Philosophy of Perception: Naïve Realism vs. Representationalism vs. Direct Transformative Process Realism

Painting of a beautiful woman in a garden - Shows the richness of perceptionWithin epistemology, which is the branch of philosophy that studies human knowledge, one of the most fundamental topics is the nature of perception. Philosophers from the Ancient Greeks to the present have offered various theories of what perception is and how it occurs. Because it is a topic so fundamental to human knowledge, specialized natural science can’t answer the basics about it. Science relies on perceptual observations of reality. Thus science itself relies on the idea that perception allows us to be aware of an external reality. If perception does not give us awareness of external reality, then scientific study of the external world is not possible: We would always, at best, be inspecting the contents of our own minds.

Thus, it is the job of philosophy to answer the most basic question: Does perception give us an awareness of reality, and if so, at the most basic level, how?

In this essay, I’ll explain three different theories of perception. To the question of whether perception gives us an awareness of reality, all three of them attempt to answer, “Yes.” Where they disagree is on the “how,” or the basic nature of perception. The three basic theories are naïve realism, representative realism, and Direct Transformative Process Realism (DTPR.) (“Representative realism” here is a synonym for representationalism. Note that these theories are all variants of “realism” in perception. Theories that answer “No” to the question of whether we can observe mind-independent reality would be variants of “idealism.”)

Philosophers sometimes use “naïve realism” as a synonym for “direct realism,” and there are many different theories that could be called “direct realist.” But here I will take “naïve realism” to be one specific sort of direct realist theory: the sort of approach to perception exemplified by the Ancient Greek philosopher, Aristotle. Representationalism too has quite a few variants, but they all share a common thread. I will focus largely on the version of representationalism associated with the English philosopher, John Locke. Direct Transformative Process Realism (DTPR) is my term for the theory of perception put forward by Ayn Rand and Objectivist intellectuals after her. It’s a form of direct realism that is very different from Aristotle’s approach. I’ll explain this term in more detail when I explain this theory later in this essay.

Naïve Realism



Aristotle’s theory of perception is “direct realist,” which means that perception is taken to be the process by which material objects directly impact one’s consciousness through the sense organs. Aristotle says that a perceived object affects the sense organs in such a way that they become like the object in form. When you see a red apple, your eyes acquire a tinge of red. When you taste a strawberry, your tongue acquires a little of the strawberry flavor. And so on. This alteration of your sense organ constitutes your sense experience of the object, and is immediately transferred to your soul or consciousness as a sensory memory.

Thus, we could say that, for Aristotle, (and other naïve realists) perception is where one’s consciousness “mirrors” or “mimics” the outside world in some way. If you see a red apple, there will be something in your consciousness that mimics the shape and color of that apple, and this constitutes your sensory experience of the real apple. If you feel a smooth, cold surface, something in your consciousness somehow mirrors the smoothness and coldness, and this constitutes your experience of that surface.

All the qualities that we sense, such as shape, color, feel, odor, sound, etc., are “out there” in the object. These qualities then make their way into one’s consciousness through the sense organs.

This theory takes perception as basically a passive process of absorption of the external world into consciousness. And under this theory, it makes sense to compare the contents of a given perceptual experience to the external world. (That is, the experience and the outside world are commensurable.) If the form of the experience matches the form of the external object, then the experience is “accurate,” or “true” to the object. In the technical vocabulary of philosophers, the experience is “veridical.” (From the Latin for “true.”) If one’s experience did not match the object, the perception would be “non-veridical.”

Naive Realism in visual perception diagram.

Naïve Realism in visual perception. Aristotle’s theory of perception.

Aristotle himself thought that all sense perceptions automatically and infallibly matched the objects that generated them: He likened the sense experience to “the way in which a piece of wax takes on the impress of a signet-ring without the iron or gold”–a literal “sense impression,” strictly determined by the form of the object that made it. (De Anima ii 12)

But many of his contemporaries, as well as later philosophers, were not so sure. His account of perception leaves open the possibility that something could happen to cause the form in one’s consciousness to not match the form in reality. Couldn’t one use a signet-ring to make a “bad impression”? (I.e. a distorted one?) Or couldn’t the received form be distorted on its way to another portion of the body in which the experience actually takes place?

