Direct realism about perception

Clarendon

To perceive something is to be in unmediated contact with it. I take that to be a conceptual truth that all involved in this debate will agree on.

With that in mind, a 'direct realist' is someone who holds that we are sometimes perceive the mind external world. That is, when I look at the ship I am directly aware of the ship itself. Thus, I perceive the ship.

This is as opposed to indirect realists who hold that we are only directly aware - and so only perceiving - mental states of our own, rather than the world out there.

I think most contemporary philosophers will want to describe themselves as direct realists of one sort or another. But having read some of their views and overviews of their views, it seems to me that most of those who call themselves direct realists are not really direct realists at all.

The point they typically make at first is to note that when we have a visual sensation of a ship, it is not the visual sensation that we perceive, but the ship itself by means of it. The mistake they accuse indirect realists of making is to confuse a 'vehicle' of awareness with an 'object' of awareness. Fair enough: that does seem to be the mistake the indirect realists are making (one of them, anyway).

What they then do is point out that the visual experiences have 'content' - that they are 'about' a ship or 'represent' there to be a ship. And they then think that this is somehow enough to secure the direct contact with the ship needed for that experience to constitute a perception of the ship.

That's where I get puzzled. Fair enough that the indirect realists are making a mistake. But it seems to me that these direct realists are making one too. As in order to secure direct contact with the mind-external ship, the experience would surely have literally to contain the ship. It's not enough that it's 'about' a ship. A note about a ship is about a ship, but it can't thereby be a means by which we perceive a ship. A thought about a ship is about a ship, but again one can't perceive a ship by thinking about a ship. So it won't help at all to make a view 'direct' just to focus on the way in which a sensation is 'about' or 'of' a ship. The sensation would have to include the ship itself.

Some direct realists acknowledge that in order to perceive a ship the perceptual relation needs to put us in direct contact with the ship - so acknowledge that the ship itself must be included in our perceptual experience. But they insist that the ship itself is somehow a constituent of the experience.

But I can make no sense of that idea. An experience is mental, so how can it include an actual ship? It's like proposing that the number 5 has a door in it - it just makes no sense.

So I think that currently direct realists are either guilty of being indirect realists in disguise (if they admit that experiences do not contain actual ships and such like but are mediators between us and ships), or they have an incoherent view (that mental experiences can somehow incorporate ships, even though a ship is a thing and a mental experience is a state of a thing).

I think that direct realism 'proper' would have to be the view that perceptual relations have 2 and only 2 relata: the perceiver and the perceived. That is, no mental experience features as a relata within it (for then you automatically get indirect realism). That doesn't mean that there is no mental experience associated with perception - for clearly there is, as there's something it is like to perceive a ship - but that the experience is 'of' the perceptual relation, rather than a constituent part of it. It might even be the case that it is invariably the case that there is an experience of the relation whenever the relation obtains. The point would be that to experience perceiving and perceiving are distinct nevertheless

I am interested in hearing any objections to this 'proper' form of direct realism - perhaps it is not coherent or perhaps it has unacceptable implications. (I am not interested in defending indirect realism - my interest is in investigating the viability of direct realism so only mentioned indirect realism because I think other forms of direct realism collapse into it).

jkop

↪Clarendon

Objections to direct realism are typically based on arguments from illusion or hallucination. For example, a straw that looks bent in a glass of water is used as an argument against the idea that we see the straw as it really is. One is then supposed to conclude (incorrectly) that the visual experience is an illusion. Yet we see the straw as the light is refracted by the water, and the refraction is real. Turns out one sees the straw exactly as it is under such conditions of observation.

Tom Storm

Objections to direct realism are typically based on arguments from illusion or hallucination. — jkop

Kant’s concern was more structural and general: he focused on how the mind contributes to experience. Our sensibility provides raw intuitions, structured by space and time, while our understanding organizes these intuitions into coherent experience through concepts, or categories. As a result, everything we experience: the phenomenal world, is filtered through these mental faculties. This framework implies that our perceptions can misrepresent the “thing-in-itself,” whether through error, illusion, a range of factors. However, Kant didn't need to discuss specific perceptual errors to make this point, his argument is systematic: we never access things directly, only through the mind’s structuring. This argument (in a simple form) occurred to me when I was very young and always seemed more convincing to me than arguments from illusions or hallucinations.

