SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. 1. Preliminaries
  4. 2. The Phenomenal Representation of Ethical Properties
  5. 3. Ethical Perception and Cognitive Penetration
  6. 4. Objections and Responses
  7. 5. Epistemic Dependency
  8. References

In the recent metaethical literature there has been significant interest in the prospects for what I am denoting ‘Perceptual Intuitionism’: the view that normal ethical agents can and do have non-inferential justification for first-order ethical beliefs by having ethical perceptual experiences, e.g., Cullison 2010, McBrayer 2010, Vayrynen 2008. If true, it promises to constitute an independent a posteriori intuitionist epistemology, providing an alternative to intuitionist accounts which posit a priori intuition and/or emotion as sources of non-inferentially justified ethical beliefs. As it is formulated, it is plausible that a necessary condition for the view is the truth of Ethical Perception: normal ethical agents can and do have perceptual experiences (at least some of which are veridical) as of the instantiation of ethical properties. In this paper a sophisticated and promising account of Ethical Perception is offered. Extant objections are shown to fail. However, it will be argued that it is far from obvious that the account of Perceptual Intuitionism which emerges constitutes an independent alternative to other intuitionist accounts. This is because we have reason to think that ethical perceptual experience may be epistemically dependent on other epistemic sources, e.g. a priori intuition or emotion.

In the recent metaethical literature there has been significant interest in the prospects for Ethical Intuitionism:1

Ethical Intuitionism (EI): normal ethical agents have at least some non-inferentially justified first-order ethical beliefs.2

EI says that normal ethical agents have at least some ethical beliefs that are justified independently of the justification they have for believing other propositions.3

In general epistemology, there is a reasonable degree of consensus that if anything is, the following are sources of non-inferential justification: perception, a priori intuition, introspection, and memory.4 Despite this, no-one defends an intuitionism that is based upon introspection or memory. The explanation for the lack of an Introspective Intuitionism in contemporary ethical epistemology seems clear: introspection doesn't plausibly provide non-inferential justification for first-order ethical beliefs, i.e., one could not be justified in believing that torture is wrong simply on the basis of introspection, even if one can be introspectively non-inferentially justified in believing, e.g., that it seems to me that torture is wrong. The explanation for the absence of a Memory Intuitionism is more interesting. Here is one suggestion: on a plausible epistemology of memory, Preservationism5, memory can only preserve a previous justification that a subject had for belief, but cannot generate it, i.e., if I seem to remember that p then I am only justified in believing p if I had some previous non-memorial justification for believing that p. This is just to say that memory is an epistemically dependent source of non-inferential justification. Here is an account of epistemic dependency:

Epistemic Dependency (ED): a state, d, epistemically depends on another state, e, with respect to content c iff e must be justified or justification-conferring in order for d to be justified or justification-conferring with respect to content c.6

Even if memory does ground non-inferentially justified ethical beliefs, one plausible reason why no-one defends Memory Intuitionism is that such a view is, by itself, inadequate. Its proponent would still owe us an explanation of how ethical justification (non-inferential or otherwise) gets generated.

Independently of the plausibility of this epistemology of memory, a general lesson can be learnt: an adequate intuitionist epistemology must ultimately account for non-inferentially justified belief in terms of epistemically independent sources.

With this in mind, consider the following account which has been the subject of recent discussion:7

Perceptual Intuitionism (PI): normal ethical agents can and do have non-inferential justification for8 first-order ethical beliefs by having ethical perceptual experiences.9

If true, Perceptual Intuitionism would seem to constitute an independent a posteriori ethical intuitionism, providing an alternative to intuitionist accounts which posit a priori intuition and/or emotion as sources of non-inferentially justified ethical beliefs, e.g., a particularist intuitionism which counts ethical beliefs about particular cases as a priori justified.10 PI may therefore hold considerable attraction for ethical naturalists.11

As it is formulated, it seems plausible that the following non-epistemological view is necessary for Perceptual Intuitionism:12

Ethical Perception (EP): normal ethical agents can and do have perceptual experiences (at least some of which are veridical) as of the instantiation of ethical properties.

In this paper a sophisticated and promising account of how EP might be true will be offered. Extant objections will be shown to fail. However, it will be argued that it is far from obvious that the account of Perceptual Intuitionism which emerges constitutes an independent alternative to other intuitionist accounts. This is because we have reason to think that ethical perceptual experience may be epistemically dependent on other epistemic sources, e.g., a priori intuition or emotion.