Even in Aristotle’s time, others like the Sophists had observed that sense experience of a given object varied with variations in the sense organs. So it didn’t seem to make sense to say that the senses could reliably reproduce the form of an external object. Also, illusions seem to be a case where the experience of an object clearly does not match its form in reality. Hallucinations seem to be an instance of perceptual experience where there’s not even any real object to impart its form to one’s consciousness.

These objections to naïve realism were part of the motivation behind the development of the next theory I’ll discuss: representationalism.


John Locke

John Locke

Representative realism is an “indirect realist” theory of perception. This means that real objects are only perceived indirectly, through intermediate “representations” in one’s consciousness. These representations may be called “ideas” (as with John Locke) or “sense data.” They are what is perceived directly, but they are held to be caused by real, physical objects.

In Locke’s version of representationalism, physical objects have two types of qualities directly relevant to perception: primary qualities and secondary qualities. Primary qualities are those qualities of objects that produce “ideas” (sensory experiences) in us that resemble them. They are in the real object, as we perceive them. These qualities include the object’s shape, structure, mass, texture, and motion. So when you see that an apple is round and that it is moving, your experiences of these things match the way the apple is “in itself,” apart from you.

According to Locke, secondary qualities are those qualities of an object that cause in us “ideas” (sensory experiences) that do not resemble the object as it is, “in itself.” They are the abilities of objects to generate in us experiences of color, odor, taste, warmth, cold, physical pain, etc. The secondary qualities in an object are actually a subset of the objects’s primary qualities that aren’t translated into experiences isomorphically–i.e. in a way that resembles them in the object.

An example of what Locke has in mind here would be the surface composition and texture of an apple, versus that of a peach. In the apple, the surface is smooth and composed in such a way as to reflect mostly “red” (low frequency) light. Thus, we see the apple as “red and shiny,” because it reflects “red” light at most angles, and all “colors” of light at certain specific angles (specular reflection.) The surface of a peach, on the other hand, is not smooth or “red-light-reflecting.” It has tiny hairs that disperse light, and a surface that reflects slightly higher frequencies of light, such that it appears “matte” and “pinkish-orange” to us.

For Locke, the “ideas” (sensory experiences) we have of an object’s overall size, shape and motion could be said to represent the real object “faithfully,” whereas the ideas we have of its color, temperature and odor represent the real object “unfaithfully.” The overall shape and motion of the representation that we experience is basically a copy of what is intrinsic in the real object, whereas the color, temperature and odor of the representation is a subjective invention of the individual’s mind.

Representationalism / indirect realism perception diagram - John Locke

Representationalism or indirect realism in visual perception, as theorized by John Locke.

This sort of representational theory of perception is meant to solve some of the problems with naïve realism. Unlike naïve realism, it can account for the problem of colorblind people, who see different (or fewer) colors in objects, or the fact that putting one’s hand in ice-water before touching a rock can make the rock feel warmer than normal. This is because, in representationalism, part of what makes our sensory experiences what they are is our own sensory processing. Elements of our sensory experiences “come from us.” Illusions would happen when the primary qualities as seen in our representation don’t match those in the object, because of what we add to the representation. Hallucinations would happen when we have a “representation” that doesn’t actually represent any physical object, because it is completely generated by us.

But representationalism also generates its own problems. If our ideas of secondary qualities don’t match those in the object, aren’t our senses lying to us? And if our senses lie to us about secondary qualities, how do we know our representations of primary qualities match those in objects? They don’t match when we have an illusion or hallucination, but how do we know we’re not in that state in a given experience, or indeed, all the time? We can never get “outside” our representations to compare them to the object “in itself,” so we seemingly have no way of knowing. This problem has traditionally been called the “veil of perception.”