Banno

To perceive something is to be in unmediated contact with it. — Clarendon

There's a lot going on in that. Why should we accept it?

Isn't such statement ruling out representationalism, sense-data theories, intentionalist and inferential accounts by fiat?

And "unmediated" is doing a lot of work. I hope no one will deny that vision is mediated by light, hearing by sound.

And we know that perceptual content is structured, that we see a chair, not bare colours and textures.

And what is that relation, being in contact? spatial? causal? intentional? normative?

I think most contemporary philosophers will want to describe themselves as direct realists of one sort or another. — Clarendon

Maybe not in those terms: Survey Results: Metaontology: heavyweight realism, anti-realism, or deflationary realism?

Kant’s concern was more structural and general: he focused on how the mind contributes to experience. — Tom Storm

And we know that perceptual content is structured, that we see a chair, not bare colours and textures. — Banno

Picking up on these observations: By starting with the idea of "looking at a ship", we can be misled into believing that to perceive a ship is always to do so under that description.. A child has to learn, quite literally, to look at a ship -- to learn what to look for, how to recognize one, what the fuzzy cases are. Direct perception would instead be something like "bare colors and textures" -- a very unnatural thing for the human species to experience, past infancy. I think that to defend direct realism, you have to argue that those unmediated (?) experiences are what we perceive, full stop.

Banno

Picking up on these observations: By starting with the idea of "looking at a ship", we can be misled into believing that to perceive a ship is always to do so under that description.. A child has to learn, quite literally, to look at a ship -- to learn what to look for, how to recognize one, what the fuzzy cases are. Direct perception would instead be something like "bare colors and textures" -- a very unnatural thing for the human species to experience, past infancy. I think that to defend direct realism, you have to argue that those unmediated (?) experiences are what we perceive, full stop. — J

Why not just say that the babe has not yet learned to see the ship, and doesn't do so until they do so under a description? That "seeing a ship" just amounts to applying that set of games and rules. Before learning, the baby sees shapes, colours, textures — not “ships”. After learning, the baby sees ships, and not by constructing an internal model or representation, but by participating in a set of practices: naming, recognising, sorting, using ships.

Seeing the ship is unmediated... Seeing it through a telescope might be called unmediated. What we call a "ship" just is the sort of thing that we see. We don't see it "indirectly" in any ordinary sense.

So the claim is: when I see a ship, I am directly in contact with the ship itself, not with a representation, sense datum, or mental model of it. And what "the ship itself" is, is an aspect of the games we play with words and the world.

Summarising, when a babe learns to “see ships” it is learning to participate in the practice of identifying them. And seeing a ship is direct: you are in contact with the ship, not a mental model. What counts as a ship is socially and linguistically mediated, but that mediation does not make perception indirect. Direct realism = contact with objects; practice/learning = the conditions under which contact with conceptually-defined objects is possible.

jkop

As a result, everything we experience: the phenomenal world, is filtered through these mental faculties. — Tom Storm

Kant doesn't explicitly reject direct realism. His empirical realism and transcendental idealism can be interpreted as two worlds, or two perspectives. I think there's only one world that can be seen in many ways under various conditions of observation.

Tom Storm

↪J

I can see that.

Kant doesn't explicitly reject direct realism. His empirical realism his transcendental idealism — jkop

I’m no Kant expert but I was referring to that argument specifically. Mind you, if we what we see is phenomena not noumena then what meaning does realism have?

jkop

Mind you if we what we see is phenomena not noumena then what meaning does realism have? — Tom Storm

Seeing is part of what's real. No need to split the world in one that we see and another that we supposedly never see.