1. Preliminaries

  1. Top of page
  2. Abstract
  3. 1. Preliminaries
  4. 2. The Phenomenal Representation of Ethical Properties
  5. 3. Ethical Perception and Cognitive Penetration
  6. 4. Objections and Responses
  7. 5. Epistemic Dependency
  8. References

Discussions of Ethical Perception typically assume a representational theory of perception, i.e., (roughly) the view that to have a perceptual experience of an object O as having a property F, is to be in a perceptual mental state with phenomenal character which has representational content O is F. This view of perception is almost always accompanied by the view that perceptual experiences, like beliefs, have accuracy conditions, and that specifying the accuracy conditions of a given perceptual experience is a way of specifying the content of that experience, i.e., if we know that the experience will be accurate iff there is a red apple on the table, then we also know that the perceptual experience has the representational content that there is a red apple on the table (it is however a point of intense debate what the accuracy conditions of perceptual experiences are). This sits comfortably with the popular distinction between perceptions which are factive and perceptual experiences which are not.13

Before sketching a plausible account of Ethical Perception, it is worth briefly mentioning a couple of initial motivations for thinking that some version of it is true. The first of these is phenomenological. Consider the following oft-quoted example from Gilbert Harman:

Cat: If you round a corner and see a group of young hoodlums pour gasoline on a cat and ignite it, you do not need to conclude that what they are doing is wrong; you do not need to figure anything out; you can see that it is wrong.14

From this phenomenological datum defenders of EP think that we have something like a prima facie case for the view that the ethical seeing described is non-metaphorical. However, sceptics might object that all such scenarios establish is a case against the view that, when making particular ethical judgments, subjects always consciously (or unconsciously) draw inferences from general principles, and, that by itself, Cat doesn't provide us with any reason to consider Ethical Perception.

In order to show that EP really does constitute a reasonable hypothesis about what goes on in cases like Cat it is important to consider the high-level-perception motivation for EP which arises from recent developments in the philosophy of perception. Almost everyone who thinks that perceptual experiences have representational contents agrees that the contents of visual experience will include properties pertaining to shape, size, position, colour and object-hood. However, some philosophers15 have argued that human beings might be capable of having perceptual experiences which represent ‘high-level’ properties (H-properties) such as natural kind properties and causal relations. In this context, it seems that if suitably trained perceptual agents can come to perceive that, e.g., the tree over there is a pine, then, plausibly, it becomes somewhat less incredible to think that in scenarios like Cat, conceptually and cognitively sophisticated agents could perceive the wrongness of the hoodlum's actions. In sum, the thought is that, in conjunction with the phenomenological motivation, the high-level-perception motivation provides a sufficient reason to consider and take seriously Ethical Perception.

2. The Phenomenal Representation of Ethical Properties

  1. Top of page
  2. Abstract
  3. 1. Preliminaries
  4. 2. The Phenomenal Representation of Ethical Properties
  5. 3. Ethical Perception and Cognitive Penetration
  6. 4. Objections and Responses
  7. 5. Epistemic Dependency
  8. References

If we assume that ethical properties are H-properties, then EP depends on the truth of the high-level view of perceptual content. It is not, however, the aim of this paper to provide a defense of that view.16 Rather, the working assumption will be that at least some H-properties can be represented in perceptual experience. Ethical Perception will here be defended against an objection which is apparently independent of a denial of the high-level view of perceptual content: the Looks Objection17 (LO). As shall become clear, responding to this objection will involve providing an account of the nature of the phenomenal representation of ethical properties, and H-properties more generally.

The LO is supposed to boil down to the deceptively straightforward idea that ethical properties, e.g., wrongness, cruelty etc, don't look (or sound, smell, feel, taste) like anything. More formally, the LO says:

P1: There is such a thing as ethical perceptual experience only if there is a way that ethical properties look.

P2: Ethical properties, e.g., wrongness, cruelty, don't look a certain way.

C: There is no such thing as ethical perceptual experience (EP is false).

McBrayer (2010a) has recently attempted to defend Ethical Perception from the Looks Objection. His method is to consider extant accounts of ‘looks’ (which is clearly an ambiguous term) and to contend that whichever we assume, the LO is unsound. Although McBrayer's response is on the right track, it overlooks the sense of looks which is of crucial import; phenomenal-looks, i.e., what we associate with phenomenal experience in perception. Rather than going into details, consider the following improved response which does tackle the main thrust of the objection.