Representative realist theories other than Locke’s will differ in their details, such as whether secondary qualities are what Locke said they were, or whether there’s even a distinction between primary and secondary qualities, or in the exact nature of the representations, or in what the representations are called, (e.g. “sense data.”) But in my understanding, the “veil of perception” problem is pervasive among these theories. For a pure representative realist, all we ever really deal with, or are directly aware of, are elements of our own consciousness. We never really perceive or deal with the physical world as such, only representations in our consciousness. So why not dispense with the physical world and focus on our representations as our reality?

Thus, in my estimation, pure representative realism leads naturally into de facto idealism. The only reality we can ever deal with is our representations, which are all elements of our consciousness. Reality “in itself” is inaccessible. So “our reality” is effectively composed of our own consciousness, which is the doctrine of subjective idealism.

The next theory I will discuss is one that, in my estimation, solves the conceptual problems of both naïve realism and representative realism, without lapsing into idealism.

Direct Transformative Process Realism

Ayn Rand, novelist and philosopher of Objectivism, a philosophy for living on Earth.

Ayn Rand

Direct Transformative Process Realism (DTPR) is a theory that contrasts with both naïve realism and representative realism. To understand how it’s different, let’s start with a simplified analogy:

Imagine a robot that uses a laser to sense what’s in front of it. It emits this laser, which then bounces off an object and back to a sensor on its front. When the laser light strikes this sensor, it triggers a series of steps that convert the light into an electrical signal that bounces back and forth along a wire in the robot’s head. The frequency of the signal varies with the distance between the robot and the object that reflects the laser. Let’s say that when the object is closer, the frequency is higher, when it’s further, the frequency is lower.

Now the robot also has wheels that it travels on and controls. It can sense when those wheels are turning and how many rotations they make, according to electrical impulses in other wires in its head. The robot can rotate its wheels until it gets to the object in front of it. By comparing the initial frequency of its laser sensor to the number of turns its wheels make before it’s stopped, the robot can determine the distance a given laser sensor frequency corresponds to, in terms of wheel rotations.

So, the first thing to notice about this situation, is that robot’s “awareness” of the object in front of it is not any sort of copy of the object in front of it. The robot’s laser-based awareness of the object is in the form of a resonating electrical impulse with a certain frequency. (And the robot’s wheel-based awareness is also in the form of an electrical impulse.) So it doesn’t make sense to ask if the electrical impulse matches the object’s distance from the robot or not. In other words, it doesn’t make sense to ask if the robot’s laser-sense is “veridical.” The resonating electrical impulse is not commensurable with the distance to the object: The two cannot be compared and judged as “matching” or “not matching.”

This is analogous to the way DTPR understands human sense perception. The signals that impact the senses undergo a process that transforms them into an awareness that is incommensurable with the physical object out in reality. This is why I refer to the theory as a “transformative process realism.” DTPR also regards the whole causal chain that leads from the physical object to a human’s awareness of it, as fundamentally part of the process of perception. All the steps of this chain directly participate in making the perceptual experience what it is. DTPR does not regard states or elements of consciousness as the objects of perception in the direct or fundamental sense. So this theory is a direct theory of perception, rather than an indirect one. Thus I call it Direct Transformative Process Realism.

Now, as against the simplified example of the robot, real human perception is very rich in information. The sense of vision provides, not just one distance, but a whole field of objects with depth, shape, angular size and color. The process of visual perception starts when light is emitted from a source. It then either enters the eyes directly, or is reflected off of one or more objects in one’s visual field. It is focused by the eyes onto the retina. This triggers neural signals and these travel to the brain for processing. This processing then generates an awareness of the solid, physical object that emitted or last reflected the light. This awareness of the object is not isolated, but includes the object’s immediate environment, insofar as it is visually perceivable.

This whole sequence of events constitutes the process of visual perception. The object of the perception is the physical object out in the world. The form of the perception is the way the object is perceived. The form of the perception includes things like the color, shape, texture, brightness, size and depth of the object as perceived. These are the ways in which one experiences the object. If I see an apple, then I am not seeing a “red blob in my mind.” I am seeing the apple, and I am seeing it as red. I am also seeing it as shiny, as rounded, as of a size similar to my hand, and as at a certain depth in my visual field.

