Perception as a source of knowledge

Direct realism and the immediate object of perception#

FoundationAQA 7172 3.1 Epistemology — Perception as a source of knowledgeDfE GCE Philosophy subject content

The theories of perception

Fig. 1The three theories differ over whether perception is direct and whether its objects are mind-independent.

Direct (or naive) realism is the common-sense theory of perception: the immediate objects of perception are ordinary mind-independent physical objects and their properties. When I see a tomato, what I am directly aware of is the tomato itself — its redness, its roundness — not some intermediary. The world I perceive exists independently of any mind, and I perceive it directly, without the mediation of ideas, images or sense-data. This is the default view most people hold before doing philosophy, and its great strength is precisely its fidelity to how perception seems and to the way we talk: I see the cat, not a picture of the cat.

The theory has two components worth separating. The realism component says the objects of perception are mind-independent — they exist and have their properties whether or not anyone perceives them, and would continue to exist in an unobserved room. The directness component says these mind-independent objects are the IMMEDIATE objects of awareness — there is no perceptual intermediary standing between the perceiver and the world. Indirect realism will keep the realism but drop the directness; idealism will drop the realism. Locating a theory on these two axes is the key to the whole topic.

Direct realism enjoys three main advantages. It is the simplest theory (it posits no additional entities such as sense-data), it accords with common sense and ordinary language, and it makes knowledge of the external world unproblematic: since we perceive the world directly, there is no gap between appearance and reality for scepticism to exploit. It also fits the phenomenology of perception, which does not present itself as the viewing of inner pictures but as open contact with things. For these reasons contemporary direct realists (such as John McDowell and, in a different key, the disjunctivists) have tried hard to defend a sophisticated version of the view.

The difficulty is that perception is not always veridical, and the non-veridical cases seem to show that the immediate object of awareness cannot always be a mind-independent object as it is. Illusions (a straight stick looks bent in water), perceptual variation (a round coin looks elliptical from an angle), hallucinations (seeing something that is not there) and the time-lag (we see long-dead stars) all suggest that what we are immediately aware of can differ from any physical object's actual properties — and if so, the immediate object of awareness is not, or not always, the physical object. These arguments, reconstructed next, are the classic route from direct realism to its rivals.

Worked example

Placing a theory on the two axes

A theory says we are directly aware of the physical world, which exists independently of us. Which theory is it, and how does it differ from indirect realism?

Test directness
'Directly aware of the physical world' — no intermediary — so the theory affirms directness.
Test realism
'Exists independently of us' — mind-independent objects — so the theory affirms realism.
Identify
Affirming both directness and realism is direct realism.
Contrast
Indirect realism keeps realism but denies directness: we are immediately aware of mind-dependent sense-data that represent the physical world.

Result: The theory is direct realism; it differs from indirect realism precisely on directness — whether the immediate object of awareness is the physical thing itself or an intervening sense-datum.

Typical mistakes

Equating direct realism with 'the world is exactly as it appears' — the sophisticated view allows appearances to vary while insisting the direct object is still the physical thing.
Confusing direct realism with idealism because both are 'about what we perceive directly'; direct realism is realist (mind-independent objects), idealism is not.

Active revision

Outline direct realism and explain one advantage it has over indirect realism. (5-mark AO1 response)

Active recall

Recall the key points — then reveal.

Sources: GCE AS and A level subject content for philosophy (Department for Education)

The arguments from illusion and perceptual variation#

StandardAQA 7172 3.1 Epistemology — Perception as a source of knowledgeDfE GCE Philosophy subject content

The argument from illusion

Fig. 2From the bent-looking stick to the conclusion that the immediate object of perception is a sense-datum.

The argument from illusion is the classic challenge to direct realism. Its structure is: (P1) in cases of illusion — a straight stick half-submerged in water looks bent — we are aware of something that has the property we perceive (something bent); (P2) but the physical object (the stick) does not have that property (it is straight); (P3) so the immediate object of awareness in the illusion is not the physical object, but something else, which philosophers call a sense-datum (a mind-dependent bearer of the perceived properties); (P4) illusory and veridical perceptions can be subjectively indistinguishable — the bent-looking stick can seem just like a really bent stick — so, by parity, we must be aware of a sense-datum in the veridical case too; therefore (C) the immediate object of perception is always a sense-datum, and direct realism is false.

