Saturday, 30 September 2017

The Structure of Representations

[C]lassification cannot be embodied in, or reduced to, or dissolved in, discriminant behaviour directed towards the classified objects, but involves the use of symbols specific to the purpose. These symbols will be explicit symbols: they will not be mistaken for the things classified; they will stand off from the world whose contents are being classified. (Raymond Tallis 2000)

It’s almost a platitude to say that representations come in all shapes and sizes, but it isn’t a platitude to observe that there is no representation without structure. Even in their most ephemeral and ethereal forms, words, images, models and performances are far from immaterial, and when we consume them, their contours rarely escape our notice. If not for this awareness of the differences between representations and the things they are used to represent, there could be little, if anything in the way of representations or acts of representing. A clone of an apple is just another apple[1]. If on waking, we were to find ourselves in an identical copy of this world, we would accept it as readily as this one. Interestingly, such a copy would not qualify as an illusion because the concept of sameness logically excludes illusion. If you surreptitiously swap your dog’s dinner bowl with another of the same design, there is no illusion involved and no representation either. So much for the suggestion that we might be dwelling in a simulated universe!

Not only are representations structured, they must be crafted, manufactured or otherwise performed, and thus the skills necessary to do so must first be discovered, developed and refined through social and environmental feedback. Nothing as sophisticated as representation could just spring out of the evolutionary sandbox fully formed. Deaf songbirds never learn to sing, let alone to pick-up the localised “dialects” paraded by their chirruping friends (Marler 2004). Without feedback, there could be no opportunity for the shaping or “grounding” (Harnad 1990) of communication. There could be no right way and no wrong way of depicting, characterising or even mimicking aspects of the environment. Even camouflage, which isn’t strictly a form of representation[2], would be impossible, because there would be no constraining mechanisms guiding its evolutionary development.

Perceptibility in Principle

Must representations always be perceptible? Can imperceptibly small images, models or words still be representations? So long as a candidate representation remains perceptible in principle, it qualifies as a representation. But if we divide a representation into fragments and rearrange these in an unrecognisable form, then they would no longer qualify.

"Imagine a painting cut up into small, almost monochromatic bits which are then used as pieces in a jig-saw puzzle. Even when such a piece is not monochromatic it would not indicate any three-dimensional shape, but should appear as the flat colour patch. Only together with the other pieces does it become a bit of blue sky, a shadow, a highlight, transparent or opaque, etc. Do the individual pieces show us the real colours of the parts of the picture?" (Wittgenstein, “Remarks on Colour” §60, 1977)

Data Processing and Storage

It might be argued that the data stored on a hard drive is representational. At best, this would be “representation” dressed up in a lab coat, not representation in its ordinary non-technical attire. Many objects and patterns have representational utility (i.e. they are apt for representational interpretation and use), but they do not constitute representations unless we treat them as such. So for example, the highly patterned structure of DNA is widely regarded as a code, but few self-respecting biologists would wager that the “genetic code” is anything other than a convenient metaphor for the exquisitely computable structures of Guanine, Adenine, Thymine and Cytosine. Similarly, the facial movements of ordinary speech are not representations in addition to the words we produce, but they are nonetheless a visible component of speech and can be reliably interpreted as representations by lip readers.

So when we say that photographs, videos and emails etc. are stored on digital media, it needn’t follow that this storage is representational too. Computation relies on the fact that we can convert analogue signals into the meaningless binary states of digital processing and storage media. Think of Wittgenstein’s jigsaw again. When the picture is disassembled, it no longer represents anything. Only when the fragments are reassembled do they become an image once more. So even though computers disassemble their various inputs in highly structured ways (into what we commonly refer to as “data’), this structuring alone is not sufficient to constitute representation. Data stored on digital media has to be converted back into representational form before we can recognise it as such.

