15) rely on "the Conjunctive Analysis" of seeing. According to (14) and (15), the difference among veridical and delusive visual awarenesses of x is that the former are appropriately caused by x, while the latter are not. But having a mind-independent object as the appropriate cause cannot be the factor for distinguishing between perceptions and hallucinations a naïve realist makes appeal to. For according to the naïve realist, external mind-independent objects constitute perception rather than causing it. Remind that naïve realism takes perceptual experiences to be constitutively relational states whose relata are external mind-independent objects (e.g., Martin 1997Martin , 2006. This means that a state of visually perceiving x is constituted, among other things, by there being x in front of the subject's eyes. The perceptual state cannot be caused by the external object, because in this case it would be an effect induced by, and hence separate from, the object itself. This is why naïve realists as Snowdon (1990) and Johnston (2004) reject a causalist theory of perception: a causal relation among an object and an inner experience is not the kind of relation we can refer to if we want to claim that we are in a direct and immediate relation to external mind-independent objects when we perceive them. 9 Snowdon argues that the reasons for doubting that a causal theory of vision is correct are at least in part identical to the reasons for doubting that "there is a visual 9 Moran (2018) has proposed to incorporate elements of a causal theory of perception within a naïve realist framework. In particular, he has maintained that there are causal constraints that must be met if a hallucinatory experience is to occur that are never met in perceptual cases. This would solve our problem. Unfortunately, these causal constraints merely amount to the fact that "it lies in the nature of [hallucinations] to be generated by deviant or non-standard causation" (p. 375). The first problem is that standard and deviant causation are very vague notions, as Moran himself acknowledges. The second and more fundamental problem is that this way of characterising hallucinations presupposes characterising genuine perceptions by appealing to standard causation-a view that, in fact, Moran explicitly endorses. But, again, thinking of perceptions as (standardly) caused by the perceived external objects seems incompatible with thinking of them as constituted by those objects. non-world-involving experience common to both hallucinations and perceptions" (Snowdon 1990, p. 55), which is the core claim disjunctivism rejects; and as a disjunctivist he denies that the concept of seeing "is a causal concept with a separable experience required as the effect end" (Snowdon 1990, p. 61). 10 4.
Of course, disjunctivists agree that for a subject S to visually perceive object x, x must causally affect S some way. Only, they deny that we can characterise the perceptual visual experience as an effect of these causal processes. The idea, as said, is that the perceptual experience is constituted by x, among other things. So, it can be said to be also constituted by the whole causal chain going from x to S: The mental state of affairs, o's looking F to S, is not a state or event at the end of a causal chain of events initiated by o; it is, rather, a (larger-sized) event or state of affairs which itself consists in the whole chain of physical events (not merely events within S) by which o causally affects S. The experience is the complete state of affairs, o causally affecting S. The ultimate effect in this causal state of affairs -the state or event which lies at the end of the causal chain which starts with o -is something physical in S; but that ultimate effect is neither identical with not constitutive of the experience itself (Child 1992, p. 309).
On this basis, we could try to reformulate (14) and (15) as such: (16) A visual experience of x at t counts as an illusion if x visually appears at t as having a particular visible quality that x has not at t, unless (i) x has had that quality in the past of t, and (ii) the visual experience of x consists of a causal chain of events starting from x in the moment of the past of t when x had that quality, continuing and largely consisting in the travel of the light originally diffusely reflected or emitted by x, and ending with some neural events in the visual cortex of S at t. 10 Child (1992) has tried to show that the causal theory of vision and the disjunctivist position can coexist. In his view, the disjunctivist holds that "there is a single sort of characterization which can be applied to cases of vision and to cases of hallucination: in both, it looks to S as if something is F. But there is no common type of state of affairs, no common ingredient; rather, any case of its looking to S as if something is F will be characterizable, more fundamentally and specifically, either as a case of something's looking F to S, or else as a case of its merely being for S as if something looked F to him" (p. 300). This said, Child claims that there is nothing incoherent in a disjunctivist holding that object o causes o's looking F to S. Still, he says that "the effect in vision […] can be described in mental terms as an experienceo's looking F to S" (p. 308). So his proposal amounts to admitting that o causes the most specific kind of mental event that occurs when having a veridical perception. This being the case, however, o, as a cause of this kind of event, cannot be a constituent of it. Child is aware of this difficulty. He rebuts, first, that disjunctivists can accept to drop the idea that o constitutes the perception of o, because what really matters is the epistemological achievement that the most fundamental mental characterization of a perceptual experience is world-dependent in that it involves mentioning a worldly object as a cause. Secondly, Child says that-following David Lewis (1986)'s idea of piecemeal causation, where a whole can possibly be said to cause one of its later parts-we could perhaps make sense of the idea of an external object causing the relational state which it is a component of. One could be forgiven for remaining sceptical about the compatibility between the causal theory of vision and the disjunctivist position even after considering Child's attempt.