The argument from perceptual variation, associated with Bertrand Russell in The Problems of Philosophy, makes the same point without invoking error. A round coin looks elliptical from an angle; a white wall looks yellow in evening light; the same table looks different colours and shapes from different positions. Russell argues that the 'real' shape and colour of the table are not what we immediately see — what we immediately see varies continuously with our viewpoint and the lighting, while the physical table does not change. So the immediate objects of perception (the varying apparent shapes and colours) cannot be identical with the constant physical object; they are sense-data, and the physical table is inferred, not directly seen.

The move from (P3) to the conclusion — the 'generalising step' (P4) — is where the argument does its heavy lifting and where it is most contestable. It reasons that because illusory and veridical experiences can be qualitatively identical 'from the inside', they must have the same kind of immediate object; so if the illusory object is a sense-datum, so is the veridical one. This is sometimes called the 'phenomenal principle': if there sensibly appears to be something with a property F, then there is something of which the subject is aware that really is F. Grant the phenomenal principle and the argument is powerful; deny it and direct realism can survive.

Direct realists have two main replies. The sense-datum reply can be resisted by giving an adverbial or intentional account of illusion: in seeing the bent-looking stick I am not aware of a bent OBJECT but am 'appeared to bently' — the bentness characterises the WAY I perceive, not an inner object that is bent. Alternatively, disjunctivism (McDowell, Hinton) denies (P4) by holding that veridical perception and illusion/hallucination are fundamentally different mental states that merely seem the same, so nothing follows about the veridical case from the illusory one. Both replies aim to block the generalising step and keep the physical object as the direct object of veridical perception; whether they succeed is a central evaluative question.

Worked example

Testing the generalising step

Assess the inference from 'we are aware of a sense-datum in illusion' to 'we are aware of a sense-datum in veridical perception'.

State the step
Because illusory and veridical experiences can be subjectively identical, they are claimed to have the same kind of immediate object (the phenomenal principle).
The adverbial reply
Deny that illusion requires an inner bent object: I am 'appeared to bently', so no sense-datum is needed even in illusion, blocking the step at its source.
The disjunctivist reply
Deny that sameness of appearance entails sameness of object: veridical seeing and mere appearance are different states, so the veridical case need not inherit the illusory case's object.
Weigh
If either reply holds, direct realism survives; the sense-datum theorist must defend the phenomenal principle, which is not self-evident.

Result: The argument from illusion is only as strong as its generalising step; adverbial and disjunctivist accounts give the direct realist a principled way to resist it, so the argument is powerful but not decisive.

Typical mistakes

Treating the argument from illusion as merely 'the senses sometimes deceive us' — the philosophical claim is about the immediate OBJECT of awareness, not about the reliability of the senses.
Forgetting the generalising step: from illusion alone nothing follows about veridical perception unless the two are assimilated (P4), so an answer that omits it misstates the argument.

Active revision

Explain the argument from illusion and one direct realist response to it. (12-mark AO1 response)

Active recall

Recall the key points — then reveal.

Sources: GCE AS and A level subject content for philosophy (Department for Education)

Indirect realism and sense-data#

StandardAQA 7172 3.1 Epistemology — Perception as a source of knowledgeDfE GCE Philosophy subject content

Primary and secondary qualities

Fig. 3Ideas of primary qualities resemble the object; ideas of secondary qualities are caused by powers but do not resemble.

Indirect (representative) realism accepts the conclusion of the argument from illusion: the immediate objects of perception are mind-dependent sense-data, but these are caused by, and represent, mind-independent physical objects. John Locke is the classic proponent. On his view perception has a two-stage structure: physical objects act on our sense organs and cause 'ideas' (sense-data) in the mind; we are immediately aware of these ideas and, through them, indirectly aware of the external objects that cause them. The theory keeps realism (the external world exists mind-independently) but drops directness (we reach the world only via a representational intermediary).

