Bence Laszlo Tarr
The Theory of Direct Realism
- In the light of Theories of Perception -
Realism in medieval philosophy meant that universals have real existence, meaning that general ideas have an ontological status too, just as the things they were abstracted from. In modern philosophy, the term stands for a theory of knowledge according to which the world is full of independent objects and that it is these we perceive directly when we train our senses on them.
This view is opposed to a variety of views, grouped under the general term positivism , according to which we perceive something else: whether appearances, sense-data or some other intermediate entity that stands between us and the world. If this is true, then we have no direct perception of the independent objects, therefore we must remain skeptic about the reality of the external world. This view seems to be supported by the argument from illusion: Normally we see, feel, hear things as they are, but exceptionally our senses deceive us. We seem to see things that are not there, or things that appears otherwise than they are.
The classical examples: that an oar seems to be broken when put into water, or things look smaller when we move away from them, etc. are known to everybody. Since there are more than one examples to the argument from illusion, it is right to ask: what then is it that we see in such cases?
I would like to give an account of how the theory of perception is formulated around this question, although I will not try to answer the question itself, since that would mean the re-formulation of all the former theories, and that is clearly beyond my reach.
I. Argument from Illusion:
Let’s start with the proposed problem, the ‘argument from illusion’. This argument is in fact not one, but two arguments merged into one. The first argument is that if some of our experiences are illusory then we ought not to trust any of them. This ignores the way in which we settle whether an experience is illusory. This ignorance comes from the difficult and highly criticized definition of what we call experience, which is the root of all misunderstanding. This approach to defining experience in this manner has such a deep root in the history of philosophy, that it seems sensible to retrace it’s major characteristics.
Problematic definition of experience:
According to Aristotle, having an experience of something, is not simply a matter of perceiving things; judgment and memory are needed, and experience is ‘with a view to action’ . For Locke, on the other hand, having an experience of something may simply be a matter of having sensations, for sensation is the ‘great source of most of the ideas we have’ ; and Locke’s concern is with experience as the source of ideas (as opposed to ideas being innate), not with experience (as opposed to knowledge of theory) as a guide to action. Both of Aristotle’s and Locke’s approach follow Plato’s, in imagining that our minds contain a block of wax which, when we perceive something, has an impression imprinted on it as a seal imprints something, an impression on wax ; but they used this image differently. Aristotle, like Plato, used it in his account of memory. Memory is the persistence of the ‘sense-impression’ . To explain sense-perception itself he used his notions of form and matter: a sense is ‘what has the power of receiving into itself the sensible forms of things without the matter’ .
Locke used the image not only in his account of memory, but also in what he said about sense-perception. This led him to postulate, within sense-perception an element, sensation, in respect of which he is active; and to distinguish between them in terms of whether or not the perceiver can properly be said to be right or wrong: he cannot be if he only senses, he must be if he judges. Locke’s use, to explain perception itself, of the image of a seal imprinting an impression on wax, raises a question. How does the perceiver, given the sense-impression or sensation, arrive at a judgment? The answer is a development of the imprinting image. A seal may imprint words on wax in a foreign language. They have to be interpreted. Likewise, in perception, the perceiver ‘interprets’ the sensation. He interprets in the light of past experience, and judges accordingly. The act of interpretation takes place so quickly that he doesn’t notice it, and mistakes what is actually judgment for sensation .
Unfortunately Locke’s introduction of sensations, as elements of perception not only made the definition of experience even harder, but gave a new ground to a whole new series of questions like: where can the sensations be located, and what is their cause, etc. We will have to come back to the problem of definition and location of sensations, but for now let’s try to get back to the main stream of thought.
It must be clear from this small survey, that the main difficulty is that sensation, judgment and experience is very hard to separate, when we have to apply these terms to perception. If we go back to our ‘two pencils’ example, it is hard to say what we exactly perceive. We have a sensation in our crossed fingers (it may be located elsewhere - see Ch. III.), which tells us that there are two pencils, but we have a second visual-sensation in our eyes which tells us the opposite. So if we want to know what we perceive, we have to make a judgment to construct our experience. This is, what is overlooked by the first argument of illusion: It is only by trusting some experiences that we can identify others as illusory. A pencil held between two crossed fingers feels like two pencils, yet we call this perception illusory. We only do this, because we trust our eyes better, than the feeling of there being two pencils between our crossed fingers. The problem from now is as follows: which experience should be regarded as real? Since it would be the most unnerving thing to think, that all of our experiences are illusory. Just because there are perceptions which are illusory, we mustn’t conclude that all are.
