The Problem of Perception

Perception appears to place subjects in direct contact with a mind-independent world. In seeing a tree, one seems aware of the tree and its colours, shape, and distance, not of an inner image from which the tree is inferred. Illusion and hallucination disturb this simple picture. An experience can present an object as having a property it lacks, and a subject can undergo an experience subjectively indistinguishable from perception when no corresponding object is present.

The problem of perception asks how veridical perception, illusion, and hallucination are related. Do they share one fundamental mental state whose accuracy varies, or does genuine perception have a world-involving nature absent from hallucination? The answer divides sense-data theories, adverbialism, intentionalism, and disjunctivist or naive realist views.

Three cases

Veridical perception. A subject sees a red apple, and an apple with the presented properties is before them. The experience is accurate and appropriately caused.

Illusion. A real object is perceived but appears to have a property it lacks: a straight stick looks bent in water, a white wall looks yellow under illumination, or an object appears smaller at a distance.

Hallucination. An experience occurs without the ordinary perceived object. It may be indistinguishable to the subject from a veridical perception.

These categories concern success, not merely phenomenology. An illusion can be phenomenally ordinary once lighting and context are understood; a hallucination can be subjectively vivid. A theory must explain both how the cases differ in world-involvement and why they can resemble one another from within.

The argument from illusion

A traditional argument begins with variable appearance:

1. In illusion, an object appears to have a property it does not have.
2. The subject is directly aware of something possessing the apparent property.
3. That item cannot be the ordinary object, since the object lacks the property.
4. The same kind of immediate object is present in veridical perception.
5. Therefore subjects directly perceive sense-data or appearances, and perceive ordinary objects only indirectly.

The second premise is controversial. From "the stick looks bent" it does not follow that the subject is aware of a bent item. Looking bent may be a way the straight stick is presented under refraction. A representational state can be inaccurate without an object instantiating its represented property.

The move from illusion to all perception is also substantial. Even if illusion involves an intermediary, why must veridical perception have the same structure? The answer usually appeals to a common factor shared by subjectively indistinguishable cases.

The argument from hallucination

Hallucination sharpens the problem because no ordinary object need be present:

1. A hallucination can be introspectively indistinguishable from a veridical perception.
2. Indistinguishable experiences have the same immediate mental nature.
3. In hallucination, the subject is not directly related to an ordinary external object.
4. Therefore the veridical experience is not fundamentally a direct relation to such an object either.

Premise 2 is the common-kind assumption. It is natural if phenomenal character is wholly internal and shared by duplicates. It is denied by disjunctivism: two experiences can be indiscriminable without sharing a positive fundamental nature, just as two objects can be visually indiscriminable without sharing composition.

Rejecting common kind preserves direct perception but creates an explanatory demand. Why are the cases indiscriminable, and how can psychology generalise across them?

The causal argument

Perception involves a causal chain: light reflects from an object, stimulates receptors, and produces neural processing. The same proximal neural state could, in principle, be produced without the distal object through stimulation or abnormal activity. If mental state is fixed by the proximal cause, veridical perception and hallucination share an internal state.

This causal argument supports a common factor more strongly than mere imaginative indistinguishability. Yet it assumes that the neural state wholly fixes the perceptual state. A relationalist can say the same neural condition constitutes different mental events depending on whether an object stands in the right relation, much as internally similar movements can be a successful action or an attempt.

The causal chain therefore establishes physical dependence, not by itself the metaphysics of perception.

Direct and indirect realism

Direct realism says ordinary objects and their properties are the immediate objects of perception. Perceptual contact is not mediated by an item that is itself perceived.

Indirect realism says subjects directly encounter sense-data, impressions, or inner representations and perceive the external world through them. The view explains a common phenomenal factor easily: the same intermediary can occur with or without its normal cause.

"Direct" can be misleading. A direct realist accepts retinal and neural mediation in the causal process; the claim is not that perception occurs without mechanisms. The issue is whether those mechanisms produce an inner object that is what awareness immediately concerns, or realise a relation or representation directed at the world.

Likewise, a representationalist may count as direct if the subject perceives the apple by being in a state that represents it, without first perceiving the representation. The contrast is metaphysical and epistemic, not a count of causal stages.

Sense-data

A sense-datum is a mind-dependent object bearing the properties an experience immediately presents. In veridical perception it is caused by and resembles the ordinary object; in hallucination a similar datum occurs without that object.

This gives all three cases one structure and guarantees an object for awareness. Its cost is a veil of perception. If subjects are directly aware only of sense-data, their evidence for external objects is mediated, and scepticism threatens. The theory also duplicates the visible world with private objects whose identity and causal role are obscure.

The next page treats sense-data and their classical non-objectual rival, adverbialism, in detail.

Intentionalism

Intentionalism treats perception as a representational state with accuracy conditions. A veridical perception accurately represents an object; illusion attributes the wrong property; hallucination represents an object or scene that is absent. No inner coloured item is required.

This supplies a common mental factor while preserving world-directed content. It fits the representational theory of consciousness, though one can hold that perception represents without reducing phenomenal character to content.

Its central tasks are explaining how content is fixed, why some representations are conscious, and whether content exhausts phenomenal differences. If a hallucination and perception share content, their different relation to reality is not part of their narrow mental nature.

Adverbialism

Adverbialism denies that experience involves either an external object in every case or an inner object. To experience a red circle is to sense redly and circularly: the subject's act is modified in a certain way.

This avoids sense-data while preserving a common internal state across perception and hallucination. It struggles with complexity. Sensing redly and circularly could describe a red circle, a red square beside a blue circle, or two differently organised fields. Without object-like structure, the theory needs resources to bind qualities into the right experienced arrangement.

Disjunctivism and naive realism

Naive realism or relationalism says veridical perception is fundamentally a relation in which ordinary objects and properties partly constitute the experience. Remove the object and that very experience cannot occur.

Disjunctivism says an apparently unified category such as "an experience as of a red apple" contains fundamentally different disjuncts:

$$\text{either seeing a red apple}\vee \text{merely seeming to see one}.$$

The view rejects the inference from indistinguishability to a common positive mental nature. In the good case, the world itself explains phenomenal and epistemic contact.

Its challenge is the bad case. Describing hallucination merely as indiscriminable from seeing can seem negative and incomplete. Causally matching hallucinations also appear to support common psychological explanations. Disjunctivists differ over whether they allow common representational or neural properties while denying a common fundamental experiential kind.

Perception and knowledge

The metaphysics affects epistemology. If veridical perception and hallucination share the same evidence from the subject's point of view, how can perception provide knowledge rather than only a justified guess? Indirect and common-factor theories often answer through reliability: the same type of state usually has an appropriate external cause.

Epistemological disjunctivists say a subject who sees that $p$ can possess factive evidence - the fact that $p$ itself - unavailable in hallucination. The experiences may be indiscriminable to the subject without supplying the same epistemic position.

The benefit is a strong answer to scepticism; the cost is explaining why reflective access cannot reveal the difference in evidence. Metaphysical disjunctivism about experience and epistemological disjunctivism about reasons are related but not identical.

Perceptual constancy

Objects project changing sensory patterns as distance, angle, and illumination vary, yet are experienced as relatively constant in size, shape, and colour. This perceptual constancy shows that perception is not a passive copy of proximal stimulation.

Intentionalists treat constancy as representation of stable distal properties through variable sensory information. Naive realists treat it as skilled access to objects across conditions. Predictive and enactive accounts emphasise expectations and sensorimotor dependencies. Sense-data theories must explain why changing private items present one stable object.

Constancy does not decide the metaphysics, but it imposes a constraint: perception is an achievement of a system embedded in an environment, not a sequence of inner snapshots.

The decision structure

The problem of perception is not solved by saying that perception is constructed or brain-dependent. Every view accepts causal processing. The dispute is what the resulting mental episode fundamentally is and whether the external object enters its nature.

Selected references

• Ayer, A. J. The Foundations of Empirical Knowledge (1940).
• Austin, J. L. Sense and Sensibilia (1962).
• Fish, William. Philosophy of Perception (2010).
• Martin, M. G. F. "The Limits of Self-Awareness" (2004).
• Snowdon, Paul. "Perception, Vision and Causation" (1980-81).