Locke sharpens the view with the distinction between primary and secondary qualities. Primary qualities — extension, shape, size, motion, number, solidity — are qualities the object has in itself; our ideas of them resemble the qualities in the object. Secondary qualities — colour, sound, taste, smell, temperature — are powers in the object (arrangements of its primary qualities) to produce sensations in us; our ideas of them do not resemble anything in the object, which is not itself coloured or warm in the way our sensation is. This explains perceptual variation neatly: the coin's real (primary) shape is constant, while the (secondary-quality-like) apparent shape varies with viewpoint, and the felt warmth of lukewarm water differs for a cold and a warm hand because warmth is a power, not an intrinsic quality.

The theory's advantages are that it accommodates illusion, hallucination, perceptual variation and the time-lag without strain (all are facts about our sense-data and the causal process), and that it fits the scientific picture, on which perception is a causal chain from object to sense organ to brain. But it incurs a heavy cost: if we are only ever immediately aware of our own sense-data, the mind-independent world lies behind a 'veil of perception', and it becomes a genuine question how we can know it exists or what it is like. This is the sceptical problem the next section develops, and it is the price indirect realism pays for accepting the argument from illusion.

A related objection presses on the notion of representation itself. If I am aware only of my ideas, I can never step outside them to compare an idea with the object it supposedly resembles, so I cannot verify that my primary-quality ideas resemble anything or that my sense-data represent an external cause at all. Berkeley pushed this hard: the very idea that an idea could resemble a mind-independent, unperceivable object is, he argued, incoherent — 'an idea can be like nothing but an idea'. Indirect realism thus stands on a knife-edge: it is driven to sense-data by the argument from illusion, but sense-data threaten to cut us off from the world it wants to be realist about, opening the door both to scepticism and to Berkeley's idealism.

Worked example

Explaining perceptual variation with Locke's distinction

Lukewarm water feels warm to a cold hand and cool to a warm hand. Use indirect realism to explain this.

Classify the quality
Felt temperature is a secondary quality — a power of the water (its molecular motion) to cause a sensation, not an intrinsic warmth.
Locate the variation
The sensation depends on the relation between the water and the perceiver's hand, so it can differ between hands without the water changing.
Preserve the object
The water's primary qualities (its actual temperature/energy) are constant; only the secondary-quality SENSATION varies.
Draw the moral
We are immediately aware of the sensation (sense-datum), and infer the constant physical cause — indirect realism at work.

Result: The case is explained without contradiction: variation lies in the secondary-quality sensations we are directly aware of, while the mind-independent water, characterised by its primary qualities, stays fixed.

Typical mistakes

Saying secondary qualities are 'not real' — for Locke they are real POWERS in the object; what fails to resemble is our idea of them, not their existence.
Describing indirect realism as anti-realist; it is realist about the external world — the mind-dependence is only of the sense-data, not of the physical objects.

Active revision

'Indirect realism cannot be defended against scepticism about the external world.' Assess this claim. (25-mark essay)

Active recall

Recall the key points — then reveal.

Sources: GCE AS and A level subject content for philosophy (Department for Education)

The veil of perception and scepticism#

AdvancedAQA 7172 3.1 Epistemology — Perception as a source of knowledgeDfE GCE Philosophy subject content

Responses to the veil of perception

Fig. 4Realists infer the world (causally or abductively); idealism dissolves the gap by denying mind-independent objects.

The gravest objection to indirect realism is that it traps us behind a 'veil of perception'. If the only things I am ever immediately aware of are my own mind-dependent sense-data, then I never directly encounter the external world; I only ever have the ideas. But then two sceptical questions arise that I seem unable to answer from within my experience: how do I know there is any mind-independent world causing my sense-data at all (rather than, say, a demon or a dream), and how do I know it is anything like the way my sense-data represent it? Since I cannot get outside my ideas to check them against the world, indirect realism appears to make knowledge of the external world impossible — the very knowledge it wanted to secure.

Locke's own response is a causal one: our ideas must have some external cause, since they come to us involuntarily and in an orderly, coherent way we do not choose, which is best explained by mind-independent objects acting on us. He also appeals to the coherence of different senses (sight and touch agree about shape) and to the involuntariness of perception (I cannot help seeing the page). But the sceptic replies that a dream or a deceiving demon could produce equally coherent, involuntary experiences, so coherence and involuntariness do not entail an external physical cause; Locke's inference is at best probable, not certain.

