The following outlines some ill-considered thoughts on perceptual experience. It begins with emphasizing certain important features of our perceptual experience. The observations made there will be brought to bear on how we come to have beliefs about the unobserved. Subsequently, I will discuss the perception of ethical properties.
Consider the features of your perceptual experience. The most salient features are what is immediately detected by the senses. Touch – whether something is rough or smooth, soft or hard, things feel a certain way. They also taste and smell a certain way. Audible sensations, too, have a unique character. Visual experience gives us a manifold of colors arranged in various ways, which we then try to make sense of. It is sometimes thought that this is the totality of our immediate awareness. All else is inference.
But this is not the whole of our perceptual experience. Note that we also have a certain discriminatory capacity that is characteristic of our perceptual experience. The ability to pick out objects and see them as objects. There is more to visual perception than a swirling tapestry of color; our visual field is populated by objects. Consider your visual experience of a disheveled pile of books. You can pick out each book distinctly, you are aware of where one book starts and ends, you see each book as an independent object. And it is difficult to do otherwise. Try to see two books as a unified object. If it is not impossible, it is at least difficult; as difficult as looking at a blue wall and trying to imagine it yellow. The effort runs counter to the perceptual experience – I can’t help but see them as objects. I am sitting down to write, when a small black dot intrudes upon the periphery of my vision. My attention is immediately attracted to the movement of the black dot, and now I can track it. I saw it, picked it out as an object, and now, realizing that it is a gnat, I raise my hand to swat it. We can pick out and track objects, our attention is drawn toward the movement of objects. 1
Now I’m sitting at home, starting to watch “Pulp Fiction” on Netflix. The cursor does not disappear as the movie begins to play. I am irked by the presence of a foreign object on my screen. So I slide it all the way to the bottom right corner of the screen. Now just the very tip of the cursor is present, the rest having disappeared “off the screen”, as it were. I am feeling better, but still the tip is bothering me, even though it is not in the way of my seeing the movie. Why does the tip still bother me? Because even though the cursor is off the screen and only shows the tip, I am perceptually aware of the cursor as a whole object. The presence of the entire cursor is manifested in my perceptual awareness – I get a distinct sense of the cursor as a whole, even though I see just the tip.
Let’s say that the cursor now actually disappears. There is nothing to indicate the object’s presence; I do not have a sense of it. I am no longer disturbed by its presence. Suppose instead there is a very tiny crack in the bottom right corner of the screen. Would this be as disturbing as the cursor? I do not think so. This is because it doesn’t impose itself on the screen as an object.2
This is arguably a case of amodal completion. Amodal completion is a phenomenon that suggests we get a perceptual awareness of an entire object when the entirety of the object is not observable. A dollar bill is slipped underneath a newspaper, so that only a thumb-length of the bill is visible. Looking at the newspaper on top of the dollar: you see the newspaper, you see the end of the bill sticking out. But you also get a sense of the whole dollar. If you wanted to, you could probably take a pen and draw out the rest of the dollar on top of the newspaper such that it matches the bill beneath, and you do this just from your sense of the object.
This is a somewhat richer sense of perception than some may be used to. But I think that upon reflection, we really do get more from perception than Russellian sense-data. I think this richer conception has many consequences in a variety of philosophical subjects and problems, particularly when (1) it comes to the status of causal relations and (2) it comes to the status of ethical judgments.
Hume said that our idea of causal connection comes from constant conjunction, which does not give us reason to believe in a causal connection. Goodman in Fact, Fiction, and Forecast explains his new riddle of induction in the following way. Inductive hypotheses based on lawlike correlation are confirmed by their positive instances. 3 Inductive hypotheses based on coincidental correlation are not confirmed by their positive instances. The trouble is in knowing how to distinguish between hypotheses expressing a lawlike correlation and hypotheses expressing a coincidental correlation.
These mysteries thrive on a restricted view of what we get in perceptual experience. On a richer picture, these difficulties can be overcome.