Paine

As in order to secure direct contact with the mind-external ship, the experience would surely have literally to contain the ship. It's not enough that it's 'about' a ship. A note about a ship is about a ship, but it can't thereby be a means by which we perceive a ship. A thought about a ship is about a ship, but again one can't perceive a ship by thinking about a ship. So it won't help at all to make a view 'direct' just to focus on the way in which a sensation is 'about' or 'of' a ship. The sensation would have to include the ship itself. — Clarendon

There is a problem here with comparisons. If one invokes "experience", that includes all that we do not understand about it happening. Thinking about how perception works does not require a zero sum game where the "real" is real or not. If we do not stand on both sides, we cannot judge.

Tom Storm

Seeing is part of what's real. No need to split the world in one that we see and another that we supposedly never see. — jkop

But isn’t that the question which matters? How do we cocreate our reality as opposed to see reality?

Whether the question matters is a separate one. As Simon Blackburn put it: An idealist is a realist whenever he walks out the front door.

Clarendon

↪jkop

Yes, those are objections that are used to motivate indirect realism (the idea that what we perceive are mental states). They don't seem to apply to the view I have proposed, however. With my view our experiences of perceiving are mental states, but the perceptual relationship itself is not. Thus cases of hallucination share with cases of experienced perception the same mental states, it is just that in the former there is no perceptual relationship there (and thus the experience constitutes a hallucination).

Clarendon

↪Banno

I think it's not in dispute that perception involves direct contact between the perceiver and the object of perception. What's in dispute is what's perceived - the world itself or mental states (with indirect realists saying that we perceive mental states and direct realists saying we perceive external reality).

Clarendon

↪Paine

I'm arguing that experience is not a constituent of a perceptual relationship. We do not perceive things by experience (though we have experiences of perceiving things).

Paine

↪Clarendon

Upon what basis do you make this distinction?

We do not own our experiences; we just have them.

Clarendon

↪Paine

A perceptual experience is 'of' something, namely a perceptual relationship.

My argument is that a perceptual relationship cannot possibly involve a mental state, as then there would be no direct contact between the object of perception and the perceiver. The perceptual relationship must have just 2 relata: the perceiver and the perceived. No doubt experience occurs too, but not as a constituent of the relationship. The experience of perception is 'of' it, not constitutive of it.

Corvus

I am interested in hearing any objections to this 'proper' form of direct realism — Clarendon

I think I have addressed this point some time ago to @RussellA.

It depends on the mode of our perception - how we perceive the object.

When I am seeing a guitar in front of me, and can touch it, pick it up and play it, it is directly I am perceiving or interacting. But when I am thinking about it, imagining it, or remembering it in my mind, it is indirect perception I am having. The real flesh guitar is not available for me - I cannot touch it, pick it up or play it, but I can still see it by imagining, remembering or thinking about it.

So, both indirectly and directly we perceive and interact with objects. It depends on the existence and availability of the object in flesh we are perceiving or interacting.

Paine

↪Clarendon

You set up those conditions of what a "mental state" involved. You presume the difference that you hope to demonstrate.

Clarendon

↪Corvus

I am a direct realist and so i am not disputing that we perceive the actual objects out there - guitars and such like. But when I read other direct realists they talk about experience and think experiences are somehow constituents of perceptual relationships - which seems to me to transform them into indirect realists by another name - or they think the actual objects out there somehow feature as constituents of the experience - which seems incoherent.

Given perceptual relationships must be direct, then it seems true by definition that a perceptual relationship can only have two relata: the perceiver and the perceived. There is no room for any mental states. On my view 'experience' (and thus mental states) enter indirectly: we have experiences 'of' perceptual relationships between ourselves and objects, but these experiences are in no way constitutive of the perceptual relationships themselves.

Clarendon

↪Paine

I do not understand. I take for granted that experiences are mental states. But that's not unorthodox or something the other direct realists would dispute, I think. Admittedly, some of them think mind-external objects can feature as constituents of mind-internal mental states - but my point is that this is incoherent. (I'm sympathetic to what they're trying to do, as they recognize that for perception to be occurring the object does indeed need to get inside the experience, but I just don't see how the notion can be coherent).

Banno

What's in dispute is what's perceived - the world itself or mental states — Clarendon

When you see a boat, it' a boat that you see. If what you see is not a boat - if it is an illusion of an hallucination - then by that very fact what you see is not a boat.

