Skip to main content



Most of the major philosophical problems of perception derive from the fact of "illusions." These problems center on the question whether perception can give us true and direct knowledge of the world, and thus they are basic to epistemology. This entry will describe illusions and set forth and examine the argument from illusion that perception cannot be trusted as a source of knowledge of the external world but affords direct awareness only of appearances or sensa.

Three Kinds of Illusory Experience

The term illusion is used by philosophers to cover a range of phenomena approximately classifiable as follows.

illusions proper

Illusions proper occur when the percipient is deceived or is liable to be deceived in identifying the object perceived or its properties. Psychologists have produced a number of optical illusions, such as equal lines that appear to be of unequal length; a stationary balloon that when inflated and then deflated seems to advance and then recede; and a specially constructed Distorted Room, in which a man looks smaller than a boy. Diseases or drugs, including alcohol, may produce other illusions, such as double images or the unearthly colors and multiple shapes an object may assume for one who has taken mescaline. Other examples are mirages, mirror effects, and conjurers' tricks. The perception of motion introduces many more: At the cinema a rapid succession of slightly different stills on a flat screen makes us see a scene with a three-dimensional perspective in which people move about; the wheels of a coach may seem to be going backward when really they are moving rapidly forward (stroboscopic effect).

relativity of perceptions

A round plate that looks elliptical when seen from an angle and a square table that looks diamond shaped illustrate the relativity of perception. The same water may feel cool to one person and warm to another; the same wine may taste sweet or dry, depending on what one has just been eating; green hills may look blue in the distance; and as a train rushes past, the pitch of its whistle may seem to vary. Further examples are color blindness, shortsightedness, and other physical defects that alter the appearance of things. In all these cases the apparent properties of an object vary relative to the position of the percipient, the distance and media between him and the object, the lighting, the state of his health, body, or sense organs, etc. These are not strictly illusions (they usually do not deceive), and they vary around a norm in which the objects are perceived accurately.


In pure hallucinationsfor example, the pink elephant a drunkard sees, the apparitions of delirium, Macbeth's daggersome physical object is "perceived" when neither it nor anything at all like it is present. In contrast are illusions where the mistake is about the properties, position, or identity of some object actually in view.

Some, perhaps even most, hallucinations are triggered by some perceived feature of a very different character; for instance, a beam of light may be taken to be a person. Many hallucinations are integrated; they fit well with the real background, cast shadows, and vary in size and perspective as they move. One may also class phantom limbs as hallucinations. Pain or other sensations are felt "in the toes," for example, of a leg that has been amputatedthe victim still feels he possesses the missing limb.

Argument from Illusion

The main aim of the argument from illusion is to show by means of illusions that the senses are not to be trusted and that perception is not direct and certain awareness of the real properties of material objects but awareness of appearances only. In fact, this argument involves three subarguments.

(a) a skeptical claim

However sure we are about our perceiving, it is always possible that we are being deceived by one of the many kinds of illusion or hallucination, since it is characteristic of such states that we cannot tell that we are suffering from them. This may in practice be a negligible possibility, but philosophy is concerned with the highest standard of exactitude, and from this strict position perceiving is not absolutely certain because there is always some theoretical possibility of error. Various conclusions can then be drawn. One is that for certain knowledge we must rely not on the senses but on some other faculty, such as intellectual intuition (as in René Descartes); another is that we must abandon commonsense realism.

(b) nature of appearances

In all these illusions there is some thing or quality that does not coincide with the object or object-properties that are in fact presentfor example, the apparitions of hallucinations, the elliptical appearance we see when we look at a round plate, the black shape the color-blind person sees when looking at a red box, the oasis of a mirage, and the second bottle in double vision. All these are merely appearances and cannot be identified with real objects or properties. What then are these appearances? In some cases, and probably in all, they must be sensa, private, probably mental, objects of awareness quite distinct from external material objects, although no doubt they are caused by or resemble material objects.

(c) significance of continuity

If one were to change from seeing an appearance, a private and transitory sensum, to seeing a public, enduring physical object ("public" meaning observable by several persons at one time), one would expect a sudden change in the character of one's sensory experience. But no such jump occurs: There is normally an unbroken continuity between situations where we cannot actually be seeing the material object but are aware only of appearances and situations where we think we see the material object. As we move from where the plate looks elliptical to where it looks round, or as the drunkard looks first at the pink rat and then at the real bed on which it sits, there is a smooth transition. Consequently, even in these seemingly genuine or veridical perceptions we must also be aware of appearances or sensa and not directly of the object itself.