Any alteration in the way the process of perception occurs can alter the way the object is perceived (i.e. the form of the perception.) If an object is turned to a different orientation relative to the observer, the form of the observer’s perception of it will likely change. If the color (visible frequency composition) of the light reflecting off of an object changes, then the form of the perception will change. The form of visual perception will be different between a colorblind person and a normally sighted person: Where the normally sighted person may see an apple as red, a totally colorblind person may see the apple as grey. Any alteration in neural processing of retinal signals may also affect the form of the perception.

In each of these cases, the perceived object remains the same, but the way it is seen changes. There is no “right” or “wrong” form in which to perceive the object, because perception is not a process of “copying” or “mirroring” the object. Perception is a transformative process that generates an awareness of the physical object that is incommensurable with it.

In the DTPR theory, the process of perception must also be sharply distinguished from conception. That is, perception must be distinguished from the practice of classifying objects under concepts like “apple” and “table.” Perception is a wordless experience of physical objects. Any words, classifications, or judgments are a separate step beyond sheer perception. Thus, in the terms of contemporary philosophy, DTPR is a non-conceptualist theory of perception.

Diagram of Objectivist Theory of Perception - Direct Transformative Process Realism

Direct Transformative Process Realism in visual perception. Note that this diagram is simplified. Visual perception doesn’t generally present the observer with single objects in isolation, but a whole spread of multiple objects in relation to each other and to the observer.

So, as one might gather from what has been said, DTPR regards sheer perception of objects as infallible. It cannot be “wrong” in the sense of being “non-veridical,” nor can it be wrong about the conceptual classification of an object, since it provides no conceptual classification. (Nor can it be “right” in either of these senses.) Perception is an automatic process in response to physical objects that has no ability to “lie” or “distort.” Perception gives a presentation of an object in a given situation as it gives it, and that’s it. Misidentifications of things, like mistaking a wax apple for a real apple, are mistakes of conceptual judgment, not of perception. (Perception only tells one that there’s something there and that it looks a certain way to one’s eyes, not how to classify it.)

As you can see from the diagram on visual perception, light emission from a source is part of the process of perception. Variations in the color of light striking an object can alter the color of the object-as-perceived. In sheer perception, there is only ever the experience of the color of the object-as-perceived, never an experience of an “intrinsic color” in the object alone.

It’s a separate, conceptual task to define and judge the “intrinsic color” of the object. This can be done scientifically by measuring what proportion of what frequencies of visible light the object reflects. (I.e. the reflectance spectrum.)

So what does DTPR have to say about the arguments from illusion and hallucination that we discussed in regard to the other theories of perception? The cases of illusion and hallucination are very different. In the case of an illusion, a person is actually perceiving something in the world, just in an unusual form. The classic example of a pencil being submerged in water so that it “looks bent” is one such illusion.

Bent Pencil in Water Illusion

A pencil partially submerged in water.

This is a case where the context the pencil is in alters the path of the light it reflects, and thus affects the form in which one perceives the pencil. It is not a case of a perception being “wrong.” Whether the pencil is “straight” or “bent” is a conceptual judgment. It is not given to an observer by sheer perception. There is no single way that a straight pencil is “supposed to look.” It can look many different ways depending on its situation in its environment and relative to the observer:

Pencil Images

The same pencil in 4 different contexts, with different relations to the observer.

A judgment of the pencil as “straight” or “bent” is not directly implied by a single moment’s perception of it. To make this judgment, we would properly take many instances of perception of the pencil–or an extended interaction with it–then use that to apply our definition of what makes something straight or bent to it. We can make an error in doing this, but that is an error in judgment, not perception.