The strongest realist response is abductive, developed by Bertrand Russell: the existence of the external world is not something we can prove, but it is the best hypothesis — the simplest and most powerful explanation of the patterns in our sense-data. That the cat appears in the doorway, then later in the middle of the room, is far more simply explained by a persisting mind-independent cat that moved while unobserved than by a series of disconnected sense-data popping in and out of existence. On this view knowledge of the external world is like a well-confirmed scientific theory: not certain, but rationally compelling as an inference to the best explanation, and vastly better than the sceptical alternatives.

Evaluation turns on how much this abductive defence concedes. The realist grants that the external world cannot be known with certainty and can only be inferred, which is exactly what the direct realist warned against: by accepting the argument from illusion and interposing sense-data, indirect realism created the sceptical gap in the first place. A direct realist will say this shows the wrong turn was taken earlier — better to block the argument from illusion (adverbially or disjunctively) and keep the world as the direct object of perception, than to accept sense-data and then labour to reach the world by inference. The idealist, by contrast, will say the lesson is that the mind-independent world was a confused posit all along, and should be dropped. Which response is best is the topic's central evaluative fork, and it links directly to 'the limits of knowledge', where the same sceptical structure recurs in Descartes.

Worked example

Model 25-mark plan: does indirect realism lead to scepticism?

Plan an essay assessing whether indirect realism leads to scepticism about the external world.

Thesis
Indirect realism does create a sceptical gap, and Russell's best-explanation reply narrows but does not close it, so the theory is defensible only at the cost of certainty.
AO1
Explain indirect realism, sense-data, and the veil-of-perception problem (two sceptical questions).
The realist replies
Locke's causal argument and Russell's inference to the best explanation; the external world explains the coherence and involuntariness of experience.
The sceptic's rejoinder
A dream/demon could produce coherent, involuntary experience, so the inference is probabilistic, not certain — the gap remains.
Judgement
Conclude that the abductive reply makes external-world belief rational but not certain; whether that is enough depends on whether knowledge requires certainty (link to reliabilism).

Result: A defensible judgement: indirect realism does invite scepticism, but the best-explanation response secures rational, if not certain, knowledge — unless one demands certainty, in which case direct realism or idealism is preferable.

Typical mistakes

Treating Russell's best-explanation defence as a proof; it is an inference to the best explanation that explicitly gives up certainty.
Assuming the veil objection refutes realism as such; it targets INDIRECT realism specifically — direct realism and idealism avoid it in opposite ways.

Active revision

'The best response to scepticism about the external world is that it is the best hypothesis.' Discuss. (25-mark essay)

Active recall

Recall the key points — then reveal.

Sources: GCE AS and A level subject content for philosophy (Department for Education)

Berkeley's idealism#

AdvancedAQA 7172 3.1 Epistemology — Perception as a source of knowledgeDfE GCE Philosophy subject content

The master argument

Fig. 5Berkeley argues that trying to conceive an unperceived object is self-refuting, so esse est percipi.

George Berkeley accepts that we are immediately aware only of ideas, but draws the radical conclusion that there are no mind-independent material objects at all: to be is to be perceived (esse est percipi). Ordinary objects — tables, trees, apples — are not unperceivable material substances but collections of ideas (sensible qualities) that exist only in so far as they are perceived by a mind. Idealism thus abolishes the veil of perception by abolishing what lies behind it: there is no gap between our ideas and the world because the world just IS ideas, so perception cannot misrepresent a hidden material reality.

Berkeley's central positive argument is the 'master argument': try to conceive of an object existing wholly unperceived — say, a tree in a forest that no one perceives. Berkeley claims that in the very act of conceiving it, you are perceiving (imagining) it, so you cannot actually conceive of a genuinely unperceived object; the supposition is self-refuting. He reinforces this with an attack on the primary/secondary quality distinction: the arguments from perceptual variation apply just as well to shape and size (primary qualities) as to colour and heat, so if secondary qualities are mind-dependent, so are primary ones — there is no residue of mind-independent 'primary' reality for matter to consist in. And, he argues, the notion of 'matter' as an unperceivable, inert substrate is incoherent, since an idea can resemble nothing but another idea.