We may note three things concerning our subarguments: (1) Argument (b ), unlike (a ), does not depend on there being error; even if one is not deceived by perspectival distortion, double vision, and so on, the argument that what is really perceived must be sensa is unaffected. (2) The claim in (b ), that the appearances are private and mental existents, depends to some extent on considerations of continuity. Almost all hallucinations, the dark shapes a color-blind person sees or the results of diseases and drugs, are plausibly private to the percipient. But simple perspectival distortions will be private only to the viewpoint. For instance, the elliptical appearance of the plate is as public as the round one in that many may see it at once; this holds similarly for mirages and reflections. Unless causal considerations are introduced, the supposition that each person is then seeing a numerically different but qualitatively similar elliptical appearance or sensum must rely partly on similarity with cases where the content of illusion is undeniably private and partly on the assumption that if the plate is round, then the elliptical appearance must be something other than the plate; but these are hardly compelling grounds. (3) The charge may be made, How do we know that the plate is round or what its real color is? These points would normally be settled by measurement or by reference to standard lighting conditions, but the argument does not rely on this. To take the plate example, it may be put thus: The plate looks elliptical to A and round to B ; it cannot be both round and elliptical, for that would be a self-contradiction; therefore, one of these appearances at least must be quite distinct from the plateand perhaps (by continuity) both are.

Criticism of the Argument from Illusion

The argument from illusion can be countered in various ways.


The skeptical claim is often met by stressing the comparative rarity of illusion and the efficacy of the various tests that can be made to remove doubt. We can use one sense to help another. For example, wax fruit may look like real fruit, but touch and taste reveal it; sight, memory, and testimony can show that a phantom limb does not exist; measurement can settle the real shape of an object; confirmation from others can show up many hallucinations, though there are some group hallucinations; we soon learn to discount alcohol and drugs and may generally argue from known causal factors present. But although these tests reduce the possibility of error in a tested perception to extremely slight and in practice negligible proportions, the critic will still say that it is not absolutely certain and that only absolute certainty will satisfy the philosopher. To this there are two replies. (1) It is logically impossible that we suffer from hallucinations all the time; if no perception were ever certain, then there would be no way of distinguishing hallucinations and illusions from normal perception. (2) The skeptic is misusing the word certain ; well-tested perceptions are just the things we refer to as certain. If we say they are only probable we destroy the normal useful distinction between certain and probable. If nothing is certain, the word has no meaning, and we shall just have to invent a new term for that ordinary distinction.

We may comment on these replies. Reply (1) is of no help in deciding whether any particular perception is certain or notwhich is one of the main pointsand anyhow, a merely approximate certainty would serve to distinguish perceptions from hallucinations. Reply (2) seems to depend on confusing meaning and reference. It is true that perceptions are things we refer to as certain, but that may only be due to our ignorance of the possibility of illusions. The normal meaning of "certain"without any possibility of doubtis correctly adopted by the skeptic; he merely argues that it may only be used of the results of intuition or of mathematical demonstration, not of perception; that is he differs only as to the referents of the word. Also, he can still distinguish between "probable" and "practically certain" in perceptual statements. However, a modified reply to the skeptic may be made (3) that he is in fact limiting the word certain to cases of logical necessity, to those that it is self-contradictory to deny. This limitation not only has the practical disadvantage of destroying the ordinary certainty-probability distinction but also rules out a priori the possibility of any perceptual statement's being certain; thus the lack of certainty in perception is due not to any defect in perceiving but simply to its not being something quite different from what it is, namely intuition or entailment. It would therefore be much more appropriate to use a relaxed standard in dealing with perception and to allow a perceptual statement to be regarded as certain if it has passed all conceivable or all recognized tests. At any rate, there is no reason to suppose that ordinary perceptions are uncertain in the way that the result of a horse race or the nature of next year's weather is uncertain.


The argument from the significance of continuity claims (1) that hallucinations are private sensa, not public material objects, and (2) that since they are indistinguishable from the objects of perceptual consciousness, especially when integrated with them, the latter must also be groups of sensarepresentations, perhaps of external objects.