But when philosophers say "perceive" instead of "see" they quickly loose track of what is going on. "Perceive" tries to treat veridical vision, hallucinations and illusions as if they were the same; it often presumes that there is a "something" that is being perceived, even when this may not be so; and it tries its best to be a private inner process.

jkop

An idealist is a realist whenever he walks out the front door. — Tom Storm

:smile: Direct realism is a bit like idealism in the sense that experience and object are not separate entities. The visual experience that you have when you see a real ship is the real ship.

↪Clarendon

Ok, so one objection to your view is that the assumed "perceptual relation" between a "mental state" and the object means that the experience would be indirect.

Brain states are constituitive for having experiences, but since brain states can be shared in hallucinations and veridical experiences of a real object there is no relation between a brain state and the real object. Instead, the experience is the object.

RussellA

I think that direct realism 'proper' would have to be the view that perceptual relations have 2 and only 2 relata: the perceiver and the perceived......................I am interested in hearing any objections to this 'proper' form of direct realism — Clarendon

There seems to be two main forms of Direct Realism. There is Phenomenological Direct Realism (PDR), a direct perception and direct cognition of the object ship as it really is in a mind-independent world. There is also Semantic Direct Realism (SDR), an indirect perception but direct cognition of the object ship as it really is in a mind-independent world.

Your “proper” direct realism seems to be PDR, although many direct realists support SDR.

For the SDR, language is crucial. For example, David Armstrong, who emphasised the role of language in understanding reality, and Michael Dummett, who emphasised the link between language and the world.

I agree that Indirect Realism and SDR overlap in many ways.

However, PDR is far more difficult to justify, and it may be that few Direct Realists actually support PDR.

==============================

The mistake they accuse indirect realists of making is to confuse a 'vehicle' of awareness with an 'object' of awareness.........................Fair enough that the indirect realists are making a mistake. — Clarendon

I don’t think that Indirect Realists are making this mistake. For example, as an indirect Realist, when I look at a pixel on my computer screen that is red, I am well aware that I am directly looking at the colour red.

As an Indirect Realist, I am also well aware that it is the pixel that is red. If there were no pixel I would not be able to see the colour red.

Michael

To perceive something is to be in unmediated contact with it. I take that to be a conceptual truth that all involved in this debate will agree on. — Clarendon

I disagree. You seem to be defining perception of X as direct perception of X, and so this would entail that indirect perception of X isn't perception of X, and so the very concept of indirect perception would be a contradiction.

With that in mind, a 'direct realist' is someone who holds that we are sometimes perceive the mind external world. That is, when I look at the ship I am directly aware of the ship itself. Thus, I perceive the ship.

This is as opposed to indirect realists who hold that we are only directly aware - and so only perceiving - mental states of our own, rather than the world out there. — Clarendon

I understand the distinction between direct and indirect realism to be better expressed by this picture (imagine a line down the middle and them to be two separate viewpoints):

?u=https%3A%2F%2Fstatic.vecteezy.com%2Fsystem%2Fresources%2Fpreviews%2F007%2F742%2F606%2Fnon_2x%2Fknowledge-of-perception-between-direct-realism-and-indirect-realism-vector.jpg&f=1&nofb=1&ipt=99fc66cc6233b37daf3d646d7adf51180f0d5add612a03188bd6ec220de6073d

?u=https%3A%2F%2Fstatic.vecteezy.com%2Fsystem%2Fresources%2Fpreviews%2F007%2F742%2F606%2Fnon_2x%2Fknowledge-of-perception-between-direct-realism-and-indirect-realism-vector.jpg&f=1&nofb=1&ipt=99fc66cc6233b37daf3d646d7adf51180f0d5add612a03188bd6ec220de6073d

Indirect realists accept the distinction between Locke's primary and secondary qualities, and that secondary qualities are qualities that constitute conscious experience and not material objects like apples and ice cream cones, i.e. the left-side of the above picture. Primary qualities like atomic composition and electromagnetic reflectance may causally determine secondary qualities, but it is a mistake to think of these secondary qualities as being properties of the object seen.

This view contrasts with direct realist views, e.g. realist color primitivism, which believe that Locke's so-called secondary qualities are in fact primary qualities, i.e. the right-side of the above picture.

This distinction explains why there is an epistemological problem of perception. If indirect realism is true then how much of the world we experience is a product of our bodies and brains and how much is really "out there", i.e. which qualities are primary and which qualities are secondary? What if even visual distance, shape, and orientation are secondary?

Incidentally I think it's a shame that so much of this discussion focuses on vision and colour at the expense of the other senses and other qualities. I wonder if the realist color primitivist would commit to realist taste primitivism.

flannel jesus

understand the distinction between direct and indirect realism to be better expressed by this picture: — Michael

I don't understand what that picture is getting at at all. They both see an ice cream that's half red, half black. One has a mental image of it being brown and one has a mental image of it being red? That's what I'm getting from it, but that doesn't seem to have much to do with the distinction between direct and indirect realism to me.

Michael · Screenshot 2026-01-04 122803

↪flannel jesus

Cover one half of the picture, and then imagine the other half of the ice cream being a mirror of what you can see.

For the indirect realist the ice cream itself has no colour, because colour is not a property that exists outside of experience, and so it's represented as entirely black. The red, dark brown, and light brown colours are then produced by the brain in response to the eyes being stimulated by various wavelengths of light. As alluded to in my previous post, this might be better explained with reference to taste rather than vision; sweetness isn't a property inherent in sugar but a mental quality produced by the brain in response to the chemical reaction between sugar molecules and taste buds.

For the direct realist, the dark red, light red, and light brown colours are inherent in the ice cream, and in the veridical case the colours we experience "match" these inherent colours.

Michael

This may help:

flannel jesus

↪Michael

that picture doesn't make that clear, but that's fine.

My question is, don't we have a scientifically agreed upon sequence of events from "there's an ice cream in front of you" to "you're experiencing the visual sensation of the ice cream in front of you"? Like, the matter that makes up the ice cream is there, it reflects or emits photons, some of those photons hit your eyes, your eyes send signals to your brain, your brain interprets those signals and the context they're in to create your full visual-spacial-objectoriented experience of the ice cream and the space it exists in.

If we already know the physical sequence of events, what extra disagreement even is there to be had?

Michael

My question is, don't we have a scientifically agreed upon sequence of events from "there's an ice cream in front of you" to "you're experiencing the visual sensation of the ice cream in front of you"? Like, the matter that makes up the ice cream is there, it reflects or emits photons, some of those photons hit your eyes, your eyes send signals to your brain, your brain interprets those signals and the context they're in to create your full visual-spacial-objectoriented experience of the ice cream and the space it exists in. — flannel jesus

Yes, and so the relevant questions are; what and where is colour and what and where is taste? The indirect realist says that colour and taste are something like emergent mental phenomena and not qualities or properties inherent in the ice cream (even if those things which are qualities and properties inherent in the ice cream, like its chemical composition, causally determine particular mental phenomena), whereas direct realists like realist colour primitivists say that colour and taste are not (just?) emergent mental phenomena but (also?) qualities and properties inherent in the ice cream, and that our colour and taste perception is veridical if and only if our experience "matches" what's out there.

flannel jesus

↪Michael

so isn't all that answered in the physical description of the sequence of events?

Michael

so isn't all that answered in the physical description of the sequence of events? — flannel jesus

I think so, which is why the Wikipedia article on direct and indirect realism says "indirect perceptual realism is broadly equivalent to the scientific view of perception that subjects do not experience the external world as it really is."

I think the science clearly shows that colour, taste, smell, etc. are the product of our biology, causally determined by but very different to the objective nature (e.g. the chemical composition) of apples and ice creams.

So the traditional phenomenological and epistemological questions are firmly resolved in favour of indirect realism.

The current problem as I see it is that semantic direct realists have muddied the waters by trying to adapt direct realist terminology to mean something very different — something which doesn't actually contradict the phenomenology or epistemology of indirect realism.

Direct realism about perception

Welcome to The Philosophy Forum!

Categories

More Discussions