Therefore the first argument, which said that we mustn’t trust any of our experiences, is not necessarily true. If we believe that all of our experiences are illusory than we cannot come to the conclusion that there are illusory experiences. (Since we would lack the experience of real.) It is more likely that we must formulate the second argument from illusion, which would say that some of our experiences are illusory, and the real problem is in making the right judgment about which sensations to take as real.
II. Theories of Perception
The argument from illusion points out the main difficulty of all the theories of perception, including the realistic approach. If we want to formulate an elementary theory, supported by the second argument from illusion, we will arrive at what is called the representative theory of perception. Behind this theory lie three philosophical assumptions: a) only judgments can be true or false; b) a person, to be justified, must have some basis for judging as he does; c) the basis must ultimately be something other than an other judgment. Therefore there must be an element that is not a judgment (and so not true or false). This might correspond to the stimulation of the sense organs, but must be mental (such that the perceiver is conscious of it) and the basis for the judgment the perceiver makes. This element is called a sensation or idea, and is said to ‘represent’ the external object, by ‘interpretation’. It provides the basis for the judgment, and suggests that there is an external object, with such-and-such qualities. Perception is accordingly defined as ‘the interpretation of sensations to yield knowledge of the external world’. That the external world exists, is postulated in the act of interpreting, and so knowledge of it is ‘mediated’ or ‘indirect’, while the perceiver is ‘immediately’ or ‘directly’ acquainted with the sensation.
I have given the definition of realism in the foreword to this paper, as the theory of direct perception of independent objects of an external reality. This definition clearly contains one of the main realist theories of perception. If we want to formulate this simplest, and most accepted theory, which supports the realist view of the world, we will arrive at the theory of direct realism. This theory states, that we are directly aware of external objects. If we are to accept the validity of the second argument from illusion, then we must question the possibility of direct perception. This is quite evident: if we were to perceive the objects as they are (directly), we would need to have the same experiences every time we come into contact with the same object. Since the second argument from illusion seems to be true, we have to say that we have no direct perception of the object. Directly, we only perceive sensations, which seem to give a description of the object. Unfortunately sensations seem to be under the influence of the circumstances of perception.
Holders of the first theory object to direct realism that for it perception cannot be causal, and misperceiving is impossible (this stands against the argument from illusion). Direct realists object to the representative theory, that they do not experience any interpreting of sensations, and that something not true or false cannot be a basis for something that is. Representationalists hold that knowledge claims should rest on what cannot be false. The active intellect can be right or wrong in interpreting the sensation, but in sensation itself the mind is a passive recipient or what comes to it from the external world. However, this leaves open whether the external world exists at all. One might reply that the existence of material things is merely a hypothesis, or move on to phenomenalism, the view that material things are not anything over and above sensations.
One can avoid having to choose between representationalism and direct realism, by denying that only judgments can be true or false. In a visual illusion the look of a thing is non-veridical. In the well-known Müller-Lyer figure (see Ch. IV.), one line looks longer than the other, but is not. This is not a matter of judgment. They still look unequal to someone who, knowing it to be an illusion, judges them to be equal. This provides no foothold for generalized skepticism. One cannot be aware of how something looks without being aware of the thing which he looks at.
What is general in every theory of perception, is that it uses sensations as it’s basis. Since both theories face the problem of judging what sensations are to be regarded as significant or real, we should try to give an exact definition of what can be understood under this term.
III. Sensations as Elements of Perception
Before we do this, let’s try to define what we may call perception, after the former analysis. When we say perception, we not only mean the sensation of what we perceive, but the understanding of that sensation. This understanding is based on the judgment of whether what we sense is real or illusory. If these two predicates are valid, then perception is the cognitive apprehension of something. The problem is, as we have pointed out in the second argument from illusion, is that this something is always a sensation; and it is a question of judgment how we understand it. Hallucination, being merely the apparent perception of something (which is not there (there = real)), cannot be discarded as not-perception, although it doesn’t have a real basis for it’s appearing sensations. It seems that the problem surmounts to what we define as sensations.
The problem of defining sensations:
Descartes said that nature has told him, by his bodily sensations, that he was not present in his body merely as a pilot is present in a ship, but ‘as it were mixed up’ with his body, so that he and it formed a unity (Sixth Meditation). By his mind being ‘mixed up with’ his body he did not mean that his mind was in the parts of his body in which he had sensations. ‘Pain in the hand is not felt by the mind inasmuch as it is in the hand, but as it is in the brain’ . That is, for a pain caused by the hand being squeezed to be felt in the hand a nervous impulse must reach the brain, with which the mind has connections via the pineal gland. The problem raised by Descartes is central: where should one locate sensations? For it is not clear at all. Hermann Lotze’s theory of ‘Local Sign’, tried to show that it is the mind, that locates bodily sensations, by understanding the local signs applied to them. Unfortunately this isn’t necessarily true. As Oswald Külpe has pointed out, the experienced difference between a sensation in one’s right hand and one in one’s left can be simply one of the places in which it is felt, just as the experienced difference between sound heard as coming from one’s left can be simply one of the direction from which the sound is heard as coming. Like the explanation of the locating of sound-sources, the explanation of the locating of bodily sensations can be purely physiological.
Whichever is the case, one can talk of having a sensation of something one is touching, such as the furriness and warmth of a cat. We may say that this involves having bodily sensations, in one’s fingers, which give one’s sensations their character. Normally one attends to the qualities of the object, but one can attend to the sensations in one’s fingers: one infers that the cat is furry and warm when touches it. This seems more plausible in the case of the warmth than in the case of the furriness, since warmth is a recognized bodily sensation. There is a difference of category between feeling the warmth of something with one’s finger, and one’s finger feeling warm. One’s fingers feeling warm is not a matter of feeling the warmth of one’s fingers with something (or if it is, we are not talking about a bodily sensation.)
The use of words like ‘warm’, refer to both qualities of objects and to bodily sensations. This double use of all our adjectives have already been noticed by Berkeley, when he argued against calling things warm (First Argument). According to him, being sentient, we can feel warm. To be right in calling objects warm we should have to be justified in ascribing to objects something like what we feel when we feel warm. But objects are non-sentient; so there is no basis for the ascription of warmth to objects. If we accept this, then we should rather call sensations, something presented to the mind, by our senses. If this is the case then we arrive at Locke’s theory of perception, where sensations are only elements in perception. These elements still need to be interpreted, but that is purely a question of judgment. A person cannot be said to be right or wrong in having a sensation, but can be said to be right or wrong in making a judgment.
Those who distinguish this way (like Locke or Descartes) sometimes accuse the ‘vulgar’ of mistaking a judgment for a sensation, particularly as regards the distance and three dimensionality of objects. Descartes says that the size, shape and distance of a staff ‘clearly depend upon the understanding alone’, but ‘are vulgarly assigned to sense’ because ‘custom makes us reason and judge so quickly’ that ‘we fail to distinguish the difference between these two operations and simple sense perceptions.’ 
This theory is certainly the source of the ‘sense-datum’ theory of perception propounded by Russell and Moore and others in the early part of the 20th century. Of course this clearly reflects the positivist doctrine which holds that we must confine ourselves to what is ‘given’ to us in sense-experience as sources of knowledge. Now in this positivistic account, what is ‘given’ in sense-perceptions is the sense-datum. This is just a reformulation of the Locke’s mediator, the sensation. Since the positivistic approach is aware of the problem stated in the second argument from illusion, sense-datum is introduced along with the distinction between the mediated awareness of objects in the external world and immediate awareness of the sensation or sense-datum. The only novum in this theory, is whereas a sensation is by definition mental, a sense-datum might be independent from the mind. Moore, for example. sought to introduce it by a sort of ostensive definition, a ‘picking out’ of an element in one’s experience of an object, an element that might continue to exist after the experience.
IV. Cognitive Appearance as Perception
It seems, that Moore’s wishful thinking is some kind of formulation of the realist and positivist approach. What we call perception has to be more than mere presence of sensations; with the introduction of sense-datum, he tries to account for the cognitive apprehension, we defined in the beginning of the previous chapter. After having enumerated the problems around the term ‘sensation’, we have gained solid ground for saying, that perception consists of sensations, which are presented to the mind, and some kind of mental understanding, which is mainly the judging of sense-datum. This judgment, as we have shown, is about determining the nature of our sensation.