  1. One answer, based on the usual psychological account of hallucinations, would be that they are not sensa but mental images of an unusually vivid type that are confused with normal perception. To meet point (2), the unusual vividness and the lack of normal discrimination may be stressed and explained by the special circumstances in which almost all hallucinations occur, as when the victim is suffering from fever, drunkenness, drugs, starvation, religious ecstasy, or madness or is influenced by lesser factors, such as fear, acute anxiety, or drowsiness. (In the hallucinations of mescaline the person's mental powers are unimpaired, but he usually recognizes the hallucinations as such and is not deceived into thinking they are real.) It is questionable whether these factors, especially the lesser ones, can account for the integration and triggering of hallucinationscases in which the continuity argument is strong and imagery would seem to merge with genuine perceptions. Also, to be complete this answer would need to offer an explanation of the nature of mental imagery and of why it resembles perceiving. Probably imagery depends on reactivation of the kinds of brain and nervous activity that occur in perception (or in action, if it is motor imagery), and the occurrence of such activity can be detected during the imagery. However, this involves the causal processes, study of which leads by a different route to the abandonment of commonsense theories.
  2. It has been pointed out, by J. L. Austin, for example, that the argument from illusion as applied to hallucinations relies on certain dubious assumptions, namely, that if two things (i.e., an object of genuine perception and an object of hallucination) are not generically the same they cannot look alike, and that they cannot be distinguishable if we in fact fail to distinguish them. The special circumstances cited in point (a ) may come in here as providing reasons for the victim's failing to distinguish what are in fact distinguishable and quite different experiences. This criticism certainly undermines the argument from illusion as a demonstration; for it to be that, these assumptions would have to be accepted as universally true. But it can be replied that an explanation is still required for the general similarity between the two things (sufficiently close a similarity for people suffering only from anxiety to confuse them); also, we need some general theory of the nature of hallucinations and of their integration and triggering.

Phantom limbs are not covered by these points: One can hardly say that the pains and sensations involved are images of genuine onesthey are genuine enoughnor are the victims suffering from drugs or delirium. The usual physiological explanation is that the nerves from the toes, for example, remain in the untouched part of the limb and, being irritated at the stump, send impulses to the brain similar to those they would send if the toes were being crushed or the pain and other receptors in the toes were being otherwise stimulated. This seems to confirm that pain and somatic sensations are private sensa and accords with the general causal theory of representative realism. But it is still arguable that such sensations are very different from sight and hearing, so that nothing follows about the nature of the latter.

illusions and relativity

The argument from the nature of appearances relies on the odd assumption that things cannot look other than they are, that when one apparently sees as elliptical a plate that is actually round, then one cannot really be seeing the plate; one is seeing something, an appearance or sensum, which, being elliptical, cannot be the round plate. But one can simply deny the assumption and say that one is in fact seeing a round plate from such a position that it looks elliptical; its elliptical appearance is not some entity different from it. To treat appearances as entities, as though they were things, is a quite unjustified reification; when we speak of the appearance of something we speak of how it, the original object, appears, not of some other object distinct from it. One may confirm this point by noting that the elliptical shape will appear on a photograph too, so that it cannot be subjective or mental.

The same answer may be applied to the various examples of relativity and illusion. The distant green mountains are actually seen but look blue and may be so photographed; in the Doppler effect we still hear the whistle (a "public" noise), and its apparent variation in pitch may be recorded on tape. In the optical illusions we are still seeing lines on paper, balloons, or a man and a boy, and cameras will photograph them with their deceptive appearances. Again, as Austin has shown in detail, in refraction and reflection we still see the objectthe face in the mirror or the stick in the water; even in a mirage we see a real oasis, though it appears many miles nearer than it actually is. It may also be claimed that the color-blind man sees the red box, even if it looks black to him, and that the man with double vision sees the one bottle, but it looks double to him. (This last point is more dubious: It may be said that looking double is not like looking blue for it involves an extra apparent object and not a differing quality of the one object. On the other hand, the percipient is not seeing two bottles in the same way normal people see one: The bottle and the background have a doubled, slightly defocused appearance that perhaps makes it reasonable to say they look double.)

For some people this general answer is immediately convincing, and it seems incredible that the argument from illusion was ever taken seriously. But others protest that it is inadequate and neglects the immediacy of perception; in the various situations mentioned they seem clearly and directly to be aware of an elliptical shape or a blue expanse of mountain, an advancing balloon, two bottles, or, if color-blind, a black box-shaped expanse. Thus to be told that they are aware only of a round plate, green mountain, stationary balloon, one bottle, and so on, is to them unconvincing and fails to do justice to the facts of experience. This feeling for the immediacy of sensory awareness, and the belief that confrontation is so direct that its apparent object must exist as perceived, is at the bottom of the sense-datum theory. The alternative is to dismiss as illusory this apparent direct and mistakeproof confrontation in perception; perceiving is variable in quality, is affected in its accuracy by position, distance, and many other factors, and may thus be inefficient. It is more plausible to suppose that position, distance, and media distort perception of a round or green object or that color blindness and shortsightedness prevent one from seeing it properly than to suppose that these factors give one excellent and perfect awareness of some elliptical or blue sensum different from the object.

But this is not a final answer, for if one then seeks to discover how these factors affect the quality of perception, one has to go into scientific details. Angle of sight varies the pattern of light striking the eye, refraction or reflection bends the light rays, dust scatters them or absorbs some frequencies rather than others, drugs affect the activity of the nervous system, lack of certain retinal pigments alters the eye's response to light, the Distorted Room and other optical illusions rely in their effects on misleading cues. In short, the effects of illusion point beyond themselves to the causal and psychological processes that underlie perception and constitute its most serious theoretical problem.

See also Perception; Realism; Sensa.


The argument from illusion is a stock feature of introductions to philosophy and is dealt with in most of the books in the bibliographies for the entries Sensa, Realism, and Perception. The books listed below give a fuller treatment of the subject.

Henry Habberley Price's Perception (London: Methuen, 1932) and Charlie Dunbar Broad's Scientific Thought (London: K. Paul, Trench, Trubner, 1923) and The Mind and Its Place in Nature (London: Kegan Paul, 1923) develop the argument to support sense data. Alfred Jules Ayer's Foundations of Empirical Knowledge (London: Macmillan, 1940) develops it to support the introduction of a "sense-datum language." Price has some second thoughts on perspectival distortion in his article "Illusions," in Volume III of Contemporary British Philosophy, edited by Hywel David Lewis (London, 1956), and Ayer's views are presented in his clear introductory work The Problem of Knowledge (London: Macmillan, 1956).

John Langshaw Austin, Sense and Sensibilia (London: Oxford University Press, 1962), D. M. Armstrong, Perception and the Physical World (London: Routledge and Kegan Paul, 1961), and Rodney Julian Hirst, The Problems of Perception (London: Allen and Unwin, 1959), are critical of the argument from illusion. See also George McCreath Wyburn, Ralph William Pickford, and R. J. Hirst, Human Senses and Perception (Toronto: University of Toronto Press, 1964), which, besides criticism, gives some physiological and psychological details. For other attempts to meet the facts of illusion without postulating sensa, see the works listed under "Perspective Realism and the Like" in the bibliography for Realism and under "Sense Data" in the bibliography for Sensa.

The neurologists John Raymond Smythies, in Analysis of Perception (London: Routledge and Paul, 1956), and Russell Brain, in The Nature of Experience (London: Oxford University Press, 1959), give some useful references and nontechnical information on hallucinations. The psychological details of illusion in general can be found in any good introductory textbook on psychology, such as David Krech and Richard S. Crutchfield, Elements of Psychology (New York: Knopf, 1958).

other recommended titles

Alston, William P. The Reliability of Sense Perception. Ithaca, NY: Cornell University Press, 1993.

Chisholm, Roderick. Perceiving: A Philosophical Study. Ithaca, NY: Cornell University Press, 1957.

Crane, Tim. "The Waterfall Illusion (1988)." In Essays on Nonconceptual Content, edited by York H. Gunther. Cambridge, MA: MIT Press, 2003.

Dancy, Jonathan. "Arguments from Illusion." Philosophical Quarterly 45 (1995): 421438.

Dretske, Fred. Seeing and Knowing. London: Routledge and Kegan Paul, 1969.

Firth, Roderick. "Austin and the Argument from Illusion." Philosophical Review 73 (1964): 372382.

Jackson, Frank. Perception: A Representative Theory. Cambridge, U.K.: Cambridge University Press, 1977.

Kenny, A. "The Argument from Illusion in Aristotle's Metaphysics, 100910." Mind 76 (1967): 184197.

Reynolds, Steven. "The Argument from Illusion." Nous 34 (2000): 604621.

Robinson, Howard. "The Objects of Perceptual ExperienceII." Proceedings of the Aristotelian Society, supplement 64 (1990): 151166.

R. J. Hirst (1967)

Bibliography updated by Benjamin Fiedor (2005)

Cite this article
Pick a style below, and copy the text for your bibliography.

  • MLA
  • Chicago
  • APA

"Illusions." Encyclopedia of Philosophy. . 18 Aug. 2018 <>.

"Illusions." Encyclopedia of Philosophy. . (August 18, 2018).

"Illusions." Encyclopedia of Philosophy. . Retrieved August 18, 2018 from

Learn more about citation styles

Citation styles gives you the ability to cite reference entries and articles according to common styles from the Modern Language Association (MLA), The Chicago Manual of Style, and the American Psychological Association (APA).

Within the “Cite this article” tool, pick a style to see how all available information looks when formatted according to that style. Then, copy and paste the text into your bibliography or works cited list.

Because each style has its own formatting nuances that evolve over time and not all information is available for every reference entry or article, cannot guarantee each citation it generates. Therefore, it’s best to use citations as a starting point before checking the style against your school or publication’s requirements and the most-recent information available at these sites:

Modern Language Association

The Chicago Manual of Style

American Psychological Association

  • Most online reference entries and articles do not have page numbers. Therefore, that information is unavailable for most content. However, the date of retrieval is often important. Refer to each style’s convention regarding the best way to format page numbers and retrieval dates.
  • In addition to the MLA, Chicago, and APA styles, your school, university, publication, or institution may have its own requirements for citations. Therefore, be sure to refer to those guidelines when editing your bibliography or works cited list.