The response involves distinguishing between two types of phenomenal representation:

(a) phenomenal presence representation, and,

(b) phenomenal presence as absence representation.18

Roughly, the thought is that to be phenomenally present in experience is to be represented in the same sort of robust way in which colours and shape properties get represented. To be phenomenally present as absent is to be represented in experience in some other less robust sort of way, i.e., in a way different from the way colours and shapes get represented.

As will hopefully become clear, the LO involves equivocating on the term ‘looks’. That is, if we interpret ‘looks’ as referring only to phenomenally present representation, then although P2 of the LO, which would claim that ethical properties are not phenomenally present in experience, is probably true, P1, which would limit the representation of ethical properties in experience to phenomenally present representation, becomes dubious. Alternatively, if we interpret ‘looks’ as including phenomenally present as absent representation, then we have the converse result; P1 seems more plausible, but P2 no longer appears to be true. Either way we have insufficient reason to think that the argument is sound.

Before proceeding to discuss the distinction between (a) and (b) in detail, it is worth pointing out that although this distinction is potentially very closely related to distinction between low-level and high-level properties, the distinctions do not necessarily map on to one another neatly. For example, it seems that a high-level theorist who thinks that, e.g., pine trees can come to be represented in experience, could claim either that H-properties get to be represented in a phenomenally present sort of way, or that H-properties are only represented in a phenomenally present as absent sort of way (or both). That said, it will be suggested here that, if it is the case that H-properties—including ethical properties—do come to be represented in experience, then it seems most plausible that they are represented in a way which differs from the way in which colours and shapes are represented.

In order to attain a working grasp of the distinction, consider Noë's (2009) view that, in having an experience, e.g., as of a tomato, in addition to experiencing what might be loosely called the qualities of the tomato which are visible, i.e., the colour, shape of the side of the tomato one is looking at, one also may have an experience as of the backside of the tomato, which is of course in some obvious sense not seen. As he claims:

vision is not confined to the visible. We visually experience what is out of view, what is hidden or occluded… For example, you look at a tomato. You have a sense of its presence as a whole, even though the back of the tomato (for example) is hidden from view. You do not merely think that the tomato has a back, or judge or infer that it is there. You have a sense, a visual sense, of its presence.19

This is putatively a case where there is phenomenal representation of a property but where this is of a different sort from the representation of colour and shape properties, which I am referring to as presence as absence representation.20 A brief point of clarification: I am not endorsing the view that normal agents can visually perceive the backsides of objects. Perhaps they can only infer beliefs about these things. Rather, I am simply using this example to illustrate the possibility of a less robust sort of phenomenal representation, i.e., if there were phenomenal representation of the backside of objects then it would clearly be of a less robust type than that of colours and shape properties.

Consider now Siegel's (2006) case of perceiving the property of being a pine tree after developing a recognitional capacity. It does not seem terribly plausible to think that this property is visible in the same way that, e.g., the colour and shape of the leaves are. Instead, if this property does get phenomenally represented in experience it will be present in visual experience in a way that is more similar to the way the backside of the tomato is present, i.e., in a less robust sort of way. The crucial point is that the properties which are phenomenally present as absent make a difference to the phenomenology of the experience, i.e., they can get into the representational contents of visual experience, but the difference they make to phenomenology is not the same as the way, e.g., colours and shapes effect phenomenology. The suggestion here is that the defender of Ethical Perception ought to say that if there is the phenomenal representation of ethical properties in perceptual experience, then it will most plausibly be phenomenally present as absent representation.

At this point, someone may worry that there isn't obviously a lot in common between the putative perceptual representation of the backside of objects and the representation of natural kind properties and ethical properties. So far all that has been said is that these properties would be represented in a less robust sort of way, which has then been labeled as phenomenal presence as absence. This may seem unsatisfactory and in need of further elucidation.