First note that we all possess a concept of causation. How did we achieve this? Well, how did you acquire other concepts that you possess? For instance, consider your concept of a chair. Perhaps you were young, and your mother pointed at a chair and said, “chair”. You looked to where her finger was pointing, saw the chair, and abstracted the concept of chair from it. Granted, maybe it took more than one iteration of this process to acquire the concept, but eventually you did, by associating a word with an object in your perceptual experience. Now consider a pot of boiling water on a flaming stovetop. You certainly see the pot. You see the water starting to bubble. You see the fire and the stovetop. On an impoverished view of perceptual experience, this is all you get. But in addition to seeing all these individual isolated elements, you also see the fire boiling the water. In this way, you see the causal relation between the fire and the water. You don’t just see a ball moving through the air and then a broken window; you see the ball shattering the window. You have a perceptual experience of a causal relation. This is how we get the idea of cause, precisely because we see more than one thing being constantly conjoined to another thing; we see the relation between them. In seeing the causal relation between them, we acquire the concept of causation.
At this point there are two objections that are worth mentioning. (1) If we can see a causal relation, then there must be something it looks like. (2) There are many times we see a causal relation and take it to be coincidental, and vice versa. I will respond to each of these in turn.
Here is a response to (1). What should a causal relation look like? When we say that A caused B, we say A affected B in such-and-such a way. What would it look like for A to affect B in such-and-such a way? We see A act on B, and we observe an effect in B. We see a cause and we see an effect – the two ends of the causal relation. This wasn’t enough for Hume, who thought that you saw the conjunction of two completely independent events. What would satisfy Hume? Some observable unity between the cause and the effect? But the cause and the effect are fundamentally distinct. What if there were some sort of strange orangey-glow that appeared around wherever a causal relation is instantiated? Such an idea is ludicrous. There is no way to come to associate the glow with a causal relation – this would be nothing more than mere constant conjunction. You see the cause and you see the effect – that’s all you need to see a causal relation. One billiard ball is rolling toward another. You see it strike the second ball, and the second ball moves. This is what it is to see the one ball striking the other ball, to see the causal relation. There is nothing that it looks like over and above the observation of the cause and the observation of the effect. There doesn’t need to be anything in particular that it looks like.
Here is a response to (2). Consider Jack, who is looking out into the distance. He turns to his friend, Jill, and says, “Hey, look at that bird way over there”. Jill looks and says, “That isn’t any bird; that’s a plane!” And sure enough, it is. Jack’s judgment of his perception is mistaken. How is Jill to respond? One way she won’t respond is by dismissing Jack as a capable perceiver. She won’t say, “Are you sure you had a caprese sandwich for lunch, too? If you looked closer, it might have been turkey.” Later, when Jack tells Jill that he saw two big rigs on the way to work, Jill is not going to doubt him. In general, Jill will still trust Jack’s reports on his perceptual experience. Much in the same way, we can be mistaken about our perception of a causal relation. Sometimes we will be wrong about there being a causal relation. But then we come to correct it and learn something.
This provides us with a way for distinguishing between lawlike and coincidental hypotheses. If we can have perceptual awareness of causal relations, then we can come to believe things like A caused B. Consider Goodman’s use of “grue”. We do not believe the hypothesis that all emeralds are grue because we do not see a causal connection between emeraldness and grueness. 4 What we see is a connection between emeraldness and greenness. This explains our belief in green hypothesis over the grue hypothesis. Moreover, it explains how we come to have beliefs about the unobserved. We get a sense of which things are causally connected and which aren’t. If the ascription of causal connection were based merely on inference, then to infer the causal relation would require already having a concept of cause. To explain how we have this concept antecedently would indeed be mysterious. If we acquire our concept of cause through perception, then we can better explain how we come to have causal beliefs and beliefs about the unobserved.
Let’s explore some of the other features of our richer conception of perceptual experience. Notice that we have concepts of things like right or wrong. These must have been acquired in some way, but how? It is important to note that these concepts are different from concepts like pleasure or pain. Getting rolled over by a boulder would be rather painful, but it is neither right nor wrong. So right and wrong is of a different quality than pleasure and pain. We do not get rightness or wrongness out of just pain or pleasure. So there must be some other way that we acquire the concepts. It is not as though the concepts could simply be taught. They are too complex and too wide in scope for that. It is too challenging to give a general definition of right such that a person would be able to consistently apply the concept in a wide variety of circumstances or that the overarching majority of people would assent to. (I invite you to try.) You might try to teach the concept of right or wrong through example. But this, too, has its problems. Consider showing a child a video of stealing and saying to the child “This is wrong”. The child could think (1) it is wrong because stealing is something that is wrong (or violates our customs) or (2) it is wrong because it caused the other man harm or distress. But this is inadequate. (1) would not teach the child the resources for determining which other actions are wrong. He only learns that stealing constitutes some sort of violation, but not the reason why. (2) is problematic because the “causes harm” becomes the criterion that the child will use to evaluate the rightness or wrongness of an action. But as we have discussed, causing harm is not constitutive of something’s being wrong; there are times when harm is caused but the action was good or neutral. So we do not come to right or wrong through our pleasures or pains or through the process of education.
In keeping with the theme of this paper, I propose that we acquire our moral concepts through perceptual experience. We see that an action is right or wrong. It is my hope that I can make this view a little more plausible.
Consider a peaceful island society with no written history. Since the formation of the society, there has never been any crime. The people have no concept of right or wrong because nobody has ever done any wrong on the island. Hundreds of years pass and they continue to live this way. But one day, a stranger comes to the island. He behaves maliciously. When a citizen of the island addresses him, the stranger responds with great hostility and verbal abuse. The stranger sets fire to an islander’s crops. Surely the islanders would think that the stranger has behaved badly. But I suspect that they will feel something more specific than that. They won’t just think that his actions are bad. In witnessing the stranger’s malicious acts, the islanders would see that the action is wrong, and thus acquire the concept of wrong and come to think of the stranger’s actions as wrong. They will have perceived wrongness for the first time. And that’s a different sensation than something being bad; there’s a certain revulsion that comes with experiencing wrongness, a distinct-felt quality. The islanders would feel this revulsion, even though they previously had no concept of it. They acquired it from their perceptual experience of the stranger’s actions. Notice the immediacy of this visceral reaction.
When I experience wrongness, I feel it in a certain way distinct from how I experience badness or pain. There is a unique, felt quality. And this is a part of my perceptual experience. It is not as though I come to the conclusion that something is wrong through some operation of reason. I see a man steal a bike on a university campus. This is not how the sequence of events works. I see the theft. I identify the action as a theft. I check theft against my favorite moral system. After performing the necessary deductions in that system, I reason to the conclusion that: that is wrong or morally impermissible or something. Thus I have come to my moral judgment of the situation. This is much too unwieldy and long a process to accurately describe how we make moral judgments. It seems like I see the man steal the bike and feel that it is wrong and so judge it so. If afterward I wish to rationalize my judgment by inputting the action into my favorite moral system, I may do so. But I intuited the action as wrong prior to reasoning it to be wrong. I sensed it before I judged it. The wrongness of the action is present in my perceptual experience.
I can imagine that there must be several objections at this point. (1) Some feral children do not have a sense of right or wrong – if wrongness was present in our perceptual experience, then they should be able to have this sense. (2) There is nothing that a moral property looks like. (3) If we experience moral properties, then how can we account for deviations in ethical judgment? I will respond to each of these in turn.
That feral children do not have a moral sense is not a problem. Consider a person born and raised in a dark room. Such a person would not develop his sense of sight. Maybe he would see shadows or blurs, but for the most part he is blind. Because his development was devoid of visual stimuli, he did not develop the sense modality. Such cases have been known to happen. Likewise, if a feral child does not have a moral sense, this does not mean that most people do not have a moral sense. The feral child simply developed in an environment that was devoid of ethical stimuli. This is especially worth noting if you object to the peaceful island example. You might reject the intuition that the islanders see the wrong in the stranger’s actions, but this is not a devastating criticism. The islanders would have developed in a society devoid of ethical properties, and so it is reasonable to think that they would not see the wrong in the stranger’s actions, because their moral sense never adequately developed.
That a moral property doesn’t look like anything is a similar complaint to that discussed earlier, that a causal relation doesn’t look like anything. A similar strategy will be used here. Just because something does not look a particular way is not reason to reject the presence of that thing. Causal relations are one example. What should a moral property look like? It is hard to say. But what would any concept/property look like? If someone were to ask, “what does the green look like?” it would be difficult to respond. You might show her a green thing and say, “this is what green looks like”. But you aren’t showing her greenness, you are showing her an object that is green. To show someone a concept, you show someone an object bearing the property. If someone asked, “what does wrongness look like?” I might respond by showing a picture of people being maltreated in an internment camp. I show an object that bears the property, not the property itself. There is nothing that greenness qua greenness looks like, just as there is nothing that an ethical property qua ethical property looks like. Or think about walking through your local farmer’s market. You see twelve tomatoes, ten kumquats, and eight pears. But how were you able to see how many? It is not as though you see some platonic Two instantiated. There is no “two” that you could point to. Still, you could count the objects, one, two… You could see that they were two despite no “two” actually being instantiated. We simply see the wrong just as we saw that one thing causes another.5
We can explain deviations in moral judgment through our notion of moral sense. An analogy to other senses will be useful. Let’s consider our sense of smell and a collection of a variety of fine wine. Wine x and wine y do not have the exact same scent, though it is similar. I cannot tell the difference, for I have a weak nose. But suppose there is a wine connoisseur who could. In fact, this wine connoisseur is so discerning such that he might be labelled “the perfect smeller”. With ease he distinguishes between x and y. I disagree with his assessment, but then that’s just attributable to the fact I cannot smell as well as him. If I could, I would have agreed with him. The imperfections of my nose affect my judgment. Deviations in moral judgment work much the same way. Suppose there is an event p and q, such that they instantiate the moral properties Φ(x) and Ψ(y), respectively. Perhaps I had the misfortune to not just be born with a poor nose, but also an imperfect moral sense. It is possible that I (mis)perceive Φ(x) and Ψ(y) to be instantiating the same moral property. A person with an acute, discerning moral sense would be able to accurately distinguish between Φ(x) and Ψ(y). So there moral properties are really there to perceive, but due to an imperfect moral sense, like an imperfect nose, we perceive the situations differently.6 There is more or less general agreement on how things smell. Likewise, there is more or less general agreement on ethical properties. Stealing is bad, murder is wrong, etc… The disagreement is usually on the fringes. There is a sort of general agreement, just like with other senses.
It is worth noting that I do not think we have a distinct faculty of moral sense, like a nose or eyes or some other appendage. This view does not require that. What it may require is some view as to how ethical properties are actually instantiated – that is, what constitutes the instantiation of an ethical property.
Ethical properties are instantiated in virtue of instantiating the right psychological and physical properties. The instantiation of ethical properties will not be a matter of looking toward the physical properties alone. If a large boulder rolled down a hill and crushed a civilian, that would be unfortunate, but it would not be moral in character. In contrast, if a large boulder were pushed down a hill by a questionable character with ill-intent, then the event resulting in the civilian being crushed will be moral in character. We find the questionable character morally responsible for the flattening of the civilian. The psychological component of the questionable character is necessary for the instantiation of some ethical property. Were these merely physical events, we would perceive no moral properties. By the same token, the instantiation of an ethical property is not based only on the instantiation of the right psychological properties. Should the questionable man form the intention to harm the civilian by pushing the boulder, but he does not actually act on it (that is, lift his arms and push the boulder), then we would not perceive any wrongness in the situation. So the instantiation of a perceptible ethical property requires the instantiation of both psychological and physical properties. This has been a rather quick treatment of how ethical properties are instantiated. A more robust account would take us too far afield, and would be more appropriate as the subject of a future paper.
I will conclude this paper by highlighting some of its key points. We get more in sense perception than Russellian sense-data suggests. We have unique capacities like the ability to discriminate between objects, and to be aware of objects that are only partially in view. We acquire most of our concepts from the world, including our concept of cause, the acquisition of which is made possible by the richness of our perceptual experience. We acquire ethical concepts from the world in a similar way, by directly perceiving them. Ethical properties are instantiated in virtue of the instantiation of a certain set of physical and psychological properties. Like our other sense modalities, we may be wrong but at least there is enough general agreement.
- This is also true of many animals ↩
- We can predicate x of a crack, which means that the crack is an object in some sense. But it is not a discrete object. Suppose there were a rent in the cover of a book. We can predicate things of the book and the rent separately, but we cannot consider the rent independently of the book. The rent inheres in the book, in a sense, and is a part of the book. ↩
- We can take lawlike to mean something like causal or explanatory. I subscribe to Dretske’s view of scientific law, where a law states conditional relation between properties of objects, e.g. if A-ness then B-ness (not if A then B). ↩
- It is worth noting that it is difficult to see how we would ever acquire a concept of grue through empirical investigation. We could only contrive it. A man, looking at a single emerald, would never acquire the concept of grue, for he could never perceive the temporal feature. Thus grue is an artificial concept. ↩
- Moreover, if we see wrongness as a two-place relation (e.g. x wrongs y), then we can also see wrongness as a causal relation, e.g. x wrongs y by way of x producing such-and-such effect in y. ↩
- As another example, think about the Rorschach Test. There is little agreement over what any particular image represents. Each person sees the same inkblot(s). Yet each person perceives it differently and will have to struggle to see it another way. So too with ethical properties. ↩