(17) A visual experience of x at t counts as a hallucination if x at t is not there where it visually appears to be at t, unless (i) x has been there in the past of t, and (ii) the visual experience of x consists of a causal chain of events starting from x in the moment of the past of t when x was there, continuing and largely consisting in the travel of the light originally diffusely reflected or emitted by x, and ending with some neural events in the visual cortex of S at t.
I assume that (16) and (17) do solve the time lag problem in a satisfying way. In particular, there seems to be no difficulty in having a direct perceptual relation with an object temporally located in the past.
For, since the perception of x by S consists in the whole causal chain going from x to S, it does not matter at all whether the causal chain is long or short, and how temporally distant from S x is. Inasmuch as there is a causal chain of the right kind going from x to S and consequently S perceives x, S necessarily is in a direct relation to x, simply because the perception that S is having is the causal chain, and the causal chain, as a whole, is obviously in a direct relation with its origin, i.e., x; and we can assume that if the perception that S is having consists in something that is in a direct relation to x, then also S is.
To see it another way, let us resort to how a direct realist like Snowdon (1992) puts it. First, Snowdon offers a definition of what exactly is to perceive something "directly": S directly perceives x if and only if S stands, in virtue of her perceptual experience, in such a relation to x that, if S could make demonstrative judgements, then it would be possible for S to make the true nondependent demonstrative judgement "That is x" (a dependent true demonstrative judgement being a demonstrative judgement that is true only in dependence on the truth of a more basic demonstrative judgement not depending on the first one). The thesis of direct realism can therefore be paraphrased as the thesis that "in ordinary perception we are so related to external objects that we can (nondependently) demonstratively pick them out" (Snowdon 1992, p. 61; see also Campbell 2002). In Snowdon's terms, the time lag argument threatens direct realism in holding that the external objects cannot be the things, if any, that our nondependent demonstrative judgements are true of. Here is the time lag argument opportunely reformulated: (18) That (the nondependently demonstratively graspable and visually presented thing, whatever it is) exists now.
(19) The dead star D, which would be the appropriately placed external object, does not exist now.
It follows that: (20) That is not identical with the dead star D.
The argument can be easily generalised to all external objects, because all of them are located in the past, however slightly, because of the finite speed of light, and might have ceased to exist at the time we visually experience them.
The solution that Snowdon provides to this difficulty consists in denying (18) and conceding that we can nondependently demonstratively grasp things into the past. We can accept as true: (21) That (the nondependently demonstratively graspable and visually presented thing, whatever it is) does not exist now.
This is the equivalent of conceding that we can directly perceive into the past. If we can nondependently demonstratively pick out things into the past in virtue of our perceptual experiences, then there is no difficulty in identifying these things with external objects like dead stars -what, in Snowdon's conception, is equivalent to saying that we can directly perceive external objects like dead stars.
In Snowdon's own words: It is possible to treat the truth of [ (21)] both as enabling us to avoid the time lag problem, and also as revealing that the dogma about the impossibility of truth for such sentences should not be accepted. Thus I am suggesting we should regard the finite speed of light as enabling us at t to [nondependently] demonstratively think about items which no longer exist at t. That is why a sentence like [(21)] could express a truth. (Snowdon 1992, p. 77) Now, if we interpret a perception by S of an object temporally located in the past as consisting in the whole causal chain starting from the object in the moment of the past of t when it was there, continuing and largely consisting in the travel of the light originally reflected or emitted by it, and ending with some neural events in the visual cortex of S at t, we have a metaphysical account of how (21) can be true. Accordingly, we can accept: (22) That (the nondependently demonstratively graspable and visually presented thing) is identical with the dead star D.
This amounts to say that we can directly visually perceive D.
If conceiving perceptions as consisting in the whole causal chains going from x to S solves the problem raised by the time lag argument, however, we should not take it for granted that such a conception is immune from troubles, or at least that no further philosophical work is needed to clarify some obscure points in it. In the remaining part of the paper, I will try to highlight some possibly problematic aspects of this view of perceptions.

5.
Perceptions are commonly taken to be mental states with a certain phenomenal character. Naïve realists do not normally distance themselves from this assumption.
The naïve realist's thesis is that we can visually perceive into the past, or better, we always visually perceive into the past; and, visual perceptions should be regarded as consisting in the causal chains going from the material objects in the past to us now.
But should we then regard the mental states that perceptions are as so temporally extended? Should we take the mental state corresponding to a perception of a dead star as something temporally extended from one hundred million years ago to now? Should we accordingly say that such a mental state supervenes not merely on a neural state, but on a causal chain temporally extended from one hundred million years ago to the present, only one of whose final components is a neural state of S? Is this the right way to account for the naïve realist's thesis that veridical perceptions are constitutively relational mental states, where the relation is to be intended as one between the perceiver and a worldly subject-matter?
There are many difficulties in such a view. For example, it seems to entail that necessarily my perception of the dead star started one hundred million years ago. But in which sense, then, can I account for my belief that I only started to perceive the dead star seven seconds ago, when I first looked in the telescope? As Moran (2019) notes, it seems absurd to hold that, when I look to the star for just a few seconds, it is literally true that my visual experience of the star goes on for millions of years. Also, it seems absurd that my own perceptual experiences begin long before I was born.
The naïve realist could try to arrange an answer by saying that, when we want to state that the perception started just seven seconds ago, we are referring to that part of the whole causal chain the perception consists in that happens within me. It is only this part of the perception that started seven seconds ago, while the whole perception started one hundred million years ago.
But if that is so, would not it be preferable to describe the situation by saying that perception are relational states extended into the past, and they are partly constituted-as by the final state of the causal chain they consist of-by a neural state of S on which the perceptual mental state of S supervenes?
This way, we would be differentiating perceptions (which would be temporally extended from the position in time of x to that of S) and the mental states they necessarily include (which would be just one component of the causal chain the perception consists in and would be as temporally extended as the neural event in S they supervene on).
But if this is the case, the naïve realist's fundamental idea that "no instance of the most specific kind of mental event that occurs when having a veridical perception occurs when having a (causally matching) hallucination" (Martin 2006, p. 361) is in danger, because only perceptions-not also their constitutive mental states-would now be essentially relational; and although it would still be possible to argue that also for mental states, as so characterised, external objects "figure non-causally as essential constituents of them" (Martin 2004, p. 56), this move may appear difficult to defend. As Papineau (2021, p. 19) puts it, these mental states being conscious sensory experiences starting when the subject starts experiencing and ending when the subject stops experiencing, events occurring outside this time interval cannot make a constitutive contribution to them, although they could of course make a causal contribution-but causation is not constitution. In the same spirit, Moran (2019) acknowledges that it seems plausible that, if an event takes place in a specific time interval, then its constituent parts must be temporally located in this time interval, too; perhaps naïve realists could try to deny that it is absurd for an experience to have a constituent that is not compresent to the experience itself, but it is not easy to imagine how this could be done, and the burden of proof is definitely on them. 11 More basically, considering perceptions as being no mental states-and, in addition, consisting in items even 99,99995% of which (as in the dead star case) had already occurred or happened before my birth-would frustrate naïve realism's ambition of being the least revisionary among all theories of perception (Martin 2002, p. 421).
Even if the naïve realist decides to go for the thesis that perceptions are mental states-and these mental states, in virtue of their being no other than perceptions, are temporally extended from the position in time of x to that of S-there is another problem to be solved. For, these mental states consisting of the whole causal chain going from x to the neural events in the visual cortex of S, they would include these neural events as well. We may see these mental states to supervene on the whole causal chain; in this case, they would supervene on a physical process partly constituted by a neural state in the visual cortex of S. But consider, now, the neural state in the visual cortex of S in isolation. It would be difficult to hold that it subvenes to no mental state. For, the same neural state (at least as to its intrinsic properties) does subvene to a mental state in a causally matching hallucinatory case. 12 It would be strange that, in the perceptual case, no mental state (irrespective of its being the same or not as to the hallucinatory case) does supervene on this neural state in isolation.
But if a mental state supervenes on the neural state in the visual cortex of S taken in isolation, we implausibly have two distinct mental states occurring in the veridical perceptual case, one supervening on the neural state of S and another supervening on the whole causal chain including, among other things, the neural state of S itself (and, arguably, its supervening counterpart, too). 11 Moran (2019, p. 222) says that it seems absurd to hold that, say, a violinist could be a constituent of a musical performance without existing at the same time as the performance; nonetheless, he concedes maybe ironically, "perhaps objects of perception are not constituents of experiences in the same way that violin players are constituents of performances". Gu (2021, p. 11243) tries to argue that when a constituent is a necessary rather than contingent component of the constituted, as the external object must be with regards to the act of perceiving it, then the constitutive relation is nontemporal, as one can deduce by examining the case of the father-and-son relationship, which "is such a constitutive relation" and in fact holds also if the father has passed away. There are many flaws that can be found in Gu's argument, but one is surely that the father-and-son relationship is not a relation between a constituent and a constituted, because an individual cannot be seen as constituted by his father. In fact, Gu says that "any actual pair of a father and a son constitute a special father-and-son relationship", thus confounding a particular father's constituting a particular instantiation of the father-and-son relationship with that particular father's constituting his son. In general, Gu provides no convincing argument for thinking that if something is a necessary constituent of an experience, then it can be non-existent for all the time at which that experience takes place. 12 Visual hallucinations are today thought to have neural correlates involving the activation of specialised functional units also serving normal visual perception (see ffytche (2013)). See Moran (2018) for an attempt to deny that hallucinatory experiences occur whenever the right kind of brain state is produced, based on the idea that they are essentially (and so necessarily) caused in a certain way and that these causal conditions cannot be met in perceptual cases. For an explanation of why Moran's attempt is unconvincing, see footnote no. 9.
It is no consolation that the nested mental state is not a perception itself. The problem, here, is that a visual perception is a mental state possibly mostly occurring at times when S is non-existent (or at least, mostly temporally located at times where S is not), and including in its supervenience base (the supervenience base of) another mental state of S.
Moreover, since the supervenience base of this latter mental state is the same, at least as to its intrinsic properties, as of a causally matching hallucinatory mental state of S (and notably, a mental state consisting in S' hallucinating x), we have the embarrassing view that all perceptions of x are mental states supervening on a set of states one of which is (the supervenience base of) another mental state belonging to the kind that is the most fundamental kind of a hallucination of x (by S), but which cannot be a perceptual mental state by definition. In short, all perceptual mental states do have another mental state subjectively indistinguishable from themselves nested in them and having a shorter duration. Needless to say, this would be a very disconcerting conclusion for the naïve realist -who would have to determine, in addition, which of the two states is the primary bearer of the phenomenal character of the perception.
Whatever is the strategy adopted by the naïve realist for setting the metaphysical relation among perceptions (regarded as consisting of the whole causal chains of events from x to S) and the corresponding mental states, it seems that many troubles appear.
True, the naïve realist might be tempted to identify both the perception of x and the corresponding mental state with the psychological state supervening on the neural state in the visual cortex of S. To avoid falling into the Conjunctive Analysis of seeing, however, it would be prohibited to characterise this state in causal terms: its being appropriately caused by external particulars, in fact, would amount to saying that it-as an act of awareness-has those external particulars as its indirect objects (Johnston 2004, p. 211). The naïve realist should rather try to say that external objects "figure non-causally as essential constituents of them" (Martin 2004, p. 56). But if the external object in question is a dead star, which is temporally located one hundred million years ago, it seems that the perceptual relation between S and the star would remain indirect inasmuch as this state is not temporally extended into the past. For, it seems that S could be only perceptually directly aware of something that is temporally located at least in part where its act of perceiving is temporally located. But the psychological state supervening on the neural state in the visual cortex of S is not temporally extended into the past enough to be in touch with the dead star. It falls short of the star. As previously said, it seems that whatever is to be a constituent of this psychological state, it must be temporally located in the time interval in which the psychological state occurs (Moran 2019;Papineau 2021), which in turn seems to coincide with the time interval in which the neural state in the visual cortex of S subvening to the psychological state occurs. But the star is not temporally located in this time interval. This is why the naïve realist needs to identify the perceptions of external objects temporally located in the past with the whole causal chains of events starting with these objects and ending in the visual cortex of S. But then this rules out the possibility to identify both the perception of x and the corresponding mental state with the psychological state supervening on the neural state in the visual cortex of S.

6.
On the view that perceptions consist of the whole causal chains going from x to me, another problematic point is that related to my being the subject of what is considered to be an experience of mine, and how this is supposed to be transmitted into the past.
Consider a causal chain of events C going from star D one hundred million years ago to my telescope now. Suppose that C is adequate for becoming a perception of D if only I look in the telescope. Nonetheless, of course, C is not a perceptual experience of mine until it is adequately "completed" by me. Let us call "CC" the causal chain "completed" by me, and "CI" the same causal chain that remains forever incomplete. What are, exactly, the differences between CC and CI? Sure, CC "ends" in a neural state in my visual cortex (indeed, it continues indefinitely also after this step: but we are interested in taking it as an "end" here), while CI goes on-let us suppose-without including a neural or psychological state ever. The external world is populated not just by material objects and their properties, but also by a plethora of potential (visual) perceptions that never get actual. But can we say that CC and CI have C in common?
From a physical standpoint, the answer is "yes." But in a metaphysical sense, C must be somehow changed by its becoming CC in a way in which C clearly is not when it becomes CI. For, when C is "transformed" into a perception of D by S, it is necessary that S becomes the subject not only of an event or a state which either is identical to or supervenes on {CC minus C}, but of an event or a state which either is identical to or supervenes on CC as a whole.
The reason is this: the perception of D by S, intended as consisting in the whole causal chain going from D to S, is supposed to put S in a direct relation to D. For this to happen, S-which is never temporally located where D is, and a fortiori is not so temporally located while is visually perceiving D-must be the subject of all of the parts of the temporally extended perception, included those that "touch" or even "include" D. If this does not happen and S remains the subject of an event or a state which either is identical to or supervenes on {CC minus C} only-that is, S remains the subject of the sole part of the perception occurring within S-then we can hardly say that the whole perception is a perception of S. Either the events or states belonging to C cannot belong to what an act of S' perceiving D consists in, or-at best-S would be in a direct contact only with anything "touched" by or included in the proper parts of (the supervenience base of) the perception that S is a subject of, and in an indirect contact with the rest of the things "touched" or included in (the supervenience base of) the perception. Either way, regarding perceptions as consisting of the whole causal chain going from x to S would not vindicate naïve realism.
The naïve realist needs the property of 'being something that S is the subject of' to be transmitted backwards along the temporal parts of the perception that correspond to the events or states being the components of the causal chain that by hypothesis the perception as a whole consists in, until D, in virtue of the mere C's becoming CC. And not only that: the transmission should be instantaneous, otherwise the relation between S and D would not be direct at least for part of the duration of the (alleged) perception (indeed, the perception could become direct at a time at which it has already ceased to occur. But I assume this to sound weird to many of us).
It is not clear, however, how the mere occurring now of some neural events in my visual cortex and, arguably, of its supervening counterpart, can instantaneously and effectively expand my 'being the subject of this' through millions of years. Consider that the events or states of affairs that are the supervenience bases of the temporal parts of the perception that must be so affected have already fully occurred subjectlessly along the causal chain.
Of course, we should not presuppose that the transmission of the property 'being something that S is the subject of' is tantamount to the transmission of a physical perturbation. As previously said, from a physical standpoint C is one and the same in CC and CI.
Nor should we put in question the bare fact that a mere physical event or state (such as, for example, a neural state under a physical description) can be the supervenience base of a higher-level state instantiating the property of 'being something that I am the subject of'. Indeed, we should pay attention not to deny that also a physical state occurring outside my body (like, for example, a physical state which my computer is in) can be part of the supervenience base of a higher-level state of which I am the subject (according to the extended mind thesis; see Clark and Chalmers 1998). But in the latter case it is not difficult to tell a story according to which the physical or functional state which my computer is in enters into a relationship with one or more physical states which are either identical to or the supervenience base of states that I am the subject of, and in virtue of this becomes a proper part of a larger-sized physical state that is either identical to or the supervenience base of an extended state I am the subject of. In any case, it seems that all the parts of the so-conceived extended experience-included those consisting of, or corresponding to, the physical states occurring outside my body-must instantiate the property of 'being something that I am the subject of'.
In the case of perceptions as temporally extended into the past of objects, however, it is not at all clear how the equivalent story could be told. How could the part of my perception of the star that occurs outside me one hundred million years ago instantiate the property of "being something that I am the subject of" now? And in virtue of what? The naïve realist must claim that, while whatever is identical to or supervenes on C before its "completion" is subjectless, C instantaneously becomes the base of a state which I am the subject of when it becomes a part of CC, that is, when it is "completed" into CC. But how can this be accounted for? 13 13 One anonymous reviewer objected that there is nothing particularly odd in an item instantaneously gaining a property, because items don't always gain properties by causal means. Suppose you are the only object in the world with a certain set of intrinsic properties. You have the property of being the unique bearer of those properties. But then God creates an intrinsic duplicate of you. You instantaneously lose the property of being the unique bearer of those properties, and instantaneously gain its negative counterpart (the property of not being the unique bearer of those properties). We can modify the story so that a property is instantaneously transmitted back in time. Suppose that it is true at t 1 that you are the only thing in the universe with F, then at t 2 a new F is created, and hence it is no longer true-on certain views of time-even at t 1 that you are uniquely F. I concede that there is nothing wrong in these examples. In the case at issue, however, the difficulty comes from the particularity of the property that is supposed to travel instantaneously back in time, that is, the property of 'being something that I am the subject of'. One could be forgiven for being reluctant to accept that an hourlong event which occurred subjectlessly one hundred million years ago and has never ceased to have occurred subjectlessly until present day, can become now something I am the subject of.
Perhaps naïve realists should not be requested to come with a complete explanation of what it is to be the subject of an experience, since this is an open question for many if not all theories of perception. But unlike the others, their position requires that whatever it is to be the subject of an experience, it can swim upstream instantaneously through time. We can pretend that they take responsibility for exploring the issue of how such an immediate and instantaneous colonisation is supposed to occur.

7.
According to Martin (2002, p. 421), the disjunctivist position scores a point against the rival theories of perception in that it does not convict common sense of any fundamental error about perception. Martin says that "the idea that introspection will lead us into error about how things seem to us is hardly an attractive one," and notes that "in contrast to the kind of global errors in introspection posited by sensedatum theories and intentional accounts, the disjunctivist can claim that veridical perceptual experiences are exactly as they seem to us to be: states in which parts of how the world is are manifest to us." Sure, also the disjunctivist position is forced to concede that we may be in error when we have a hallucination or an illusion, since we may be unable to distinguish these experiences from a perception. But if we come to perceptions, the disjunctivist position is not an "error-theory" like the other main theories of perception (the term 'error-theory' originating with J.L. Mackie's views of secondary qualities and of moral properties). 14 If being not revisionary is a point in favor of the disjunctivist position and naïve realism, which the disjunctivist position is supposed to vindicate, it should be considered as a point against both these theories that they posit that we do always perceive things into the past and our perceptions are temporally extended. For, it is true that under these theories what we actually perceive are the worldly material objects, just like it seems to us; but it is also true that these objects are temporally located in the past-sometimes in a very far away past-, while they seem to us to be temporally located in the present (sense-datum theories and intentional accounts being, by contrast, error-theories with respect to the kind of things that are presented to us, but not to their temporal positions). 15 It is perhaps worth adding that under naïve realism or disjunctivism, while no veridical perception is a perception of something being temporally located exactly where it seems to me to be (the light speed being finite), there might be at the same time many different potential perceptions 16 of the same object each of which only differs from the others as for the temporal distance between the object and me. This is possible because it seems that seeing an object through one or more plain mirrors rather than without them should not preempt the visual experience from being a genuine direct perceptual relation with that object, especially so if the object appears as spatially located where it really is. Taking a different position would mean for the 14 See Mackie (1975Mackie ( , 1977. naïve realist to be forced to admit that seeing a star in a reflecting telescope (i.e., one that uses mirrors to gather and focus incoming light) is no directly relational perception: and I assume that no naïve realist would accept such a conclusion.
Indeed, it seems reasonable to agree that introducing one or more plain mirrors along the light travel from x to S should not preempt the visual experience of x by S from being a genuine directly relational perception of x by S, provided that these two important requirements continue to hold: Light Identity condition: the light hitting the foveas of S and causing the subsequent adequate neural events in the visual cortex of S at t is the same light that was diffusely reflected or emitted by x at the time in the past of t when x originated the causal chain ending with the neural events in S at t. 17 Spatial Location Identity condition: the spatial location at which x appears to be to S at t is the same where x was at the time in the past of t when x originated the causal chain ending with the neural events subvening to that visual appearance in S at t. Now, a skillful use of a relevant number of plain mirrors can delay at will the arrival of light from the object to my foveas. 18 As a result, we have the possibility to create a huge number of (potential) direct and relational perceptions of the same object x, all making x appearing as temporally located in the present, each differing from the others as to the real temporal location of x.
Ironically, then, we seem to be able-at least in principle-to genuinely directly and relationally visually perceive at t an external material object in an enormous number of temporal positions 19 except the temporal position it seems to us to be in (that is, the very same temporal position we occupy while we are perceiving it). This is not the same as for the spatial position of an object. Also if we concede that in a hall of mirrors like the one we can see in the climactic shootout of the film The Lady from Shanghai (1947) all the visual appearances of Rita Hayworth had by Orson Welles are genuine direct and relational perceptions of Rita Hayworth, nonetheless at least one of them-i.e., the one that does not involve the presence of any mirror along the light travel from Rita Hayworth to Orson Welles-is such that she appears to be spatially located where she really is. Indeed, also some 17 In spite of the photons coming out of a transparent glass being not the same that went into it, I assume that we all would agree that the light is the same. This appears to me as a reasonable assumption about the criteria for light identity. 18 Imagine that light coming from a star located one hundred million light years away from us that enters the Earth atmosphere without being "completed" into a visual perception by any subject S; some of it passes just a few miles from a mountain chain, leaves again terrestrial atmosphere and travels another one hundred million light years until a mirror reflects it back towards Earth, when-after entering and leaving our atmosphere a second time-a second mirror placed on an artificial satellite finally reflects it into our foveas, making us having a visual appearance of the star as located in the same place (or, at least, in the same direction) as if we had had a visual appearance of it two hundred million years before. 19 In case time is infinitely divisible, this number is obviously infinite. If time is not infinitely divisible, this number is as high as the number of different temporal positions the object has occupied in the past, provided that the object is very close to our eyes. In general, the object can be genuinely directly and relationally visually perceived now as occupying each of the temporal positions it has occupied in the past that is more distant in time from me than the time taken by the most direct light travel starting from the spatial position the object occupied when it occupied that temporal position and ending in the spatial position we are occupying now. mirror-involving appearances may have the very same property. But if we come to the temporal position of a material object, no perception of it is such that it appears to be temporally located where it really is. It is like we were in a hall of mirrors where, if we get rid one after another of all the perceptions of x such that x appears to be temporally located where it is not (as Orson Welles does in the movie with all his perceptions involving an incorrect appearance of the spatial location of Rita Hayworth), unlike Orson Welles we are left with nothing.
That being the case, we may start to doubt that naïve realists and disjunctivists have any right to flaunt the fact that they alone "can claim that veridical perceptual experiences are exactly as they seem to us to be," or even to use the expression "naïve realism" to refer to their own positions. As much as they might be the least revisionary positions among all the possible responses to the problem of perception, they are nonetheless revisionary to some degree. 20 8.
One final point concerns the relation between naïve realism and the metaphysical conceptions of time. The very idea that we can directly and relationally perceive external objects in the past, and that there can be a direct relation between the perceiver and a merely past worldly subject-matter, is only acceptable if presentism, as a thesis about the reality of time, is rejected (Power 2013). According to presentism, only what is present is real (Sider 1999;Crisp 2004). But if only present items exist, and by contrast there are no such things as merely past and merely future items, then a necessary condition for an experience to take something real as its object is for it to take something present as its object. Since a defining feature of a perception is that it takes something real as its object, it follows from presentism that it is impossible to perceive into the past.
Even if we want to leave room for someone to argue that, under presentism, we can indirectly perceive into the past, the principle that presentism rules out the possibility of direct relational perceptions of merely past things seems unassailable. Consider the view of perceptions as the whole causal chains of events going from the objects which existed in the past to S now. If presentism is true, then it is not possible for such a causal chain to be wholly real at one time: only the state or the event in the chain that is presently occurring is real, all the others being presently non-existent. Thus, either a perception is something that occurs wholly now-and in this case, we cannot identify it with a temporally extended causal chain-or we must concede that the merely past things are necessarily non-existent now while S is having a visual appearance of them. In either case, then, taking a perception of a merely past thing by S as a temporally extended causal chain going from the thing to S cannot ground the fact that S is able to perceive that thing directly and relationally (we are reasonably assuming it to be impossible for someone to be in a direct relation with something which is not real).
We must consider, however, that presentism is not the only position on offer in the current metaphysical debate over the philosophy of time. Most important, some researchers argue that, irrespective of how strongly our pre-scientific intuitions may recommend it, presentism is incompatible with the relativity theory, which is widely accepted today (Savitt 2000;Sider 1999Sider , 2001Wüthrich 2013). Other scholars have stressed that presentism can be criticised under many other points of view, and in particular can be accused to get in trouble when it comes to account for the different kinds of true propositions we endorse, or, for cross-temporal relations and the passage of time itself (Ingram and Tallant 2018). Under eternalism, which is the position opposing presentism and holding that also past and future things exist, the time lag argument is commonly seen to lose all of its force against realism.
Take, for example, Houts' point against (1). According to Houts, (1) should not be considered an overwhelming answer to the time lag argument. The problem with (1) is that the past things I am supposedly having a visual perception of are presently non-existent; and, presently non-existent things cannot be at any spatial distance to presently existent things. Since I am (or, if you prefer, my body is) a presently existent thing, then (1) implies that the things I perceive cannot be at any spatial distance from me. But this is absurd, because it entails that they cannot even be external to me. What does perception amount to if it is no longer perception of external things?
Moreover, since most arrays of things I am supposedly having a visual perception of are at different times, they cannot even be spatially related to each other. This means that (1) undermines any perception of three-dimensional space (Power 2013). Yet, this is only true if we assume that presentism is true. By contrast, if we assume that presentism is false, these objections no longer stand, and the time lag argument seems rather inoffensive against realism-and indeed, also against naïve realism, provided that one agrees to see perceptions as temporally extended (a conception that, as I have showed, has some independent defects).
What I want to highlight, however, is that naïve realism can only make the time lag argument inoffensive if it takes a specific position in the debate over the metaphysical nature of time. It has not been stressed enough so far that naïve realism must take a specific stand about the metaphysics of time in order to strive for immunity from the time lag argument. 21 Presentism, in particular, is lethal to naïve realism, because it is lethal to the naïve realism's defensive strategy against the time lag argument. Indeed, this is just one of a number of interesting ways in which the answers to the most important metaphysical questions about time are intertwined with those to the most important metaphysical questions about perception (Le Poidevin 2015).
We should add that presentism has been traditionally considered by both presentists and non-presentists as the most intuitive position (e.g., Sider, 2001;Markosian 2004). This being the case, we have another reason for refusing the title of "naïve realism" to naïve realism, and to believe it to be more revisionary a position than it may seem at a first impression (especially if we consider, as remarked by Moran 21 Again, with the notable exception of Moran (2019).
(2019), that the various rival 'conjunctive' theories of perception do seem to be able to adopt the common sense presentist view). True, recently some scholars (e.g., Torrengo 2017) have claimed that it is not true that common-sense favors presentism over eternalism, and that certain intuitions we all share that do seem to be supportive of presentism are actually either neutral or counterbalanced by rival shared intuitions supporting eternalism. But even if this were the case, presentism would remain a position whose intuitiveness is second to none. The fact that naïve realism and the disjunctive position must embrace non-presentism means, at worst, that they must reject the most intuitive position in time metaphysics and, at best, that they must embrace a position in time metaphysics that cannot be defended as the most natural and intuitive. In either case, there are further reasons to doubt that the overall intuitiveness of naïve realism and of the disjunctive position is as high as their supporters pretend. Indeed, we should not forget that they intensely appeal to intuitiveness in order to defend their positions. Such defense is to be thought as partly undermined by the necessity to embrace non-presentism.

Conclusion
I have tried to show that the time lag argument hits naïve realism and the disjunctivist position more significantly than it has been thought so far.
Both those supporting naïve realism and those opposing it have generally considered that a very simple answer by the naïve realist, i.e., that "we can have a direct perceptual relation with ordinary objects in the past," could settle things once and for all.
I have argued that this is not the case. First, I have shown that even if we can have a direct perceptual relation with things in the past, we must solve the problem that perceptions being direct relations to things that have visually changed or ceased to exist during the time lag seem to count also as illusions or hallucinations, respectively. To avoid these experiences to count both as perceptions and non-perceptions, the naïve realist needs to improve the sufficient conditions for an illusion and a hallucination to occur in order for them to exclude the perception of the Sun or that of a dead star. Nonetheless, the new formulations cannot rely on a mere counterfactual or causal criterion. Characterising the perception of a dead star as adequately caused by the star, in particular, is not an option for the naïve realist, because it would mean to rely on "the Conjunctive Analysis" of seeing -a view of visual perception that the naïve realists themselves regard as denying both the directness of the perceptual relation and "the Disjunctive Analysis" of seeing on which the disjunctivist position is based.
The problem can be solved if the naïve realist agrees to consider (direct and relational) visual perceptions as temporally extended into the past and, in particular, as consisting in the whole causal chain of events or states of affairs going from an external object x to a subject S. Indeed, many naïve realists have accepted to see (direct and relational) visual perceptions this way. Again, both parties seem to have considered this answer to be the last word regarding the challenge launched by the time lag argument to naïve realism.
I have tried to argue, however, that this conception of visual perception can be called into question. I have examined some problems arising from it. A first problem concerns the metaphysical relation between the so conceived perceptions and the corresponding mental states. A second problem has to do with the explanation of how my being the subject of my visual experiences is supposed to be transmitted into the past in the perceptual case. A third problem arises when the idea that "the disjunctivist can claim that veridical perceptual experiences are exactly as they seem to us to be" is set against the fact that, among the limitless temporal positions that a material object can occupy when we visually perceive it, none can ever be the one it seems to us to be in. Finally, a fourth problem regards the fact that so-called "naïve realism" can only stand if it takes a specific position in the debate over the metaphysical nature of timeindeed, one that many have considered as less intuitive than its rival.
Since the only view of visual perception that we know can solve the problem of the existence of some experiences counting as both perceptions and non-perceptions cannot be considered as completely satisfying, and on the contrary seems to raise a number of significant concerns, I believe the conclusion to be this: the question concerning the challenge launched by the time lag argument to naïve realism is not settled yet, and-contrary to common view-the threat represented by the time lag argument to naïve realism and the disjunctivist position is still standing.
Funding Open access funding provided by Università degli Studi di Sassari within the CRUI-CARE Agreement.

Conflict of Interest The author declares no competing interests.
This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.