Another illusion is provided by the following pattern:

Optical Illusion-Rotating Circles

The way you see this pattern is not “wrong” or “an error.” The perceptual experience itself does not automatically generate the conclusion that the patterns are rotating. That would take a separate, conceptual judgement. The form in which you see this pattern is affected by the nature of the neural processing in your optic nerves and brain. The neural processing of this pattern generates a form of visual perception that is similar to the way other patterns might look when they are actually rotating. This is the way your visual processing inherently and automatically works, and so can’t be called an error. It’s the form in which you must perceive this particular pattern, given the nature of your means of perception. Nor can your perception fail because it “fails to match the object in reality,” since “successful” perception is not a copy of the real object, in the first place.

DTPR deals with hallucinations by recognizing that they are not perceptions at all. Perception occurs when an external, physical object impacts the senses and causes an awareness of it in a subject. Hallucination is not an instance of that, but an experience of content that one has already internalized. Thus, hallucination is a form of memory and imagination.

All that distinguishes a hallucination from a normal case of imagination is that it is experienced in a way that overrides genuine perception and thus is more difficult to distinguish from it. It is akin to a waking dream, and is no more a case of perception than a dream.

If someone has an occasional hallucination, then they can understand that it’s a hallucination by reference to their genuine perceptions and the conceptual knowledge they’ve built up from them. Hallucinations will tend to go against what the person understands about the world, and so won’t integrate (mesh) with their sense-based knowledge. This will typically allow them to pick out the hallucination as a case of imagination, rather than perception.

But if a person starts hallucinating all the time, then it is like the person is dreaming all the time. The person won’t really be able to make sense of anything. Such a person is no longer conscious, and is in no position to philosophize about the world. To anyone who is not in this state, it is self-evident that they are not, since they can integrate their perceptions and make sense of their experiences.


So we’ve looked at three theories of perception, all of them “realist.” They all contrast with idealism, in that they take the ultimate object of perception to be real–independent of anyone’s consciousness–while idealism takes the “object” to be “ideal”–dependent on someone’s consciousness.

The naïve realism of Aristotle takes perception to be a transfer of the form or qualities of an object to the sense organs and consciousness of the observer. The perceptual experience is like an impression or copy of the form of the external object. The form of this impression is held to infallibly match the form of the external object. (Perception itself is always “veridical.”)

The representationalism of John Locke takes perception to occur in two different stages: the causal chain that leads from the object to the observer, and the internal representation of the object by “ideas of sensation.” The initial causal chain triggers the representation, but is not actually present in the “direct perception,” which is all within the observer’s consciousness. The ideas of the primary qualities of an object “resemble” the primary qualities in the object, whereas the ideas of the secondary qualities do not “resemble” the secondary qualities in the object. This means that the perceptual experience of an object is a copy of the object’s primary qualities, mixed with subjective elements that are caused by the object’s properties, but not related to them in any other way.

The DTPR theory of Ayn Rand and other Objectivists takes perception to be an active process that generates a direct awareness of physical objects in reality. This awareness does not consist of a copy of any object’s form or qualities. It’s an experience of the object, from a certain perspective and in a certain form, that is not commensurable with the object itself. Any alteration in the causal chain that constitutes the perception process can alter the form in which the object is perceived. This theory has no place for Locke’s distinction between primary and secondary qualities, since it rejects “resemblances” between physical objects and sensory experiences.

For more on Aristotle’s theory of perception, see this article: Aristotle’s Psychology (SEP). For more on John Locke’s theory of perception, see Section 2.2. of this article: John Locke (SEP) and this page: A Guide to Locke’s Essay – Simple Ideas.

For more on Direct Transformative Process Realism, see the free online course, The Foundations of Knowledge, the audio course, Perception (ARI eStore), the book, How We Know, and/or the book, The Evidence of the Senses.


