Is Direct Realism a Convincing Theory of Perception?
Direct realism, properly understood, is a convincing theory of perception
- Direct realism is the view that the immediate objects of perception are mind-independent objects and their properties. 'Direct' means unmediated — there is no intermediary of any kind between the perceiver and the object. 'Realism' means those objects exist independently of any mind perceiving them - A distinction must be drawn between naive direct realism — that we directly perceive mind-independent objects and that those objects are always exactly as they appear — and sophisticated direct realism, which maintains that we do directly perceive mind-independent objects but does not require the world to be exactly as it appears. This distinction will do real philosophical work in what follows - The objections I will consider are perceptual variation, the argument from science, and the argument from hallucination. I will argue that none succeeds against sophisticated direct realism, and that direct realism is therefore a convincing theory of perception
Section 1: The Argument from Perceptual Variation
- The relational property response works because it shows P2 is false — the objection's key premise collapses. A property can vary with the observer while still belonging to the mind-independent object, because the physical conditions of observation are themselves objective facts about the world. A square table has the relational property of looking rectangular from 45 degrees even if no one is there to look at it — relational properties are grounded in the objective physical geometry of the object and its environment, not in any mind. This is not a definitional escape: it is a genuine claim about the structure of properties, and one that the objection gives us no reason to reject - The objection also implicitly assumes that if we cannot perceive the intrinsic properties of objects we cannot be perceiving the objects directly at all — but this assumption is unwarranted. Direct realism requires only that the immediate object of perception is the mind-independent object, not that every property we perceive is an intrinsic one - It is worth noting that the indirect realist's alternative — that what we perceive are sense data, patches of colour and shape from which we infer objects — is phenomenologically fictitious. As Wittgenstein observed, we do not perceive isolated sensory qualities and then deduce objects; we simply perceive the objects themselves. The sense data picture misrepresents the structure of ordinary perceptual experience
: The argument from perceptual variation
- P1: Direct realism holds that the immediate objects of perception are mind-independent objects and their properties - P2: If the properties we perceived were mind-independent, they would not vary with the perceiver - P3: Perceptual variation occurs — the same object produces different perceptions under different conditions. Locke's example: lukewarm water feels hot to a cold hand and cold to a warm hand simultaneously. The same object produces contradictory perceptions at the same moment - C: Therefore the properties we perceive are not mind-independent — direct realism is false
: Response — relational properties
- The direct realist distinguishes between intrinsic and relational properties of objects. Intrinsic properties are those an object has independently of any observer. Relational properties are properties an object holds in virtue of its relation to an observer under particular conditions — the property of feeling warmer than objects colder than it. Relational properties are genuine properties of the mind-independent object: they describe how the object really is in relation to certain physical conditions, not something produced by the perceiver's mind - P2 is therefore false: mind-independent properties can be perceptually variable so long as those properties are relational ones
Section 2: The Argument from Science
- The argument from science is a genuine and serious objection. It correctly shows that naive direct realism is false — the world at the microphysical level does not look like the world as we perceive it, and any theory that claims otherwise is empirically untenable - However the argument does not defeat sophisticated direct realism. The relational property move shows that perceptible properties like colour and warmth are grounded in the mind-independent object's physical structure — they are not invented by the perceiver's mind. The object and its relational properties together constitute what we perceive, and the immediate object of that perception is still the mind-independent thing - The cost is real but manageable: sophisticated direct realism must abandon the claim that objects are exactly as they appear and accept that our perception of secondary qualities is perception of relational rather than intrinsic properties. This is a significant concession from the naive version — but it is not a concession that undermines the core claim about immediacy and mind-independence
: The argument from science
- The scientific conception of physical objects conflicts with the direct realist's claim that we perceive mind-independent objects and their properties - At the microphysical level, objects are collections of particles. Those particles have no intrinsic colour, smell, warmth, or texture — these are not fundamental features of the mind-independent world but effects produced by the interaction of physical structures with perceivers - If the table has no intrinsic brownness at the microphysical level, then when we perceive its brownness we are not perceiving a mind-independent property of the object. Naive direct realism — that objects are exactly as they appear — is therefore false. And if it is false, the theory appears to require an intermediary after all: something mind-dependent that mediates between the microphysical object and the coloured world we actually perceive
: Response — naive vs sophisticated direct realism
- The argument succeeds against naive direct realism. If objects are not as they appear at the fundamental level, the naive claim that perception gives us the world exactly as it is cannot be sustained - However sophisticated direct realism survives. It concedes that objects are not exactly as they appear — the naive version is false — but maintains that the immediate object of perception is still the mind-independent object. The brownness of the table is a relational property: it is grounded in the object's microphysical structure and its interaction with light and the perceiver under standard conditions. Crucially, relational properties are constituted by facts about the microphysical object itself — they are not invented by the perceiver's mind but are fixed by the objective physical structure of the object and its environment. Perceiving a relational property is therefore still perceiving something grounded in the mind-independent object. No intermediary is required
Section 3: The Argument from Hallucination
- The first response directly undermines P2: if hallucinations are not perceptions but projections of the imagination, the argument cannot get started. The causal criterion is a principled and non-arbitrary basis for this distinction — what makes something a perception is precisely its causal origin in a mind-independent object, not its phenomenal character - The second response undermines the inference from C1 to C2. Even granting that hallucinations involve no mind-independent object, it does not follow that veridical perceptions do not. The move from 'hallucinations involve no mind-independent object' to 'therefore all perceptions involve no mind-independent object' relies on treating phenomenal similarity as evidence of ontological identity — which it is not - The disjunctive theory has a genuine limitation that must be honestly acknowledged: by categorising hallucination as not-a-genuine-perception it tells us relatively little about what hallucination actually is. The label does not constitute an explanation. Furthermore, committing to the causal criterion means direct realism can no longer define perception purely in terms of its phenomenal character — it must appeal to facts about causal origin that are not accessible from the inside. This is a real cost: it means direct realism cannot fully vindicate the first-person perspective it claims to respect. However it is a cost the theory can bear — a theory of veridical perception is not obliged to give a complete account of non-perceptual states, and the causal criterion is independently well-motivated - There is also positive introspective evidence for the theory that the hallucination objection does not touch. When we try to attend to our experience of seeing a table, we seem to be aware of the table itself rather than an experience of the table — experience is transparent, we see through it to the world. This transparency of experience supports the direct realist's claim that the immediate object of perception is the mind-independent thing, and no hallucination-based argument undermines it for the veridical case
: The argument from hallucination
- P1: According to direct realism, the immediate object of perception is a mind-independent object - P2: During a hallucination, a subject perceives an object p, but there is no mind-independent object p - C1: Therefore in hallucination the object of perception must be mind-dependent - P3: Hallucinatory experiences are phenomenally indistinguishable from veridical perceptions — what is perceived seems the same from the inside - C2: Therefore in all perceptions the immediate objects must be mind-dependent — direct realism is false - This objection is of a different character to the previous two. Perceptual variation and the argument from science left the basic structure of the perceptual relation — perceiver directly related to object — intact. Hallucination threatens that structure itself: if any perception can involve no mind-independent object at all, the claim that perception just is a direct relation to a mind-independent object becomes very hard to maintain. And via the indistinguishability premise, the objection generalises from edge cases to the entirety of perceptual experience
: First line of response — hallucinations are not genuine perceptions
- P2 can be questioned: a hallucination is not a genuine perception. Perception requires a causal relation between an external mind-independent object and the perceiver's sensory apparatus — hallucination involves no such relation. If hallucinations are not perceptions, they are not counterexamples to a theory of perception
: Second line of response — phenomenal indistinguishability does not entail ontological identity
- P3 can also be questioned. The indistinguishability premise establishes only an epistemic fact — we cannot tell hallucination and veridical perception apart from the inside — but C2 draws an ontological conclusion: that they must involve the same type of immediate object. This is a non sequitur - The disjunctive theory of perception formalises this: veridical perception and hallucination belong to categorically different psychological kinds that merely resemble each other phenomenally. In veridical perception the immediate object is the mind-independent thing itself; in hallucination there is no genuine perceptual object — it is a distinct kind of event that mimics perception from the inside - The common factor principle — that subjectively indistinguishable experiences must share the same immediate object — is therefore false. Two experiences can be phenomenally alike while being ontologically distinct
- The through-line of the essay is the naive/sophisticated distinction. Naive direct realism does not survive — the argument from science defeats it and the argument from hallucination puts serious pressure on it. But sophisticated direct realism, which abandons the claim that objects are exactly as they appear while maintaining that the immediate object of perception is the mind-independent thing, survives all three objections - One might conclude from the disjunctive theory's limitation that direct realism is only partially convincing — convincing for veridical perception but silent on hallucination. This partial verdict is available but not compelling: a theory of perception is assessed on its account of perception, and on that account sophisticated direct realism stands - Direct realism, properly understood, is a convincing theory of perception