I’ve never used the word “terminate”. If perception “terminated” at mental phenomena then we wouldn’t have indirect perception of their causes, and mental phenomena wouldn’t be considered an intermediary.
As for being a constituent of experience; see naive realism. You seem to have some understanding of what it is and how it differs from non-naive views.
Intentionality and normativity work for me the same way as they do for you. You just don’t like to call your view “indirect realism”, precisely because you seem to redefine “direct perception” in terms of this intentionality and normativity rather than in the traditional terms of phenomenology.
I wouldn’t argue with that. In fact, I would go further. To put it this way, perception needs to play its part in our actions in the world. Then there’s a feed-back loop, which is almost certainly the point of the exercise in evolutionary terms.
Yes, that’s my point as well. My protest is only against those who think it is the end of the matter.
I do see colours, and different people can see different colours when looking at the same object, e.g. with the photo of the dress. The colours that we are familiar with in visual experience are not properties of material surfaces, although they are usually causally determined by the way in which those surfaces reflect photons.
Use whatever word you want. It doesn’t alter the decisive point: if mental phenomena are intermediaries, then the perceiver’s awareness contacts them first and reaches the world only through them via inference. My question remains: what grounds that inference? You keep saying “indirect perception of their causes,” but indirect perception just is inference from intermediary to cause. So what makes that inference truth-tracking rather than just a model reshuffling its own contents?
Intentionality doesn’t work the same way for both of us. On my account, awareness is constitutively world-directed — aboutness isn’t something achieved by inference from non-intentional materials, it’s the basic structure of cognition. On your account, the basic level is non-intentional (phenomenal qualities present to consciousness), and world-directedness is reconstructed via causal inference. Those are genuinely different architectures, not different labels for the same thing. The question is which one can actually account for knowledge, and I’ve been arguing that reconstruction from non-intentional starting materials can’t get off the ground without quietly helping itself to the very thing it’s trying to reconstruct.
You say I’m redefining “direct perception” away from phenomenology. I’d say your notion of what counts as phenomenology is already theoretically loaded — it treats the having of qualia as the basic phenomenological datum and treats the world-directedness of experience as derivative. I think that gets the phenomenology backwards.
I agree, and would take even further: the feedback loop you’re describing — perception, action, correction — explains how organisms become reliably adaptive. And that’s genuinely important. But I would argue that being reliably adaptive does not yet amount to knowing you’ve got it right. The further question is: what’s the difference between a system that tracks its environment successfully and a knower who can reflectively affirm that her perceptions are trustworthy? That second thing requires something the loop itself doesn’t provide.
John and Jane both see an intact red apple 10m in front of them (in the intentional sense). Because of this they both believe that there is an intact red apple 10m in front of them. However, one of the apples was disintegrated 5 seconds ago, but because light travels at 1m/s neither of them has yet seen this happen — and so one of them has a false belief.
We both agree that the above is plausible. We just disagree over whether them seeing an intact red apple 10m in front of them (in the intentional sense) counts as direct or indirect perception of an intact red apple 10m in front of them.
I say that it counts as indirect perception because the apple is not a constituent of the experience (most obvious for the person facing the disintegrated apple). The experience is a mental phenomenon — complete with “intentional structure” and “world-directedness” — that can occur even if the apple no longer exists and even if the apple never existed (e.g. if the experience was caused by direct brain stimulation by a mad scientist).
No, the inference I’m talking about is “I see an intact red apple 10m in front of me, therefore there is an intact red apple 10m in front of me [because it’s the best explanation for why I see what I see]”. It’s how John and Jane both reason, but one of them is wrong in their conclusion.
I actually think you just conceded the point without realizing it.
Look at the inference you’ve described: “I see an intact red apple 10m in front of me, therefore there is an intact red apple 10m in front of me.” The starting point of that inference is already intentionally structured. It’s not “I have red qualia arranged apple-wise” — it’s “I see an apple.” World-directedness is built into the premise. So on your own account, intentional structure is prior to the inference, not a product of it.
But then what work is “indirect” doing? You’re describing a perceiver whose experience is already world-directed — already of an apple — who then makes a further judgement about whether the object is really there. That further judgement can go wrong, sure. John and Jane’s experiences are identical but one apple is gone. I agree with all of that. But fallibility isn’t the same as indirectness. A capacity can be genuinely world-directed and still misfire. Thats just what it means for perception to be fallible rather than infallible.
Perhaps you’ll reply: “the premise is just describing how things appear, not how they are.” But even “how things appear” is already intentionally structured. Appearing isn’t a brute phenomenal given — things appear as something, of something. If you strip the intentional structure out and retreat to something like “I have reddish roundish qualia,” you lose the resources to formulate “apple” as a hypothesis in the first place. If you keep it in — “it appears there’s an apple” — then world-directedness is doing the work before inference ever enters the picture. Either way, the intentional structure is prior, not derived.
And if the next move is “well, the intentional structure is itself just part of the appearance” — ok, but now ask: appearing to whom? Directed at what? You either have genuine directedness, in which case it’s doing real cognitive work and isn’t just another phenomenal feature, or you don’t, in which case you’ve just re-described the qualitative character of the experience in intentional-sounding language and the word “about” has become a kind-of trick or illusion. You can’t have aboutness that is genuinely doing the theoretical work of connecting you to the world while simultaneously insisting it’s nothing more than an interior phenomenal property. That’s the dilemma and I don’t think theres a stable position between those horns.
We have direct perception of X if and only if X is a constituent of the experience. We have indirect perception of X if and only if we have have perception of X and X is not a constituent of the experience.
When John and Jane see an intact red apple 10m in front of them they do not have direct perception of an intact red apple 10m in front of them because no such thing is a constituent of their experience, but one of them continues to have indirect perception of the intact red apple 10m in front of them.
The inference they both make is “there is an intact red apple 10m in front of me”, and one of them is correct.
Nothing about this has anything to do with whether or not “intentional structure” or “world-directedness” is “built in” to the experience. Indirect realism doesn’t deny this because it doesn’t address this. It is concerned only with a rejection of naive realism, the acknowledgement that sensory qualities like colour are mental phenomena, and the epistemological implication that we might perceive things that either don’t exist or are radically different to how they appear.
Well, first thing I would do is change your formulation. Tracking the environment is not sufficient. Acting in it requires more, so that changes in the environment that are caused by the system’s own actions can be distinguished from changes in the environment that have some other cause. That means that the system must have at least a rudimentary sense of self. The beginnings of self-evaluation are there.
But your alternatives are presented in different conceptual systems - language games. There’s no "straightforward way of building from impersonal systems to people reflecting. What is clear, at least to me, that one could not begin to hope to building a perceptual system on its own that could do that. It requires the complex background of a whole person.
But we learn at least some of the skills of distinguishing between reliable perceptions and unreliable ones almost as soon as we start to perceive. How? Think of the small child rushing into its parents room in the middle of the night because the house is surrounded with wolves. It’s a dream. Small mistakes and disappointments happen all the time - that’s inherent in the feedback process. So common sense doesn’t have a problem with that.
But what we don’t - and can’t - learn is how to solve the philosophical worry that none of our perceptions is trustworthy. That’s a whole different level and, imo, an incoherent project.
True, physics tells us that the world is actually very different from how we perceive it. But how did it get to that conclusion? Surely the answer is quite straightforward - there’s a feedback loop between hypothesis and observation and experiment.
I’ve seen that argument around. I would have thought that successful adaption does show that you have got something right. Knowing that does require more - roughly, the context of personhood with all the complexity that involves.
It appears that a lengthy response to this has been removed.
The thrust was that I think your description precludes your conclusion or alternately unnecessarily complicated the question into what amounts to abstraction which I think is incorrect.
If I have time later today I’ll edit in a full response here and let you know. Sorry about that.
I always think this debate is framed unproductively.
Perception clearly involves a relation between an organism and its environment, features of that environment, objects in that environment. You can carve off a bit of complex causal process and describe it as “Michael perceiving an apple." In that description, Michael and the apple are participants, top-level Michael, Michael as cognitive agent.
On the other hand, Michael has distinct subsystems that are responsible for his perceiving his environment. If such a subsystem interfaces with other subsystems, responsible for predicting and planning and so on, it seems natural enough to think of the perceptual system as producing some kind of intermediary, its output that becomes an input for other processes. That output is not an object for top-level Michael, but it might be for some subsystem of Michael.
If anything like this is correct—and no doubt there are much better descriptions available—then the question of whether Michael-the-cognitive-agent’s perception of the apple is mediated looks like semantics, in the negative sense.
You keep offering that definition — direct perception iff X is a constituent of the experience — as though it’s a neutral criterion both sides can accept. But it isn’t. The entire question is what counts as a constituent of the experience. You say: only the sensory qualities. I say: the object as presented through the sensory qualities is also constitutive of the experiential act, because the act is structurally world-directed. You’re not settling the debate with that definition, you’re restating your position as a definition.
And look — you say indirect realism “doesn’t address” intentionality and world-directedness, it just rejects naive realism and flags epistemological implications. But you can’t have the epistemological implications without addressing intentionality. The epistemological worry — “maybe things don’t exist or are radically different from how they appear” — only arises because experience purports to be about the world. If you don’t have an account of how that aboutness works, you don’t have an account of how the epistemological worry gets its grip. You’re helping yourself to intentional structure whenever you need it (“John sees an apple,” “one of them has a false belief”) while officially declaring it outside your scope. That’s not a modest restriction of focus, its a gap in the theory.
I do think we’re genuinely looping now though, so I’ll leave it here unless you have something new to add. No hard feelings — I think the exchange has been useful for clarifying where exactly we disagree, even if neither of us has budged.
It’s what both naive and indirect realists mean by the phrase “direct perception”. If you mean something else by it then you’re talking about something else.
You seem to think that “direct perception” means just one thing and that both naive and indirect realists are wrong about what is required for perception to be direct. This is a mistaken interpretation of the issue.
This is a nonsequitur. That perception is “world-directed” is not that distal objects are constituents of the experience. This is where the example with the slow light and the disintegrated apple is important; John’s experience continues to be “world-directed” even after the apple has been disintegrated, but because the apple has been disintegrated it is not a constituent of the experience.
The fact that you reject naive realism but still claim that distal objects are “constituents” of the experience shows that, like with the term “direct perception”, you mean something else by the term and so are talking about something else.
Intentionality and world-directedness and aboutness works for me the same way it works for you. I just don’t call this “direct perception” because the meaning of the term “direct perception” is such that we have direct perception of X if and only if X is a constituent of the experience, and distal objects are not constituents of the experience — hence why naive realism is false.
I’ve now asked “how does intentionality work on your framework” multiple times and the answer keeps being “the same way it works for you.” But this overlooks that we have incompatible accounts. Until this gets addressed, I don’t think we can make further progress. So I’ll bow out here. Its been a good exchange — I think anyone following can see clearly where we each stand and where the pressure points are. Thanks.
I think you’re doing something interesting, perhaps without noticing it. Every time the feedback loop seems insufficient, you enrich it — first with a “rudimentary sense of self,” then with the ability to distinguish dreams from reality, then with hypothesis-testing. But each of those enrichments is smuggling in exactly the kind of cognitional act I’m pointing to.
Take your child example. The child doesn’t just behave differently after the wolf-dream. The child makes a judgment: “that wasn’t real.” Thats not a feedback correction like a thermostat adjusting temperature. It’s a reflective act — the child grasps the difference between appearance and reality and affirms something about which side a particular experience falls on. You’ve quietly moved from describing a loop to describing a knower.
And to be clear — I’m not raising the skeptical worry that none of our perceptions might be trustworthy. I think that worry is incoherent, we agree there. My question is different: what kind of act is it when we judge that a perception is reliable? That act isn’t itself another perception, and its not just more feedback. It’s something else. And whatever account we give of cognition needs to say what that something else is, rather than just folding it back into the loop.
Right, successful adaptation shows that something has gone right — I wouldn’t deny that. But “something has gone right” is a judgment we make about the bacterium. The bacterium swimming up the glucose gradient has gotten something right, but it doesn’t know that, and not just because it lacks the word for it. It lacks the capacity to affirm it. So the question isn’t whether adaptive success tracks reality — it does — but what has to be added before tracking-reality becomes knowing-reality. And as you say yourself, that requires “the context of personhood with all the complexity that involves.” So we agree on the destination. I just think the feedback loop, however enriched, doesn’t tell us what that complexity actually is. It just gestures at it.
Not at all. We start with the naive view of perception, as described by Martin 2004:
On [the naive realist] conception of experience, when one is veridically perceiving the objects of perception are constituents of the experiential episode. The given event could not have occurred without these entities existing and being constituents of it in turn, one could not have had such a kind of event without there being relevant candidate objects of perception to be apprehended. So, even if those objects are implicated in the causes of the experience, they also figure non-causally as essential constituents of it… Mere presence of a candidate object will not be sufficient for the perceiving of it, that is true, but its absence is sufficient for the non-occurrence of such an event. The connection here is [one] of a constitutive or essential condition of a kind of event.
The important points to take from this are:
P1. If I have direct perception of an object then that object is a constituent of the experience
P2. If an object is a constituent of the experience then that object exists
P3. That an object exists and is causally responsible for the experience does not entail that it is a constituent of the experience
I then consider this thought experiment:
P4. An apple is placed 10m in front of me
P5. The light it reflects travels at 1m/s
P6. The apple is disintegrated after 5 seconds
C1. Therefore, I see an apple for 5 seconds starting 5 seconds after the apple has been disintegrated
C2. Therefore, the apple does not exist during the 5 seconds in which I see an apple
C3. Therefore, the apple is not a constituent of the experience during the 5 seconds in which I see an apple
C4. Therefore, I do not have direct perception of the apple during the 5 seconds in which I see an apple
P7. If the apple is not a constituent of the experience during the 5 seconds in which I see an apple when the light travels at 1m/s then it is not a constituent of the experience during the 5 seconds in which I see an apple when the light travels at 299,792,458 m/s
C5. Therefore, the apple is not a constituent of the experience during the 5 seconds in which I see an apple when the light travels at 299,792,458 m/s
C6. Therefore, I do not have direct perception of the apple during the 5 seconds in which I see an apple when the light travels at 299,792,458 m/s
At no point have I defined experience in such a way that naive realism is false by definition. Instead, I have taken the naive realist definition of direct perceptual experience and shown that it does not obtain.