The obvious objection is that idealism cannot explain the continued, orderly existence of objects when no human perceives them, nor the involuntariness and regularity of experience, without collapsing into solipsism (only my mind and its ideas exist). Berkeley's answer is God: objects continue to exist because they are always perceived by God, the infinite spirit; and the orderly, involuntary ideas of 'sense' (as opposed to the fainter ideas of imagination, which I do control) are ideas God causes in us according to regular laws, which just are the 'laws of nature'. So the coherence of experience that Russell explained by a mind-independent world, Berkeley explains by the constant perceiving and willing of God. This also lets him claim to save common sense: the table really is there when I leave the room, because God perceives it.

Evaluation weighs idealism's strengths — it takes the sense-datum starting point to its logical end, dissolves external-world scepticism, and offers a unified explanation of order — against serious objections. First, the master argument is widely thought to trade on an ambiguity: to conceive of an unperceived tree is not to perceive a tree but to have a thought whose CONTENT is an unperceived tree, so the argument confuses the act of conceiving with its object. Second, the appeal to God looks like an unearned deus ex machina: it imports a vast theological commitment to explain perceptual order, and it is unclear why God's perceiving should make objects mind-independent enough to count as the very tables and chairs of common sense. Third, idealism struggles with illusion and hallucination: if to perceive just is for the object to exist, it is hard to say what distinguishes a veridical perception from a hallucination, since both are collections of ideas. A strong judgement will grant idealism's ingenuity in closing the sceptical gap while questioning whether the cure — abolishing matter and invoking God — is more incredible than the disease.

Worked example

Assessing the master argument

Reconstruct and evaluate Berkeley's master argument for idealism.

Reconstruct
P1: to show matter is conceivable, conceive an object existing unperceived. P2: but in conceiving it you are perceiving/imagining it. C: so you cannot conceive a genuinely unperceived object; esse est percipi.
The act/object reply
Conceiving of an unperceived tree is a mental ACT whose CONTENT is 'an unperceived tree'; the act's occurring does not make its object perceived — Berkeley conflates the two.
Assess the reply
The reply seems decisive: I can think ABOUT unperceived things (the far side of the moon before probes) without those things thereby being perceived.
What survives
Even if the master argument fails, Berkeley's attack on the primary/secondary distinction and the incoherence of unperceivable matter remain to be answered.

Result: The master argument fails because it confuses the act of conceiving with its object; idealism must rest instead on the harder arguments against mind-independent matter, which is where the real debate lies.

Typical mistakes

Reading Berkeley as denying that objects exist — he affirms tables and trees exist, but as collections of ideas, not as mind-independent matter.
Presenting idealism as obviously absurd; its point is that it EXPLAINS the same data as realism (order, continuity) while dissolving scepticism, so it must be argued against, not merely mocked.

Active revision

'Berkeley's idealism is the best theory of perception because it defeats scepticism.' Is this claim true? Defend your answer. (25-mark essay)

Active recall

Recall the key points — then reveal.

Sources: GCE AS and A level subject content for philosophy (Department for Education)

Which theory of perception succeeds?#

AdvancedAQA 7172 3.1 Epistemology — Perception as a source of knowledgeDfE GCE Philosophy subject content

Trade-offs among the three theories

Fig. 6Each theory secures one virtue (phenomenology, illusion-handling, anti-scepticism) at a characteristic cost.

A final judgement must score the three theories against the data any theory of perception must handle: the phenomenology of open contact with the world, the facts of illusion, hallucination and perceptual variation, and the demand to secure knowledge of a mind-independent reality. Direct realism honours the phenomenology and blocks scepticism but must explain away the non-veridical cases; indirect realism handles those cases and fits the causal science of perception but opens the sceptical veil; idealism closes the veil but at the price of abolishing matter and invoking God. Each theory buys a virtue with a corresponding cost, and the evaluation is a matter of which trade is least bad.

The pivotal issue is the argument from illusion and its generalising step. If the direct realist can block that step — via an adverbial theory of appearing or via disjunctivism — then the motivation for sense-data collapses, the veil never descends, and direct realism keeps both the phenomenology and mind-independent knowledge, which is a powerful combination. If the step cannot be blocked, sense-data are forced upon us and the choice narrows to indirect realism (accept the veil, defend the world abductively) or idealism (drop the world behind the veil). So a disciplined essay traces the whole debate back to whether the phenomenal principle is true.

A defensible contemporary verdict favours a sophisticated direct realism (or disjunctivism), on the grounds that it best respects the datum that perception acquaints us with the world itself, and that the adverbial/intentional treatment of illusion is independently plausible — illusions are cases of the physical object LOOKING other than it is, not cases of our inspecting a real inner bent object. Indirect realism's fit with the causal science of perception is genuine but does not by itself settle the philosophical question of what we are AWARE of, since a causal process can deliver direct awareness of its distal object (as disjunctivists insist). Idealism, though the most ingenious answer to scepticism, asks us to give up too much of the ordinary and scientific picture.

That said, the verdict is defeasible and a strong candidate acknowledges the best case for each rival. Indirect realism can reply that only a representational intermediary explains hallucination (where there is no external object to be directly aware of) and the time-lag (we see the sun as it was eight minutes ago, so cannot be directly aware of it as it now is). Idealism can reply that 'mind-independent matter' has never been coherently specified and that its own God-based order is no more mysterious than the realist's brute external world. The mark of a first-class answer is that it reaches a judgement, states the criterion it is applying (fidelity to phenomenology, explanatory power, or anti-sceptical strength), and shows why, on balance, one theory pays the lowest total cost.

The topic also feeds forward. The sceptical structure that indirect realism exposes — awareness of inner states, doubt about their external cause — is exactly Descartes' starting point in 'the limits of knowledge', where the dreaming and demon arguments generalise it to global doubt. And the empiricist commitment that all our ideas trace to sense experience, in play here in Locke and Berkeley, is developed in 'reason as a source of knowledge', where concept empiricism and the copy principle are examined. Perception is therefore the hinge on which much of the epistemology unit turns.

Worked example

Weighing the theories against a criterion

Explain how a candidate can reach a defensible judgement on the best theory of perception.

Fix the criterion
Decide what a theory of perception must above all deliver — fidelity to experience, explanation of illusion, or knowledge of the world — and state it.
Trace to the key move
Ask whether the argument from illusion's generalising step succeeds; if it can be blocked, direct realism survives; if not, sense-data are forced on us.
Apply and compare
Score direct realism (phenomenology, anti-scepticism), indirect realism (illusion, science, but the veil), idealism (anti-scepticism, but no matter).
Judge
Conclude for the theory that pays the lowest total cost on the chosen criterion, and acknowledge the strongest rival reply.

Result: A first-class judgement fixes a criterion, hinges the comparison on the argument from illusion, and defends one theory as paying the least — typically a sophisticated direct realism — while conceding indirect realism's grip on hallucination and the time-lag.

Typical mistakes

Concluding 'they all have problems, so it is a matter of opinion' — AO2 rewards a weighed judgement against a stated criterion, not a stalemate.
Evaluating the theories in isolation without connecting them through the argument from illusion, which is the thread that determines whether sense-data are forced on us.

Active revision

'Direct realism is the most defensible theory of perception.' Discuss. (25-mark essay)

Active recall

Recall the key points — then reveal.

Sources: GCE AS and A level subject content for philosophy (Department for Education)

Key points

Placing a theory on the two axes

Key points

Testing the generalising step

Key points

Explaining perceptual variation with Locke's distinction

Key points

Model 25-mark plan: does indirect realism lead to scepticism?

Key points

Assessing the master argument

Key points

Weighing the theories against a criterion

Sources

Perception as a source of knowledge

Key points

Placing a theory on the two axes

Key points

Testing the generalising step

Key points

Explaining perceptual variation with Locke's distinction

Key points

Model 25-mark plan: does indirect realism lead to scepticism?

Key points

Assessing the master argument

Key points

Weighing the theories against a criterion

Sources