Related Posts:

Faith vs. Trust and Science vs. Religion

The Axioms of Objectivism

QuickPoint 5: Any Claim to a Probability is Also a Claim to a Certainty

The Arbitrary (from The Objectivism Seminar)

Ontology and The Problem of Universals: An Objectivist Comments

Values Are Relational, But Not Subjective

10 thoughts on “Philosophy of Perception: Naïve Realism vs. Representationalism vs. Direct Transformative Process Realism

  1. A use of the term “Naive Realism” in A. D. Smith’s Problem of Perception (2002): “the view that in perception we always perceive the world exactly as it is” (p. 22). Compare to your quote: “Perception gives a presentation of an object in a given situation as it gives it, and that’s it.” Smith also groups Aristotle, Aquinas, Locke and himself in the Direct Realist camp. I agree with Smith in that at least I think that primary qualities and the form of an object are equivalent, as the essence is the internal derivation from the form and secondary qualities are the internal processing of the object.

    Re illusion: “The perceptual experience itself does not automatically generate the conclusion that the patterns are rotating.” The patterns are rotating regardless of concepts. It’s an illusion, which means a partial misperception. Hence anything that involves a misperception, like a misjudgment, is itself based on the way we perceive. I’ve written more on this here:

  2. A comparison with the views of Thomas Reid (and followers) in the 18th and 19th centuries, and with the Oxford Realists of the early 20th century (such as Harold Prichard and Sir William David Ross) would be useful.

    • I agree this would be interesting and helpful, at least in regard to Reid. I may eventually write something comparing Rand to Reid.

      They are similar, in that Reid was a direct realist and–at least superficially–recognized the incommensurability of awareness with physical objects. But Reid has significant problems that Rand doesn’t.

      Basically, Reid kept a peculiar form of the primary/secondary quality distinction, whereas Rand does away with the distinction entirely. Reid still seems to think that the perception of primary qualities in objects presents those qualities “as they are in themselves,” as opposed to secondary qualities which present objects “as they are related to us.” I think that this distinction is a problem, because it won’t allow him to deal with certain illusions, such as simple magnification by a lens. (How can the “size” I’m seeing change when the object doesn’t?)

      The fact is that all perception is a process that produces an experience mediated by the relationship between the object and the observer. Every element of perception involves the relationship between the observer and object. We can only get at intrinsic properties by integrating multiple instances of perception into concepts, as Ayn Rand describes in Introduction to Objectivist Epistemology.

      Reid also seems to make the connection between sensations and conceptual ideas “automatic,” in that they are simply features of our “constitution.” This effectively makes simple concepts an automatic extension of perception: Both must be regarded together as the most basic foundation of knowledge. So then if we have a false belief about the world occasioned by a sensation–say that the sun is the same size as the moon–then we would have to say that our senses + constitution have failed us, and we can’t trust them: Our foundation for knowledge is unreliable, and errors will be automatic and inevitable.

      Unlike Reid, Rand does not treat basic concepts as automatic consequences of sensation/perception. Thus they don’t have to be treated as part of the ultimate foundation of knowledge. Perception alone is the infallible foundation, while concepts are developed from perception in a way that is potentially fallible. Conceptual errors are not automatic. Rather they are open to choices and to be rectified by reference to continued perception.

      • Yes there were some early 20th century American philosophers who were critical of Reid on these lines – not for being too much of a Realist, but for not being Realist enough.

  3. Would it be safe to say that DTPR considers the behavioural nature of mind/consciousness , while other two consider mind/consciousness as a separate entity from the behavior

  4. It is always risky to approach poets as if they were writing philosophy. However, in the case of William Blake I wonder what you think? He definitely had a ‘theory of knowledge’ – well more than a theory, more of a way of life! He famously wrote ‘if the doors of perception were cleansed everything would appear as it is, infinite’ which in my understanding is saying we approach life from too subjective a viewpoint; self-delusion comes into the picture and so on! He aslo blurrs the boundary between subject and object; so is very contemporary in this respect. His idea of Urizen, the over-rationalising self was a hindrance to seeing the world clearly as newly created phenomana in each moment. (Eternity in an hour!

  5. ” This effectively makes uncomplicated concepts an automatic propagation of sensing: Both must be regarded together as the most introductory origination of cognition.
    Basically, Reid kept a rum form of the main(a)/lower-ranking lineament eminence, whereas Rand does away with the eminence entirely.

  6. Pingback: Se comprendre, se rejoindre - latitude77

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s