Having to decide when we may call a sensation real, and when can it be merely regarded as appearance or illusion, we have to give a definition of what is really ‘given’ to our senses as the external reality. To do this, we should look at a psychological approach to perception. The word ‘appearance’ is usually contrasted to ‘reality’, but in the following we will have to show, that the two meanings really refer to the same thing, and the ontological distinction between the two is just an interpretation of words. First of all we’ll have to make a distinction between two uses of the word ‘appearance’:
According to the first use of the term ‘appearance’, the shape (form) something appears to be to a given point of view is determined entirely by the laws of perspective. Thus a round object, seen from a point of view at a certain angle to it’s surface, will present an elliptical appearance. Leibniz called this the ‘optical’ appearance. To determine what optical appearance an object presents to a point of view it is not good enough to ask someone to occupies the point of view. He is almost certain to err on the side of the real shape he believes the object to be. Only if he erects a transparent screen at right angles to his line of vision of the object, and traces the objects outline on it, will he get its optical appearance right.  In the case of an illusion, such as that given by the Müller-Lyer figure [Fig. 1.], lines which are the same length in the original will be the same length in the tracing (i.e. in the optical appearance). In an ambiguous figure, such as the one which can be seen as a duck looking in one direction or a rabbit looking in the other [Fig. 2.], the same ambiguity will appear in the tracing as in the original.
In a quite different sense of ‘appears’, the lines in the Müller-Lyer figure -will appear unequal in length, and the ambiguous figure will look to a perceiver either like a duck or like a rabbit. This double use of words seem to undermine every philosophical misunderstanding. The two senses of the use of ‘appearance’ can be traced in one of Reid’s writings . He remarks on an artist’s need to acquire ‘the habit of distinguishing the appearance of objects to the eye, from the judgments we form by sight, of their color, distance magnitude and figure’. If we look at the sea from a cliff-top the ‘judgment we form by sight’ may be of a sea that is uniformly blue. To create this impression in the picture the artist must use a dark shade of blue for the near sea, a light shade for the far sea, and intermediate shades in between. To know what shade to use must attend to the sea in a manner that must be learnt and may be described as attending to the optical appearance. However, Reid is not writing about the optical appearance when he says that the masters in painting ‘know how to make objects appear to be the same color by making their picture really of different color’.
This is the second use of the term appearance. He is using the word ‘appear’ with the sense it has in the true proposition ‘the person who draws the Müller-Lyer figure knows how to make lines of equal length appear to be unequal’. How things appear, in this sense, normally determines what people judge themselves to be looking at. We add ‘normally’ to provide for the exceptional cases in which a person has reason to think that things are not as they appear. This sense of ‘appears’ may therefore be called the cognitive appearance (pertaining to knowledge).
There are four related conceptual differences between cognitive and optical appearances: a) Cognitive appearances are subjective, whereas optical appearances are objective. The optical appearance of an object to a point of view is a function of the object’s real figure, color, and spatial position, but of nothing else. There would still be optical appearances of objects to points of view even if all sentient life ceased to exist. Cognitive appearance of objects appear to sentient beings that can recognize them. b) If something cognitive appears to somebody, he must know that it does; if optically, he need not. Psychological experiments have shown, that even if somebody is trying to attend to the optical appearance he may get it wrong. c) Something can cognitively appear to somebody only if he possesses the appropriate concept. For example: the duck-rabbit can look like a duck only to someone who knows what a duck looks like; not for optical appearances. d) Cognitive appearances are related to their objects by being true or false of them. For example, the lines in the Müller-Lyer figure look unequal in length, but this appearance is non-veridical; really they are equal. Optical appearances are not true or false of the objects of which they are appearances: no more than the size of one angel of a triangle is true (or false) of the other two angels, which determine it’s size.
For an understanding of sense-perception as are means of knowing about the world it is the cognitive appearance that matters. This is wholly in accord with the current psychological approach according to which perception is really a hypothesis-making process.
If we accept cognitive appearance as the basis of our sense-perception, then we cannot say that external objects are given to our perception either directly or in their full meaning. Therefore direct realism shouldn’t be more than a naive interpretation of what we perceive, that can only serve as a scientific method for a scientist in action.
[ » Read the extended version of this paper ]
 This generalization is given in Moritz Schlick’s article: Positivism and Realism
 Metaphysics, Bk. 1, Ch. 1.
 Essay, II. i/3.
 Theaetetus, 191c.
 Posterior Analytics, 99b 36.
 De Anima, 424 a 18.
 Essay, II. ix/8-10.
 Principles, IV. cxcvi.
 Reply to Objections, VI. 9.
 This example can be found in Atkinson et ali: Psychology, Ch. III.
 Inquiry into the Human Mind, VI. 3.