In order to assuage this concern, consider Macpherson's (forthcoming) attempt to delineate the notions of phenomenal presence and presence as absence. Macpherson points out that there are at least some properties in experience that we are suitably counterfactually sensitive to. For example, David's experience as of the property of being red is one for which the following holds: there are a suitably large number of counterfactual instances in which it is true that if the property of redness had not been present and the rest of the visual scene were held constant (as far as is possible) then David would not have had the experience of redness that he is enjoying. Notice that this holds for other properties such as those pertaining to the shape and position of objects, and the features of the front-sides of objects such as texture, hue, saturation and brightness.21 Macpherson's thought is that we can think of the phenomenally present as all and only those properties in experience which ordinary human agents are counterfactual sensitive to.

Now contrast this with the cases that have been classed as phenomenally present as absent. Reflection on these should reveal that ordinary human agents are not suitably counterfactually sensitive to the presence of these properties. For example, Jim's experience of the property of being a pine tree does not exhibit a similar sort of counterfactual sensitivity to that property. Importantly, if the property of being a pine tree had not been present and the rest of the visual scene were held constant (as far as is possible), then Jim would still have represented the property of being a pine tree. That is, if the pine tree that Jim is now observing were to be replaced by a clever model22 (such that the property of being a pine tree was no longer present), it is not the case that Jim's experience would change. Notice that this holds for other properties such as the backsides and occluded parts of objects, causation, and individuals. Macpherson's thought is that this feature (although, strictly speaking it is the lack of a feature) marks out those experiential properties that are represented as phenomenally present as absent in ordinary human agents.

Consider now the case of ethical properties, and Cat. There is a lack of counterfactual sensitivity here too: if the property of wrongness had not been present and the rest of the visual scene was held constant (as far as is possible), then ethical perceivers would have had the experience of wrongness. If this is correct, then there is a reason for placing the perceptual representation of ethical properties in the same category as the putative perceptual representation of, e.g., the occluded parts of objects. In all of these cases, there is a lack of counterfactual sensitivity that we find in the case of colours and shapes. Moreover, appealing to suitable counterfactual sensitivity can be thought of as one way of filling in the idea of varying degrees of robustness in phenomenal representation.

For further elucidation of phenomenal presence as absence, it is worth briefly considering how Macpherson conceives of the relationship between the two putative types of phenomenal representation. Macpherson draws on the work of Peacocke (1992), who suggests that visual experience has a fundamental content which is presupposed by all other sorts of content.23 Roughly, Macpherson's suggestion is that objects and properties that are phenomenally present provide a sort of spatial framework into which properties that are represented as phenomenally present as absent are added24, e.g., the phenomenal representation of the pine tree property is added to the space occupied by the pine-tree-making low-level properties (that explains why one would point in the direction of those properties if asked ‘Where is the pine tree?’). Importantly, the addition of content associated with presence as absence representation will typically not alter the representation of those properties that are phenomenally present, e.g., in the case of pine tree representation it seems plausible that the low-level property experiences of the expert and the novice could—in principle—be the same. Certainly this seems to be true of high-level properties generally, and hence of the ethical properties that we are interested in.

In summary: if ethical properties are represented then they will be phenomenally present as absent, which we can now think of as (i) the representation of properties in experience that ordinary human agents are not counterfactually sensitive to (in the way described), (ii) the representation of properties that is added to a given spatial framework provided by the representation of properties that are phenomenally present. Putting (i) and (ii) together, we should have a sharper sense of the notion of less robust representation that was initially gestured towards, and a justification for classing together the cases discussed.

Before explaining precisely how this distinction helps in addressing the Looks Objection to EP, it is worth saying a little about why anyone would think that a property which we are not counterfactually sensitive to is phenomenally present as absent as opposed to not being phenomenally present at all, e.g., the presence of the property may simply be judged.25 We have already encountered one attempt to establish that the backsides of objects are present as absent by appeal to phenomenology: according to Noë (see above) it just ‘seems’ to be perceptual. However, such appeals are unlikely to move those not already sympathetic to the view. One way to move beyond controversial appeals to phenomenology is to focus on putative cases where most should agree that the judgment hypothesis is unsatisfactory. A group of potential candidates are what is referred to as cases of amodal completion: cases where it seems that perceivers have some sort of ‘awareness’ of the occluded parts of shapes, such that shapes can be said to be ‘completed’.26 What apparently tells against the hypothesis that the completion is always judged, is that this process sometimes appears to be insensitive to the beliefs that subjects hold. For example, in Figure 1, subjects report that they ‘complete’ an asymmetrical shape in the occluded case, despite the presence of octagons on either side which seem to suggest that the completion ought to take place in a similar way. As has been argued, this suggests that the process of completion is independent of our beliefs and other cognitive states.27

image

Figure 1. Amodal Completion

Download figure to PowerPoint

More generally, one might try to argue in favour of the phenomenal presence hypothesis by appeal to cases where it seems that the supposed ‘awareness’ of putative phenomenally present as absent properties persists despite the formation of beliefs that the property is not present, i.e., cases of known-illusion. For example, suppose that you do have an ‘awareness’ of the backside of a 3-D object (let's say it's a tomato), and that you are told that this particular tomato doesn't have a backside. If the ‘awareness’ of the backside of the tomato persists then this might be a reason to reject the judgment hypothesis, since it is generally accepted that our beliefs are not insensitive to one another in this way. Note, however, that if there is substantial disagreement about the phenomenology of such cases then we may be left with an impasse.

What, though, of the ethical case? It is by no means impossible to imagine cases of known ethical illusion, i.e., cases where subjects know that the relevant ethical property is not instantiated, but continue to have an ‘awareness’ the presence of the ethical property. For example, our experience of cinematic and theatrical depictions of ethical wrongness might be real-life cases. Note that one might favour the alternative hypothesis that we suspend belief when engaging in fiction (and there may be independent reasons for thinking so). If, however, the belief-suspension thesis turns out to be false, then cases of known ethical illusion would at least suggest that some sort of perception-like representational state (i.e., a state that is weaker than outright belief) is persisting in the face of knowledge that there are no ethical properties instantiated. This of course doesn't mean that the perception-like state is itself a perception, e.g., it might be an intuition or an emotion, but as we shall see in the following section, the presence of these sorts of states is not incompatible with the representation of ethical properties taking place in perceptual experience. Indeed, these states might facilitate ethical perceptual experience.

Although what I have said here is by no means conclusive, it should hopefully give at least some plausibility to the idea that phenomenally present as absent representation (as I have characterised it) does indeed take place.

With the notion of phenomenal presence as absence representation in hand, we can see how this provides the basis of a response to the Looks Objection. The thought is that it is only by thinking of phenomenal presence that P2 of the LO seems plausible; it does seem false that wrongness or injustice or goodness look a certain way if we are thinking of this in terms of the way colours and shapes can look a certain way. However, as we have now seen, it is not at all obvious that this type of phenomenal representation is the only type of phenomenal representation, and further, that this is the way we ought to conceive of the representation of ethical properties. Hence it is not at all clear that P1 of the LO comes out as true. Moreover, once we acknowledge both the possibility that properties can be represented in visual experience as phenomenally present as absent, and, that this is the most plausible way to think of the representation of ethical properties, P2 becomes a lot less attractive. Therefore, we have insufficient reason to think that the LO is sound.

To clarify: defenders of EP are in much the same dialectical position vis-à-vis the LO as they are with respect to the claim that the claim that perceptual experience doesn't represent H-properties. In other words, it is not obvious that one can simply rule out ethical properties as being phenomenally represented in experience without prejudging the outcome of an ongoing debate in the philosophy of perception as to what sorts of things can come to be represented in experience, and in what sort of way they are phenomenally represented.

3. Ethical Perception and Cognitive Penetration

  1. Top of page
  2. Abstract
  3. 1. Preliminaries
  4. 2. The Phenomenal Representation of Ethical Properties
  5. 3. Ethical Perception and Cognitive Penetration
  6. 4. Objections and Responses
  7. 5. Epistemic Dependency
  8. References

Having responded to the Looks Objection by providing an account of the nature of the perceptual representation of ethical properties, the task for this section is to sketch a plausible account of how perceptual experience might come to represent ethical properties, i.e., how perceptual experience might come to represent ethical properties as phenomenally present as absent.

There are two general ways in which this could occur: ethical perceptual representation might be hard-wired, or it may be acquired and enter into experience via some psychological mechanism, i.e., either the visual system comes ready-made for phenomenally present as absent representation, or, that only after the acquisition of certain cognitive and conceptual capabilities is this sort of representation possible. Although the hard-wired view is difficult to rule-out, its adoption would appear to involve taking on serious theoretical commitments that defenders of Ethical Perception should want to avoid. Moreover, most commentators and defenders of EP seem to be in agreement that ethical perception involves the acquisition of capacities and that ethical perception plausibly requires a degree of cognitive and conceptual sophistication,28 i.e., they appear to be rejecting the innate-model (at least in a simple form). Given this the working assumption will be that ethical perception is acquired by dint of some degree of cognitive sophistication.

If the capacity for ethical perception is acquired the most plausible account of how this occurs is likely to appeal to the process of cognitive penetration.29 An initial way of thinking about cognitive penetration is simply to understand it as grounding the idea that what a subject non-perceptually thinks can alter the way in which they perceptually experience the world, or that perception is ‘theory-laden’. More specifically, a visual experience, e, is cognitively penetrated if the representational content and phenomenal character of e are altered by states in the cognitive system, e.g., beliefs, concepts, desires, emotions, memories, imaginative states, intuitions, and where this does not merely involve those cognitive states having effects on the subject's visual attention. The idea is that states in the cognitive system are having a more intimate causal effect on the phenomenal character (and content) of the experience. Further, if cognitive penetration is possible then the following scenario will be possible: two subjects could be attending to identical distal visual stimuli, s, whilst having different perceptual experiences, e and e* due to differences in their cognitive economy.30 Applying this to EP, the claim is that a subject who possesses some degree of ethical cognitive and conceptual sophistication could come to have the phenomenology and content of their visual experiences altered by cognitive states, such that they visually represent ethical properties.

Before proceeding, two distinctions ought to be made between types of cognitive penetration. Firstly, cognitive penetration may involve the alteration of phenomenally present representation or phenomenally present as absent representation. As was suggested in the previous section, cognitive penetration in the ethical case will facilitate presence as absence representation, i.e., it will involve the addition of content to a framework provided by the phenomenally present. Secondly, although the content of cognitively penetrated experience is etiologically dependent on other cognitive states, there are at least two ways in which it could be dependent. A weak sort of dependency is exemplified in cases where a subject has an experience as of, e.g., a particular shade of red, due to cognitive penetration, but where the subject does not generally require penetrating states with that sort of content in order to be in that type of perceptual state, e.g., they could have had an experience of that shade of red without cognitive penetration. A stronger sort of dependency is exemplified in cases where the subject does generally require penetrating states with that type of content in order to be in a perceptual state with that content. It seems that ethical cases of cognitive penetration will fall into the latter category of strong etiological dependence, given the assumption that ethical perception is an acquired capacity.

Although the question of whether cognitive penetration (of any type) does actually occur is by no means settled, there are many suggestive studies31 apparently supporting the claim that cognitive penetration of visual experience is indeed possible. Hence, the working assumption will be that cognitive penetration is indeed a psychologically real occurrence. The task will be to clarify how it might occur in the ethical case.

There are a number of possible models of ethical cognitive penetration. The following is not exhaustive. A good starting place is to consider the model Siegel (2006) mentions whilst discussing the possibility of pine tree perception:

A perceiver who can recognise trees by sight seems to have some sort of memory representation, and some sort of perceptual input, such that the input ‘matches’ the memory representation, and the cognitive system of the perceiver registers that this is so… Part of what's at issue… is whether visual experience is an input to such processes of recognition or an output of such processes.32

Call this model Matching. Matching begins with the pine tree expert being taught how to identify pine trees by being shown exemplars. This leads to the formation of a visual memory representation(s) corresponding to pine tree.33 Afterwards, when the perceiver is presented with a visual scene containing the low-level properties associated with pine trees, e.g., the colour and shape distribution, there is a matching with one of the pine tree representations. This is ‘registered’ by the cognitive system such that the content of visual experience includes that of being a pine tree. Now, although Siegel (2006) doesn't put it in these terms, this is one possible way of cashing how presence as absence content gets added to the framework provided by the phenomenally present.

Supposing that Matching is plausible for pine trees, the ethical analogue of the development of a pine tree recognitional capacity is that an ethical expert could come to possess a memory representation(s) corresponding to, e.g., wrongness, and that this is matched with perceptual inputs when there is some relevant correspondence (in terms of low-level features) between the two, e.g., the memory representation for wrongness could comprise of a disjunction of cases that one has previously experienced and has judged to be wrong.

One question which arises about Matching concerns what is meant by ‘registering’ in the cognitive system. There appear to be two general ways in which this might work. Either talk of ‘registering’ is really just another way of saying that there is cognitive penetration of perceptual experience as a result of the matching process, or else, the registering in the cognitive system might involve the formation of cognitive states, which then cognitively penetrate perceptual experience. Focusing on the second of these, consider these ways in which this might work in the ethical case: perhaps when presented with the visual scene in Cat, a matching process occurs that triggers an emotional experience with ethical content, e.g., indignation or horror or outrage, which then penetrates the perceptual experience (note that emotions are being assumed here to have representational contents).34 Another model would involve the matching process triggering an intuition where this may be understood as a non-inferentially formed belief or a non-doxastic intellectual seeming state.

There are other possible ways, besides Matching, in which ethical cognitive penetration might occur. Perhaps no matching needs to take place. Instead, the formation of a cognitive state such as an emotional experience or an intuitive seeming may be directly triggered by what is initially perceptually given to the subject. These cognitive states would then directly cognitively penetrate perceptual experience.35 Another possibility is that an ethical subject's standing ethical beliefs directly penetrate their experience, i.e., a belief whose formation isn't triggered by being confronted with the perceptual scene. In the case of Cat it may be that a subject's belief that torture is prima facie (or conclusively) wrong, or that causing unnecessary pain to sentient creatures is prima facie (or conclusively) wrong, penetrates their experience. Call these models Non-Matching.

Finally, another way in which ethical cognitive penetration might occur involves a subject's belief (or other cognitive state) penetrating their experience, but where this belief doesn't itself have substantive ethical content. In order to understand this type of ethical cognitive penetration, consider the following case:

Shooting: While walking through a bad part of town Brian tells Tom that he is about to see something wrong. Although slightly bewildered by this, Tom forms the belief that I am about to see something wrong. Tom and Brian then round a corner to see a man being shot by a hooded youth. Tom has a perceptual experience as of the all-things-considered wrongness.36

It is assumed in this example that Tom represents the shooting as wrong due to ethical cognitive penetration of the sort discussed. Notice, however, that this constitutes a different sort of cognitive penetration from those previously discussed because the penetrating belief has not been formed as a result of what H.A. Prichard (1912) referred to as ‘an act of moral thinking’.37 Tom might have no view on the wrongness of shooting yet perceptually represents the action as wrong due to the penetration of an expectation belief. Call this the Expectation Model.

The aim of this section has not been to establish the truth of all or any of these models of ethical cognitive penetration. Instead, on the assumption that cognitive penetration is the most plausible way of accounting for ethical perception, the project has been to sketch some hopefully plausible ways in which this might occur. Attempting to establish that ethical perceptual experience does occur is beyond the scope of this essay.

4. Objections and Responses

  1. Top of page
  2. Abstract
  3. 1. Preliminaries
  4. 2. The Phenomenal Representation of Ethical Properties
  5. 3. Ethical Perception and Cognitive Penetration
  6. 4. Objections and Responses
  7. 5. Epistemic Dependency
  8. References

This section will consider and respond to some objections arising from the account of Ethical Perception offered in the previous section.

The Unobservables Objection (UO)38 arises because the account sketched seems to imply that we could have perceptual experiences of things which seem to be, intuitively at least, unobservable. For example, a physicist may observe a vapour trail in a cloud chamber and non-inferentially judge that that is a proton. In this case, although the physicist's judgment may be psychologically non-inferential, we are reluctant to say that the physicist actually perceptually experiences the proton in the cloud chamber, since protons aren't the sorts of things that ordinary human beings are supposed to be capable of seeing. The problem for the defender of EP is that there doesn't seem to be any obvious reason for thinking that cognitive penetration couldn't produce experiences with proton content. More formally the UO says:

P1: If there are ethical perceptual experiences, then there is no reason to rule out the perceptual experience of protons.

P2: There is reason to rule out the perceptual experience of protons.

C: There are no ethical perceptual experiences.

One way of responding is to concede P2 but to attack P1 by adopting a sort of innocence by association response. The thought here is that all philosophers who want to defend the possibility of high-level perceptual representation need to find some way of showing that the H-properties which they claim can be represented in perceptual experience, are not like protons, i.e., they are not unobservable. So proponents of EP are no worse off than defenders of the view that we can perceptually represent pine trees. Put another way, in lieu of a good reason for thinking that there is some important difference between H-properties and ethical properties, it seems that a version of P1 can be brought against just about any H-property. Hence, the argument is in danger of over-generalising.

A more ambitious way of rejecting P1 is to identify some feature that distinguishes ethical properties (and other H-properties) from protons which renders the latter, but not the former, ineligible for perceptual representation. One tentative suggestion is that we cannot perceptually experience protons, electrons, and quarks, etc., outside of the laboratory, because there is not any obvious way in which the phenomenal present as absent representation of those properties could be added to the spatial framework provided by that which is phenomenally present. More specifically, it is hard to think of where in the spatial framework such representation would be added given that phenomenally present properties are ‘macro-level’ properties. Although admittedly vague, the thought would be that this stands in contrast to the case of the property of being a pine tree, and indeed to ethical properties such as being wrong or being cruel where it is easier to envisage the addition of that sort of content. Note, however, that this would not rule out the perceptual representation of protons and other scientific posits in scenarios where it is easier to conceive of how the representation of those properties could be added to the spatial framework, e.g., observing a vapour trail in a cloud chamber.

This leads directly to the possibility of denying P2. Once we have the distinction between phenomenal presence and phenomenal presence as absence representation in mind, i.e., once we have accepted the possibility of varying degrees of robustness in phenomenal representation, the possibility of the phenomenal representation of protons in perceptual experience doesn't seem outlandish. This certainly seems plausible in the vapour trail example. Indeed, if it is possible to have perceptual experiences (and possibly even perceptions) of the backside of objects, and if ‘vision is not confined to the visible’, then the possibility of the perceptual experience of protons seems more palatable (note again that I am not endorsing the view that we can perceive the backsides of objects).

Given the availability of plausible responses, I contend that we have good reasons for thinking that the Unobservables Objection is unsound.

However, even if we think that proton perceptual experiences are possible, someone might claim that proton perception is not (at least for human perceivers). The thought is that if cognitive penetration could facilitate the representation of protons in visual experience, then this just goes to show that this sort of perceptual representation is not direct enough to enable perception. There are two intuitions driving this: (i) we surely don't have direct epistemic access to the presence of protons (they are theoretical posits after-all), and, (ii) perception involves some sort of direct epistemic access (one which stands in contrast to other ways of knowing such as memory). This point might involve making a distinction between perception and what might be called veridical phenomenal representation, where the latter captures the redundant features of accurate visual representation which in some way falls short of perception. For something like this distinction, note the following from Siegel (2009):

[visual] experience could represent (even correctly represent) that causal relations obtain, but these experiences might fail to count as perception, if their relation to what they represent is never sufficiently direct.39

The significance for ethical perception is this: if the representation of ethical properties is enabled by a similar mechanism of cognitive penetration then it is insufficiently direct to ground ethical perception. Call this the Directness Objection (DO) to ethical perception. More formally:

P1: Proton perceptual representation is possible.

P2: If proton perceptual representation is possible then the mechanism which enables it must not be direct enough to enable perception.

P3: The mechanism which enables proton perceptual representation is cognitive penetration.

C1: Cognitive penetration does not enable perception.

P4: Ethical perceptual experience is only enabled by cognitive penetration.

C2: There is no ethical perception.

Given that P1, P3 and P4 are being taken for granted, in order to avoid C2 the defender of Ethical Perception has to reject P2. Here is one possibility: if the mechanism which could enable the perception of protons (cognitive penetration) underlies much of our perceptual experiences, then this might lead us to rethink how we understand perception. Less elliptically: if it turns out that perceiving, e.g., that the person in front of me is my mother, via cognitive penetration is a regular feature of our perceptual lives, then perhaps the view that perception is in some—admittedly loose—sense direct or unmediated ought to be jettisoned.

Even if this is unconvincing, it seems that the defender of EP could accept P2 and concede that ethical perceptual experiences aren't direct enough to constitute perceptions, but maintain that this isn't really a problem for the view. The thought would be that, so long as there is veridical phenomenal representation of ethical properties, this is enough to make EP true. Further, on the assumption that veridical phenomenal representation is a type of mental state that at least sometimes confers non-inferential justification and knowledge, Ethical Perception could still secure the truth of Perceptual Intuitionism.

A brief objection and response: if we have conceded that there can be perceptual experiences of protons then on some theories of perceptual justification, e.g, Perceptual Dogmatism,40 this will entail that we can have non-inferentially justified beliefs about protons. We might agree with Sturgeon (2002) in thinking that this conclusion is ