Previous versions of this paper were presented at the 2021 Copenhagen Summer School in Phenomenology and Philosophy of Mind, University of Copenhagen, Denmark, and at the 2021 San Raffaele School of Philosophy, entitled Mind, Language, and the First-Person Perspective, Vita-Salute San Raffaele University, Cesano Maderno, Italy. I am grateful to the participants at both events for helpful discussion.

1. Prelude

1Towards the end of his early book on Husserl, Speech and Phenomena, Derrida makes a baffling pronouncement: “There never was any ‘perception’” (Derrida, 1973, p. 103). What might he have meant by this? The next, and final, page of the book contains a clue: “[C]ontrary to what phenomenology – which is always phenomenology of perception – has tried to make us believe, … the thing itself always escapes” (Derrida, 1973, p. 104). Derrida’s point seems to be something like the following. First, perception is supposed to be the brute or pure presentation of an object to consciousness. Second, Husserl’s phenomenology unwittingly, and contrary to Husserl’s official announcements, shows that there is no such pure presentation – presentation is always impure, since it always encompasses the representation of something not presented, something absent (cf. Derrida, 1973, p. 103). Therefore, finally, Husserl’s phenomenology teaches us – unbeknownst to Husserl himself – that perception is a fiction.

2Interpreting Derrida is a tricky business, and I don’t want to pretend that this thumbnail sketch does his text justice. But whatever Derrida might have meant, his pronouncement, and the context surrounding it, suggests a view some possible philosopher – call him Derrida* – might take. Derrida*’s view is one way of negotiating the following inconsistent triad:

1. Perception occurs.

2. Perception is the pure presentation of objects, property instances, events (etc.) in the world.

3. There is no pure presentation of objects, property instances, events (etc.) in the world.

3Derrida*’s response to this triad – abandoning (1) – is singularly implausible. To see, hear, and smell is to perceive. Of course people see, hear, and so on. So of course perception occurs all the time. That leaves two other responses to the inconsistent triad. Husserl’s own view is that (2) has to be rejected. He is very explicit about this. In Analyses of Active and Passive Synthesis, for example Husserl insists that perception “is only possible in the form of an actually and genuinely original conscious-having of sides and a co-conscious-having of other sides that are precisely not originally there” (Husserl, 2001, p. 40). Husserl’s point is that a perceptual experience is never pure presentation: it always includes a “co-consciousness” of “sides” or aspects of objects, which are not given or presented, strictly speaking.

5I don’t think The Pure View is very plausible. As I shall suggest in this paper, The Pure View seems unable to offer a convincing account of the sorts of phenomena that led Husserl to reject (2): putative cases in which perceptual experience goes beyond what is presented, anticipating something that is not there in the stimulus. Specifically, I will focus on cases of what the Belgian vision scientist (and phenomenologist) Albert Michotte dubbed “amodal completion”.

6Defenders of The Pure View – “Purists”, as I shall call them – must show that alleged examples of impure perception are merely apparent. They must argue that the impurity, if there is any such, is not due to perception, strictly speaking, but to some other mental state or episode. I will consider three Purist strategies for explaining away amodal completion in this fashion, and suggest that none is very promising.

7The structure of the paper is the following. In the next section, I introduce The Pure View. In section 3, I present three Purist strategies for handling potential counterexamples. Section 4 argues that none of those strategies yields a convincing account of amodal completion. Section 5 addresses an objection and summarizes the argument of the paper.

2. The Pure View

8According to The Pure View, perceptual experiences never make any “claims”, or have any commitments about the environment that go beyond what is strictly presented or given in the experiences themselves. Indeed, some Purists insist that there are simply no commitments in perception to things being one way or another in the environment. In Charles Travis’ words, “perception, as such, simply places our surroundings in view; affords us awareness of them. There is no commitment to their being one way or another” (Travis, 2004, p. 65). Perceptual experiences, then, are in a sense “silent” or – in Austin’s phrase – “dumb”.

9But as I understand The Pure View here, it need not be committed to Travis’ “silence of the senses” view. In particular, a Purist can maintain that a perceptual experience can be committed to things “being one way or another” as long as their being that way is given or presented in the experience without remainder, so that there is no possibility of things not being that way.

10What no Purist can accept is the idea that perception has commitments that exceed what is given or presented. For if perception carried commitments to things being this or that way that exceeded what was presented in perception, the possibility would arise that those commitments were erroneous. And that possibility is anathema to Purists of any stripe. For Purists, there is no such thing as erroneous or non-veridical perception, strictly speaking. Thus, for Travis: “Perception cannot present things as other than they are. It cannot present some ways things are not as what is so” (Travis, 2004, p. 65). Bill Brewer concurs:

[I]n perceptual experience, a person is simply presented with the actual constituents of the physical world themselves. Any errors in her world-view which result are products of the subject’s responses to the experience, however automatic, natural or understandable in retrospect these responses might be. Error, strictly speaking, given how the world actually is, is never an essential feature of experience itself. (Brewer, 2006, p. 169)

11James Genone, another Purist, denies “that perceptual error is so much as possible” (Genone, 2014, p. 344).

do not have built-in conditions of satisfaction, nor do they possess veridical [sic!] content that is accurate or inaccurate, true or false. If basic perceptual experiences – even those of the phenomenally conscious sort – possess no content, then there is simply no question of their being true or false, veridical or unveridical. (Hutto & Myin 2013, p. 87; cf. p. 134)

13In short, then, Purists claim perception is the pure presentation of bits of a perceiver’s environment. If perception is “committal” with respect to how things are in the environment – which not all Purists accept – the relevant commitments do not go beyond what is strictly given perceptually. Hence, perception can never present things as being different from the way they are.

3. Three Purist Strategies

14Before I turn to the question of whether amodal completion could present a counterexample to The Pure View, I want to consider some ways in which Purists might attempt to handle other potential counterexamples in the vicinity of amodal completion. Amodal completion belongs to a family of phenomena that are sometimes collected under the heading of visual perceptual “filling-in” (Weil & Rees, 2011). Peter De Weerd offers the following general characterization of perceptual filling-in: “Visual perceptual filling-in refers to the interpolation of information across a region in the visual environment in the absence of any physical evidence for that information in that region” (De Weerd, 2006, p. 227). “Interpolation” means that the missing information is filled in on the basis of the information that is physically present in the surrounding area. So the basic idea is that the perceptual system adds information that is not present in the perceptual stimulus. This sounds like it might be precisely the sort of thing Purists deny is possible – viz. cases in which perception goes beyond what is strictly given or presented. However, Purists have various ways of responding, depending on the precise nature of the case.

15First, consider Troxler fading (Fig. 1). In Troxler fading – or Troxler’s effect – stimuli surrounding a fixated object slowly fade away. It seems Purists can easily handle this sort of case. For it is not clear there is any experiential filling-in going on. In Troxler fading, subjects fail to see something that is there. But that does not mean that something else – the background – is “filled in” to replace what is not seen. In general, failing to spot x is not tantamount to perceptually representing y (≠ x). If it were, then given that we overlook things all the time, virtually all our perceptual experiences would be grossly non-veridical.

16Other cases of filling-in must be handled differently, however. In “modal completion” (Michotte et al., 1991), for example, an observer has a visual impression of a figure or surface where none is present in the stimulus. The Kanisza square (Fig. 2) is a standard case of modal completion. It seems as though there is a bright white square in the foreground, partially occluding (blocking from view) four black discs. Yet the foregrounded square is illusory. The image only contains four “pac men”, suitably arranged. The illusory square seems to be added, experientially, to the image. And thus it can hardly be denied that filling-in is occurring.

17What Purists can say, however, is that the relevant filling-in is not strictly perceptual. Three strategies suggest themselves: First, Purist can maintain that beliefs or judgements are doing the filling-in. Second, they can maintain that (partial) hallucinations are responsible. Finally, they can say that filling-in is accomplished by mental imagery.

18Consider the doxastic strategy first. The idea is that in Fig. 2, it is not your visual experience that fills in the missing contours of a square. Rather, there seems to be a white square in the sense that you believe or judge – or are at least inclined to believe or judge – that there is such a thing.

19The doxastic strategy might work for modal completion. A naïve subject, going by a real-life example of a Kanisza square, would surely be inclined to believe or judge that there was a square object present. Maybe we do not need to posit an underlying perceptual representation of such an object. Maybe the doxastic representation is all we need.

21Perhaps the hallucination strategy can yield a plausible account of modal completion. It vividly seems to you that as though a bright white square is there in the image; yet there is no such thing. This seems a textbook case of visual hallucination – a “perception” or perceptual experience without an object.

22Should the hallucination strategy fail, a third strategy is available to the Purist: the imagery strategy. To opt for this strategy is to maintain that filling-in is achieved by way of mental imagery. Now, as a rule, imagery tends to be less compelling than perception. We rarely mistake our imaginations for reality. Thus, the less convincing or compelling the filling-in would be, the more plausible would be the suggestion that imagery is responsible. To the extent that the added contours in Fig. 2 are experienced as less than compelling, the imagery strategy might seem promising. Does it really look as though an actual square is present in the image? If not, perhaps that is because what is doing the filling-in here is visual imagery rather than a partial hallucination.

23Purists can also avail themselves of these strategies when trying to account for other cases of allegedly perceptual filling-in. Take filling-in at the blind spot, for example (Fig. 3).

25The question I now want to consider is whether any of these strategies yields a plausible explanation of cases of amodal completion.

4. Amodal Completion

26Cases of amodal completion are importantly different from the other kinds of perceptual filling-in that we have discussed so far. Consider how the Belgian vision scientist Albert Michotte – who coined the term “amodal completion” (or amodal “complements”: compléments amodaux; see van Lier & Gerbino, 2015 for discussion of the translation) – says about Fig. 4:

[T]here is a gap in the stimulation from the rectangle underneath, but this gap is somehow inoperative, since the subjects say that they perceive two complete rectangles. … There is no question of transparency, and the subjects themselves are emphatic that although they “see” two complete rectangles, it is quite clear that the one underneath is hidden at the crossing point. For this reason Michotte has proposed that completions that are “present” in this manner should be termed “amodal” […]. (Michotte et al., 1991, pp. 143-144)

27Notice the scare quotes around “see” and “present”. The point is that although subjects have a vivid sense of being in perceptual contact with two complete rectangles, it does not seem to them as if they actually see both rectangles in their entirety. Part of one rectangle seems to be hidden behind the other, and thus out of view.

28Given that human beings live in environments that tend to be “cluttered”, as Gibson puts it (1986, p. 78), perceived objects are frequently, indeed mostly, partially covered by other objects. As a consequence, full information about an object is rarely available, and the ability to somehow fill in the missing information is therefore crucial. As Michotte and colleagues put it:

It would be difficult to exaggerate the importance of these perceptual completions. To convince ourselves of this we need only cast a critical eye about us, at the furniture in a room, for example, or at the traffic in a street. It is astonishing how rare it is to find examples of objects where the side facing the observer is completely uncovered. Nearly all of them have parts hidden by other objects (screens), and despite this the shapes we see are neither interrupted nor breeched. Indeed it is clear that the world as it appears to us is not made up of fragments of objects but of things with a complete shape presented to us this way despite the partial and temporary concealment that happens to them. This is to a great extent due to the formation of amodal completions, which are thus seen to be an essential factor in the phenomenal permanence of things and consequently also in the adequate adaptation of the behaviour of humans and animals to the environment in which they live. (Michotte et al., 1991, p. 165)

29Stressing the crucial importance of amodal completion, however, is one thing, which presumably will find little opposition. It is quite another thing to suppose, as Michotte also seems to, that amodal completion is a perceptual phenomenon after all. The world “appears to us” as made up of “things with a complete shape”, say Michotte et al. Things are “presented to us in this way”; we “see” things in this way.

30However, that perceptual interpretation of amodal completion will be challenged by the Purist. For the commitment to some object continuing behind another object, perhaps in a determinate fashion, is clearly one that goes beyond what is given or presented experientially. Moreover, it is one that could be false. Indeed, as regards Michotte et al’s illustration (Fig. 4), it is false. There is nothing continuing behind the complete rectangle.

31Thus, the Purist must reject that amodal completion is a genuine feature of perceptual experience. The ubiquity and importance of the phenomenon makes the task all the more urgent, as this can hardly be considered an odd outlier to be contrasted with run-of-the-mill perceptual experience. If amodal completion is genuinely perceptual, it is a feature of run-of-the-mill perception.

32The question, then, is whether the Purist’s strategies work for amodal completion. First, let us consider the hallucinatory strategy. Can the Purist suggest that the amodally completed part of an object is simply hallucinated? The short answer, I think, is no.

34The doxastic strategy seems much more promising, however. Perhaps you don’t believe that there are two complete rectangles in Fig. 4, but you certainly feel the pull towards so believing. Indeed, it may only be your background knowledge that this is merely a two-dimensional drawing that is preventing you from giving in to that pull. In real-life cases, you might be hard pressed not to believe the completed portion is really there.

35The doxastic strategy can be challenged, however, by cases in which it seems we clearly do not believe what we amodally complete. Consider Fig. 5. You will not need to be persuaded that this is no ordinary triangle. Now cover the X section with a finger or pencil. The triangle is now amodally completed as though it were an ordinary triangle. Yet do you now believe that it is an ordinary triangle?

36The Purist might want to dig in her heels at this point. Perhaps you have background knowledge or beliefs about triangles, which, when the X bit is covered, leads you to infer that the triangle is an ordinary one. To be sure, this belief will not eradicate your original belief. Thus, the Purist now has to assume you hold an inconsistent set of beliefs. You firmly believe the triangle to be highly irregular – indeed not a triangle at all. And yet you also believe in some sense, or are inclined to believe, that it is a normal triangle.

37Philosophers are often keen to point out that it may not be all that strange or unusual for a person to hold inconsistent beliefs. So perhaps the Purist’s case is not implausible.

38However, let me make three points vis-à-vis this suggestion. First of all, notice how different our scenario is from a standard case of inconsistent beliefs. To take a Sartrean sort of example, I can believe firmly that I am courageous. Yet I can also know (hence believe) that I’ve run away from conflicts again and again. I may also believe that a courageous person is one who does not avoid conflict. The inconsistency can persist, however, because I somehow fail to connect the dots. It is a lot harder for blatantly inconsistent beliefs to co-exist. One tends to give way to the other.

39In the present case, however, we seem to adopt, out of the blue, a belief that blatantly contradicts a firm conviction of ours. It survives for a brief moment (as the X section is covered), and then evaporates as we remove the occluding pencil or finger. Weirder still, the Purist must maintain that as we cover the X bit, there that unruly belief is again. How it can get a second, third, indeed n^th lease of life is utterly mysterious, given our firm conviction that this is no ordinary triangle.

40Secondly, we must not underestimate the degree of irrationality this saddles us with. To believe that Fig. 5 is a normal triangle is not merely to believe something you also believe happens to be false. It is to believe something you firmly believe is nomologically impossible. What was clearly not a triangle, printed on a sheet of perfectly normal paper, cannot all by itself a moment later turn into a normal triangle, then go back to the irregular figure it was and so on. It is impossible.

41Thirdly, Bence Nanay (2010, p. 244) gives an additional, powerful reason to reject the doxastic strategy. Consider how the doxastic strategy proposes to deal with a case like Michotte’s “triangle”: we tend to believe it is normal (when partly occluded), because what we now see interacts with some background beliefs in such a way as to make us infer its normality.

42But now consider Fig. 6. Here you amodally complete an irregular shape, with which you presumably have little prior experience. And you do so despite the fact that what the image as a whole seems to suggest is that the middle figure is a regular octagon like the others. Indeed, you experience the irregular shape even if your background beliefs include the knowledge – as mine do, since I have made the drawing myself – that the figure in the middle is just another regular octagon. As Nanay comments: “Even if a belief could at least sometimes be insensitive to some of our other beliefs, it certainly cannot be insensitive to those of our beliefs it is supposed to be inferred from” (Nanay, 2010, p. 244).

if what it is like to have visual imagery is similar to what it is like to perceive, and being aware of occluded parts is having visual imagery, then, putting these two claims together, we get that what it is like to be aware of the occluded parts of perceived objects is similar to what it is like to perceive those parts that are not occluded. (Nanay, 2010, p. 252)

45The argument is logically impeccable:

1. Visual imagery is similar to perceiving non-occluded (parts of) things.

2. Being aware of occluded parts is having visual imagery.

3. Therefore, being aware of occluded parts is similar to perceiving non-occluded (parts of) things.

46However, the conclusion is false. Being aware of occluded parts is precisely not like seeing non-occluded parts. It does not seem at all to you as though the hidden bits of the middle figure in Fig. 6 are given or presented to you in anything remotely like the way in which you see the non-occluded bits of the figure. The occluded bits seem to be hidden; they precisely do not seem to be seen. Since I agree with Nanay that (1) is indisputable, it is obviously (2) that has to go. Thus, the imagery strategy fails.

5. Concluding Remarks

48I want to end by briefly considering an objection to the argument I have presented. It may seem puzzling how I can say that amodal completion is genuinely perceptual, given the objections I have raised against the hallucinatory and imagery strategies. For in arguing against these, I emphasized that amodal completion was not like perceiving. But then how can I maintain that amodal completion is perceptual?

49This only looks like an inconsistency to the extent that we are still in the grips of The Pure View. For Purists, visual perceptual experience is precisely pure seeing, pure presentation, pure givenness. To abandon The Pure View is to maintain that (some) perceptual experiences have commitments that go beyond what is given or presented – that go beyond what is seen. And there really is no tension between thinking that it is perceptual experience itself that “goes beyond” the given, and maintaining that the “beyond” in question neither is, nor in any way seems to be, seen. Rather, to affirm both of those things is simply to affirm the impurity of at least some perceptual experiences.

50Go back to the inconsistent triad with which I began:

1. Perception occurs.

2. Perception is the pure presentation of objects, property instances, events (etc.) in the world.

3. There is no pure presentation of objects, property instances, events (etc.) in the world.
   
   Since (1) is indisputable – pace Derrida, or Derrida* – the question is whether to abandon (2) or (3). Husserl embraces (3) and rejects (2). For Husserl, perception is impure across the board. But nothing I have said supports (3), at least not directly. For all I have said, there could be cases of ‘pure presentation’. I take no stand on this matter. In lieu of (3), I submit.

4. Some cases of perception – cases involving amodal completion – are impure presentations of objects (etc.) in the world.