According to Jerry Fodor (1981): “There is no computation without representation.” This assertion lies at the foundations of computationalism; the doctrine that our brains are biological computing mechanisms. Fodor is half right. There can be no computation without representation because computation is a human invention necessarily dependent upon a history of cultural innovations involving the manipulation of symbolic representations. If not for this thoroughly human knowhow we would never have discovered how to create machines capable of saving us the trouble of crafting, producing and manipulating (including computing) representations in the first place.

[1] This is not to suggest that apples are any less apt for representational use than anything else of course.
[2] Because it isn’t performed for the purposes of communication.

Thanks to John Ragin, for some very helpful discussions about digital processing and storage.

Monday, 11 September 2017

I Refute It Thus!

Some theorists take the view that we have no direct perceptual access to the world. They argue that perception is mediated by our representational skills, creative techniques and—if they are to be consistent—the raw materials we use as well (although—tellingly—many would deny this latter condition). This doctrine is known as Transcendental Idealism and was first propounded by the German philosopher Emanuel Kant in the 18th Century. Idealism also comes in a vanilla edition which takes perception to be a creation of the mind or brain for the benefit of... well for the benefit of the mind or brain I guess. Many withering and sometimes funny attempts have been made to discredit Idealism, but its followers seem to be incurable. In contrast to both forms of Idealism, Realism—which also comes in various flavours—takes the world to be very largely as we find it.

The Realist philosopher Karl Popper claimed that Idealism (including Transcendental Idealism no doubt) and Realism are “neither demonstrable nor refutable”. Perhaps he was right. However, some would argue—with a tinge of irony—that his claim itself is not beyond refutation. One famous attempt at a refutation of Idealism was performed by Samuel Johnson, who kicked a neighbouring stone and quipped: “I refute it thus!” Most philosophers find his demonstration to be thoroughly unconvincing. Nonetheless, Johnson’s perfunctory gesture may have more to commend it than is ordinarily conceded or acknowledged.
Natural or manufactured objects, like the heart [or brain], chemical agents, the planets or engines, have an action, which may be slow, complicated or beautiful; but they do not take action, they do not act, however much they may act on other things. (Alan White "The Philosophy of Action", 1968 p.2)
Johnson had to take action to kick the rock; in order to make his demonstration. There is no such thing as a representation without some actions being taken to produce it. But a brain or mind cannot take action, least of all representational action. Brains are obviously significantly involved in the taking of actions but they do not have agency or take actions on their own or anyone else's behalf. The Idealist's claim that the mind/brain creates representations simply doesn't stand up to scrutiny. Even Transcendental Idealists reject Idealism as wholly incoherent.

When we act, our actions are comprised of countless unthinking physiological processes that have been shaped by millions of years of evolutionary development. Representation is merely a very recent cultural and fully public outgrowth from a winnowing process that has left countless behaviours and sensory failures in its wake. Life in general is a testament to the undeniable efficacy of mindless sensory responsiveness. It is this sensory integration, and not our representations or even our perceptions, that determines what we reliably take for granted.
“The thesis that ‘our representational practices determine all the divisions’ is vulnerable to the criticism that the de facto success of our representational practices can only be attributable to regularities that are implicit in the relationships between the components of the universe itself.” (Donald Brook, in personal correspondence 20/09/16)
Some people find Brook a little difficult to grasp, so perhaps I can try to put it differently. If the world were not comprised of objects and circumstances in precisely the configurations that we find them, then our representations, not to mention our biological processes, would never have gotten off the ground in the first place.

Saturday, 5 August 2017


A Roman walks into a bar, holds up two fingers to the bartender, and says “Five beers please.” 
A two-fingered gesture can be used as a stand-in for two objects or (according to a different convention) five, it can symbolise a “V” for victory or be thrust in the air as a symbolic act of defiance. Evidently even the simplest gestures can be interpreted in various ways. But what may not be clear is that the meanings we ascribe to things often have much less to do with their nature than with our “nature” as communicators; with our “form of life” as Wittgenstein helpfully put it. The following discussion is intended to show that symbolic/semantic content—or “meaning” as it is ordinarily known—depends upon the shared conventions of a discursive community and is therefore fundamentally grounded in culture. I also hope to shed some light on one or two of the confusions that can arise when theorists discuss the notion of content.

In ordinary language, we frequently speak of films, books, images etc. in terms of their content, usually in reference to their meaning. At other times we might refer to their content in a different way: in respect of its ethical implications. In this sense, to describe a film say, as having explicit content, need not suggest that it has explicit meaning. Explicit, graphic, adult and other forms of what we might call “ethically sensitive content” are therefore conceptually distinct from the symbolic content (the meaning) of a film, story or picture etc. This is why Picasso's “Guernica” is not censored, because its violence is largely implied. Figurative or abstracted pain, suffering or violence of this sort is generally considered to be of less concern than its more literal, explicit or graphic incarnations. Goya’s depictions of the horrors of war, on the other hand, are unquestionably disturbing because they leave so little to the imagination. 

"Guernica", Pablo Picasso, 1937.
Whilst content is commonly associated with representational media, it is not exclusive to them. To the extent that any object or event can be measured or interpreted, it can also be said to have content. Familiar examples include the nutritional content of foodstuffs, the mineral content of soil deposits or the energy content of chemicals and other substances. Light from distant celestial events also carries content, as do our genes. Even the style and state of our clothes has content that can convey information about our preferences, social position and sometimes even our political tendencies.

Many of the sciences are concerned with the discovery, observation and measurement of quantifiable forms of naturally occurring content. The arts, on the other hand, are much more concerned with the interpretation of content of the cultural sort. Unlike quantifiable forms of content, symbolic meaning (sometimes called “semantic content”) is not an essence that can be extracted, distilled or derived from representations by probing their constituent parts. So a satisfactory answer to the question "What is the meaning of Picasso’s 'Guernica'?" would not be given by describing its depicted features, no matter how exhaustively or precisely. And a detailed appraisal of the materials used in its manufacture would miss the point entirely. Instead, the meaning of representations is largely (perhaps entirely) dependent upon the interpretive and associative abilities that we bring to bear upon them. In other words, we imbue things with meaning, and we do this according to skills that we acquire to a very significant degree through our participation in discursive culture.

It might be helpful to consider the difference between content and contents. I can read the contents-page of a book, and this may give an indication of the book’s content, but if I turn the book upside-down, it would be absurd to suggest that its content has also been turned upside-down. Similarly, I can pour the contents of a packet of nuts into a bowl, but it would be misleading to suggest that I have also tipped their energy content into the bowl. The point here is that content and contents are often liberated, released, extracted or otherwise accessed in very different ways. Where representations are concerned, it makes little sense to say that we can extract, release or liberate their meaning, because, as I have already tried to make clear, the meaning of a representation is not a quantifiable feature. Meanings can be accessed of course, but this relies on a familiarity with the ideas and associations that make things intelligible not on any form of determinable magnitude.

In the visual arts, we commonly distinguish between the form of an image, its pictured subject and it’s meaning. The terms used to describe this triangular relationship may vary, but in general, everybody understands the difference between what a depiction shows, what it is about and it’s material constituents. Interestingly, this relationship is also reflected in ordinary language: in the basic prepositions we use to describe images. We distinguish between what an image is of, what it is about and what it is made from. So even though some of the preferred terminology may vary, this need not suggest any underlying confusion over the conceptual differences involved.

Sometimes theorists use the terms “connotation” and “denotation” to discuss the content of representations. Definitions of these terms typically correlate them with literal and figurative content. So the literal/denotative content of the Jolly Roger sign is a skull and crossed bones, whereas it's figurative/connotative content is piracy. In use though, a problem arises with the notion of denotation as it applies to nonverbal representations. It is commonly claimed that the denotative content of a photograph of an apple say, is an apple and that the photograph thus denotes an apple. Notice first that there is no equivalent usage of the word “literal”. We would not say that the photograph literates an apple or is literative of an apple. Nor in fact do we say that the photograph is literally an apple although we might say that it is literally of an apple.

The denotative content of a poem is its literal or obvious meaning and is necessarily primary, but the meaning of an image is not its primary representational feature; the pictured subject is. In other words, the photograph does not mean an apple, it depicts an apple. So if the denotative subject of an image is simply the pictured subject (the thing it depicts), then there is little need for additional jargon, especially if this misleadingly characterises images as quasi-linguistic artefacts. The same applies to representational sculptures, models, maps and other forms of nonverbal communication.
Denotative communication as it occurs at the human level is only possible after the evolution of a complex set of meta-linguistic (but not verbalised) rules which govern how words and sentences shall be related to objects and events. It is therefore appropriate to look for the evolution of such meta-linguistic and/or meta-communicative rules at the pre-verbal level. (Bateson 1955)
Some philosophers, claim that experience and consciousness have representational content. Strangely, many of these same philosophers make no clear distinction between “contents” and “content” (see here). In fact they seem to take “contents” merely to be the plural form of “content”. This is equivalent to saying that the subversive content of Piero Manzoni’s infamous can of “Artist’s Shit” is the same as its unappealing contents. Something is awry. Furthermore, when philosophers speak in this way of “contents”, they misleadingly imply, and may even mistakenly believe, that it makes sense to regard this as a detectable—and thus measurable—feature of the brain. As should be clear from the “Guernica” and Manzoni examples, the content of a representation is not to be found by prodding around in its contents.

It might be argued that the brain/mind is different in this regard, that it contains our thoughts and that these are therefore rightfully described as content. If “mental content” means anything, surely it refers to our thoughts, and these happen in our heads? It is true that we sometimes talk of thoughts being “in our heads”, but consider the following question: “Where were you having that thought about buying a new phone?” Not “In my head” but “At work”, “On the way to the shops”, “In the car driving along the High Street” etc. Thinking is an activity, and it is carried out wherever we happen to be. The fact that a significant portion of its biological operations occur in the brain, does not mean that its performance can be intelligibly reduced to the neural level, even if it seems scientifically shrewd to do so. Thinking is something whole people do, not their brains, minds or neurons.

Another problem with “the content view,” as it is known in philosophy, is that it confuses the kinds of accounts we give of experiences with the kinds of accounts we give of objects experienced. To describe what an object “is like”, is to make a comparison of some kind, invariably with a familiar object or some feature of it. Interestingly, to do so is to pick out a suitable representational relationship—a likeness in fact. But to describe what an experience is like is entirely different. We don't say that our experience of a lime is like a green lemon or even that it is like a lime. We say that the experience is nice, horrible, disgusting or whatever. As Peter Hacker makes clear (here), the qualities of experiences are given in hedonic terms, not in terms of the qualities of objects.

Charles Travis (here) is also critical of the view that experience has representational content. His arguments are quite lengthy and involved, so for the sake of brevity, I will mention what I think is a decisive point: "If we are going to be represented to in experience, then the relevant representing must be something we can appreciate for what it thus is." In other words, if we fail to recognise that something is a representation, then there is no question of our grasping its intended use. As Wittgenstein famously remarked of an arrow-like “dead line on paper”: “The arrow points only in the application that a living being makes of it.” This is because representations and the symbolic content that we often ascribe to them are cultural contrivances. The representational currency of an object is necessarily secondary to its form; the “dead line on paper”. We first have to recognise the line, before we can appreciate its application.
We can apprehend the representational properties of representations only because we can perceive the non-representational ones. (Hacker 2003 p.193)
It is silly to suppose that the world is representing itself to us or that we must necessarily be representing it to ourselves in order to perceive it. Meaning is attributed to the world; we imbue things with it according to skills we learn as representation-users; as communicators. 
“Anything can be a symbol and, in human life, almost anything is.” (Noble and Davidson 1996)
Angus' "Ant City". Cambridge 1/8/17

Thursday, 26 January 2017

Illusionistic Innovations

If you see the drawing as such-and-such an animal, what I expect from you will be pretty different from what I expect when you merely know what it is meant to be. (L. Wittgenstein. Philosophical Investigations. 1953, p205e)
If you merely know that a young child’s drawing is meant to be of a cat, it is probable that the drawing shares very little in common with a cat. But if you momentarily mistook a photograph or skilful drawing of a cat for an actual cat, it would be extraordinary if the picture turned out to share nothing in common with a cat.
Now a simple phenomenological inspection of any representation, either a drawing or a photo, shows us that an image possesses none of the properties of the object represented. (U. Eco. “Critique of the Image,” 1970)
For Eco, the relation between images and “real phenomena” is “wholly arbitrary”. But this is surely mistaken. Words like “cat” certainly do have a "wholly arbitrary" relation to the things they refer to. Consequently there is no question of our mistaking the word “cat” for a four-legged animal of the feline variety. But if, as Eco claims, images also share “none of the properties” of the things they represent, then how is it possible that we can very occasionally mistake what turn out to be images for the things they represent? Eco offers no explanation. I suggest that if it is true that we can sometimes mistake the properties of one thing for the wholly different properties of another thing, then it is reasonable to suppose that we must be dealing with some form of illusion.
It is important to realise here how familiarity, so to speak, takes the edge off illusion. Is the cinema a case of illusion? Well, just possibly the first man who ever saw moving pictures may have felt inclined to say that here was a case of illusion. But in fact it's pretty unlikely that even he, even momentarily, was actually taken in; and by now the whole thing is so ordinary a part of our lives that it never occurs to us even to raise the question. We might as well ask whether producing a photograph is producing an illusion—which would plainly be just silly. (J. L. Austin. “Sense and Sensibilia.” 1960, p26)
Silly as it might be to ask such a question, it wouldn't be silly to suppose that we could use a photograph to construct an illusion. Nor would it be silly in certain circumstances to momentarily mistake a life-sized photograph of a person for an actual person. So whilst familiarity may take the edge off illusion, it doesn’t eliminate the possibility of illusion and nor might it diminish the importance of the concept.
“The application of the concept of an illusion in general presupposes a concept of being wrong in the sense that were we never wrong in what we perceive, were we never to make a false judgement about what we perceive, we should not have the concept of an illusion.” (D. W. Hamlyn. Sensation and Perception. 1961, p196)
A few years earlier, Hamlyn (no relation by the way) made some similar remarks on the same subject:
There is not necessarily anything about an illusion which tells us that it is one, for if there were it would not be appropriate to say that we were ever taken in. This is not to say that it is always right to say of someone who sees something wrongly that he is taken in; for he may see it in this way despite the fact that he knows the thing in question is not like this. But in order for it to be appropriate to talk of illusions it must sometimes be the case that people are deceived.” (D. W. Hamlyn. “The Visual Field and Perception. 1957, p112)
Now if this person “knows the thing in question is not like this” then he is not seeing it “wrongly” at all. He is seeing it as anyone else with the same perceptual faculties would see it. Moreover he sees that it is illusory in some respects and he probably also knows that other people would find it illusory in the same ways. So if I say that the moon looks like a proximate flat silvery circle, I am not making a perceptual claim that should be regarded as an instance of false or wrong perception. I am using a commonplace expression that will usually be readily accepted by anyone familiar with the ways in which illusory appearances are generally expressed in language.
Shared discriminatory capacities are a precondition for shared concepts of colour, taste, sound, smell, etc. Moreover, shared propensities for perceptual illusion are a precondition for shared concepts of perceptual appearances as distinct from actualities, viz. concepts of objects publicly looking thus-and-so although not being so. (Baker and Hacker. Wittgenstein: Rules, Grammar and Necessity. 2010, p.215)
One of my principal aims in exploring these issues is to show how these “concepts of perceptual appearances” are linguistic outgrowths from innovations in illusionistic representation. In other words, without these techniques, it would make no sense to say that yonder house looks small or that a glossy surface looks wet or that a static white cinema screen looks like a multi-coloured window onto a world of moving objects and people.
It may be that people without any experience of pictorial simulation would not say that the distant hills look blue, but even such innocents would probably be tricked, by being smuggled into a good planetarium, into believing that they were looking at the open night sky. (D. Brook. “How to Draw the Curtains.” 1985. My emphasis)

Thursday, 22 December 2016

The Achievement of Illusion

Akiyoshi Kitaoka

There’s an interesting discussion on the New York Review of Books at the moment between writer Tim Parks and realist philosopher Riccardo Manzotti on the subject of consciousness. Manzotti rejects the idea, popular in much cognitive neuroscience, that our experience of the world is a mediated product of the mind or brain. Instead Manzotti takes an “externalist” stance that conceives of experience as somehow “spread” out across the things of the world. Even if we agree with Manzotti’s realism, this needn’t commit us to either internalism or externalism. Experiences simply happen wherever they happen. A walk in the highlands certainly doesn’t happen in your head, but it makes little sense to say that it is spread over the mountains either.

One of the standard objections to realist accounts of consciousness is what is known as “the argument from illusion”. The basic idea is this. If we are susceptible to optical illusions and other illusory phenomena, then we cannot rely on perception. Manzotti responds:

“Mirages and Hallucinations are not necessarily ‘pure appearances’, one sees something that is really there, only that one takes to be something else. Yet it is not misperception, rather it is misjudgement!”

Few people would claim that mirages, hallucinations and illusions are not caused. But if one mistakes a reflection in a mirror for an actual thing, then it isn’t true that “one sees something that is really there”. If the mirror (the thing that is really there) isn’t seen and its reflection “is taken for something else”, then quite clearly the mirror has been misperceived. Such a response cannot be a “misjudgement” because judgement is not a precursor to perception. If you look before you leap, you don’t judge before you look.

Manzotti seems to have fallen into a conceptual trap. In his eagerness to refute internalism, he has adopted an opposing view when it would have made more sense to stick to his otherwise justified critique of internalism and to point out that consciousness is something we ascribe to whole people, not to their brains or minds.

Parks and Manzotti discuss an interesting optical illusion by Akiyoshi Kitaoka. What appear to be blue and green portions are in fact the same colour. Internalists argue that this illusion shows that everything we perceive is an illusion generated by the mind/brain. If so, this would be a truly miraculous achievement on the part of the brain even if it were sometimes wrong. However, a far simpler and more parsimonious explanation is that these apparent colours are merely the result of the very opposite of an achievement. They are the product of a very ordinary perceptual failure brought about by the circumstances in which the illusion is presented. Change the circumstances in the right ways and it becomes obvious that there is only one colour involved. No neural mysteries need be imputed.

The effect conforms the same principle used to make colour prints with inkjet printers or to produce colours on the screen you are currently using. In fact, the same basic principle enables us to place tiny black and white squares together to produce a patch that looks grey.

If it weren’t for the fact that all normally sighted people are susceptible to illusions in the same sorts of ways and in the same sorts of circumstances, then all illusionistic media (images, movies, representational paintings and drawings etc.) would be unacceptable as representations. There would simply be no ways and no circumstances in which such representations would be like the things they might otherwise represent.

Sometimes things can look like other things because they are genuinely alike. Two leaves from the same tree will usually be alike in a whole variety of respects. Such isomorphism is the very basis of what we take to be observer-independence, of what is objectively real. On the other hand, things can look like other things because in some regularly occurring or contrivable circumstances it can be difficult for us to discriminate between them in one or more respects. Equally, and for the same reasons, two things that are actually the same in one or more respects can seem to be different. The mastery of illusion is not an achievement of the brain, it is an achievement of human culture.

I will leave the last word to the Scottish realist philosopher Thomas Reid:

“The sceptic asks me, Why do you believe the existence of the external object which you perceive? This belief, sir, is none of my manufacture; it came from the mint of Nature; it bears her image and superscription; and, if it is not right, the fault is not mine: I even took it upon trust, and without suspicion. Reason, says the sceptic, is the only judge of truth, and you ought to throw off every opinion and every belief that is not grounded on reason. Why, sir, should I believe the faculty of reason more than that of perception?—they came both out of the same shop, and were made by the same artist; and if he puts one piece of false ware into my hands, what should hinder him from putting another?” (Thomas Reid, IHM 6.20, 168–169)

Saturday, 3 December 2016

Masters of Illusion

A Cartesian can believe that the existing world is not visible, that the only light is that of the mind, and that all vision takes place in God. A painter cannot grant that our openness to the world is illusory or indirect, that what we see is not the world itself, or that the mind has to do only with its thoughts or with another mind. (Maurice Merleau-Ponty 1964, 186)
The following discussion aims to show that our perception of the world is in no way illusory. Whilst we are obviously susceptible to illusions, our various skills and tools enable us to recognise and exploit many of these susceptibilities in a variety of powerful ways.

According to Alex Byrne: "Perception comprises, by stipulation, veridical perception and illusion" (2009). For Byrne, and many other philosophers, one “veridically perceives an object if and only if one sees it, and it is the way it appears or looks.” In other words, veridical perception is what we usually mean when we speak of perception in ordinary language. It is the perception of how objects ordinarily are, or what we typically call “ordinary perception”, “normal perception” or just plain “perception”. However, like Byrne, some philosophers claim that perception “comprises”, or at least sometimes involves, illusion or illusory perceptions. Two important points need to be made regarding this claim. Firstly, an illusory parrot is not a species of parrot—in fact it is not a parrot at all. And by the same token, an illusory perception is not a perception either, at least not in the respects in which it is illusory. The same is true of misperceptions, false perceptions or perceptual mistakes etc. A failure to perceive something in certain respects cannot be a sort of perception in those respects. Missing a train is not an instance of catching a train.

Secondly, to perceive an illusion is not to be at the mercy of an illusion. It is to recognise the illusion for what it is. To say that one length of an optical illusion "looks" shorter than the other is not to say that the line is shorter than the other. Recognising that the two lines could be mistaken as being of different lengths does not preclude our seeing that they are the same length.

In a forthcoming book chapter, John O’Dea writes: "I can think of no good reason to deny that a tilted coin could be seen as elliptical and flat with respect to the viewer. This would be tantamount to denying the possibility of illusion." It is perfectly justified to say that a tilted circular coin can be treated, regarded or considered as a flatly presented ellipse, because a tilted circular coin can be successfully depicted as a flatly presented ellipse. However, if O’Dea intends “seen” in the sense of “perceived”, which seems likely in the context of his discussion, then his claim should be examined in light of the conclusions we have already drawn about the relation between perception and illusion. Thus, if the alleged perception is an illusion, then it is not a de facto perception in the relevant respects. On the other hand, if looking at a tilted circular coin results in a perception of an illusion—in seeing the illusion for what it is—then a description of the illusion alone will not answer the question of what has been perceived. What has been perceived is a tilted circular coin that can be successfully depicted through the use of a flatly presented ellipse. 

In the concluding paragraph of the chapter, O’Dea writes:
Constancy often fails; deep shadows can make surface colour perceptually unclear; at severe angles, shapes constancy disappears; size becomes harder to judge from more distant objects; and so on. Is perceptual experience illusory in these conditions?
Evidently O’Dea has mistaken illusions of inconstancy for actual failures of constancy. Constancy is usually characterised as the stability we regularly encounter in the properties of perceived objects, despite changes in angle of view, illumination, shading etc. So if circumstances of apparent inconstancy do not constitute actual inconstancy, but rather the illusion of inconstancy, then it would be false to conclude that they are perceived as inconstant. A failure to see that a dead parrot is deceased is not a failure of constancy on the part of the parrot.

If the properties of a stable object, like a book or a table (or even a dead parrot), are perceived as having constancy, there is nothing to prevent us from also regarding, considering or treating these same objects as if they have inconstant features like being blurred when viewed at close quarters or being small when seen from a distance or being colourless in moonlight etc. This is important because it shows that we are often capable of treating things in two quite different ways, one of which involves the capacity to represent the actual properties we perceive whilst the other involves a more sophisticated knowledge of how to represent objects by way of illusory representational techniques.

The following passage from Maurice Merleau-Ponty makes a related point:
It took centuries of painting before the reflections upon the eye were seen, without which the painting remains lifeless and blind, as in the paintings by primitive peoples. The reflection is not seen for itself, since it was able to go unnoticed for so long, and yet it has its function in perception, since its mere absence is enough to remove the life and the expression from objects and from faces. [...] It is not itself seen, but makes the rest be seen. Reflections and lighting in photography are often poorly portrayed because they are transformed into things..." (2012, §364)
It is true that eyes quickly lose their lustre in the absence of lubrication, but if the reflection is not seen for itself, an obvious question arises. Is the reflection on a pond seen for itself, or the reflection in a mirror? Merleau-Ponty provides no obvious answer, but it seems reasonable to conclude that his view would be this. Until the discovery of depictive techniques capable of transforming reflections into things, people were not capable of representing them. In other words, it is by virtue of the emergence of pictorial techniques that visual phenomena like blurring, reflections, perspectival distortions etc. have become communicable (there is even evidence that this applies to the colour blue). This is not to say that reflections have not always played a part in perception. After all, if tears produced no reflections, they would be invisible. As Merleau Ponty says: the reflection “is not itself seen” but it enables the tears, the pond etc. to “be seen”. And what of mirrors? Do we not perceive the reflection in a mirror? His point is that the reflection is not perceived “for itself” as a thing. We might occasionally mistake a reflection for a thing, but this would not constitute a perception. It would be the perceptual equivalent of sitting on the platform of Glasgow Queen Street station dreaming that you are on the 8:15 to Edinburgh Waverly.

Schwitzgebel (2011) raises another reason to be wary about the claim that perception involves illusion. 
Consider the oar's looking bent in water. Could we say that the oar's appearance is an illusion? That seems natural. But if so, then presumably the look of things through a glass of water, which will be similarly distorted, is also an illusion. And if that, then also the look of things through a magnifying glass held appropriately close? Through a telescope? Through ordinary corrective lenses? (166n.8)
It makes little sense to say, for example, that when things recede into the distance, they gradually become illusory. They are simply harder to make out. When we look across a table, we see less of the far side. This isn't an illusion. It's a commonplace and unremarkable consequence of the spatial fall-off of sensory input. The more distant an object, the less we see of it. The effect is regular and gradual and leads eventually to the complete loss of input as objects recede into the distance. If our sensory systems were perfect, there would be no fall-off. But then again, if our sensory systems were perfect, there would be nothing we could not perceive and there would be no such thing as partially seeing or barely hearing something etc. Partial perception and the gradual failure to make things out in certain respects are just normal characteristics of our sensory relation to the world.

In ordinary circumstances, when we perceive things normally, we often describe them as if they have properties that they do not actually possess. We might say that shiny surfaces commonly "look silvery or wet", that rainclouds "look leaden or grey" or that fast-moving objects "appear to be blurred". These are not perceptual reports (or at least not perceptual success reports), but they do not preclude other (arguably more objective) ways of reporting things in terms of their actual perceived properties. In normal usage though, it doesn't really matter which strategy of representation we invoke in describing the objects we see (or "seem to see" on account of the ordinary fallibilities of all sensory systems), because we all familiar with the same representational techniques and we are all subject to very much the same perceptual strengths and weaknesses.
Very much less theoretical attention has been paid to those perceptual failures that are the logical corollary of success. […] In the wake of each positive perceptual advance the reciprocal logic of discrimination failure opens up new prospects for influential representational substitution. (Brook 1997)
This is a vitally important observation with profound implications. To be under the spell of an illusion is to be incapable either of recognising it or exploiting it. It is simply to make a mistake. However, to recognise an illusion is not only to be capable of recognising and perhaps avoiding other similar illusions. It is to have an insight into how other similar illusions might be staged. While many creatures are capable of learning from their mistakes, only tool-users are capable of recognising and exploiting their mistakes in the form of illusionistic tokens.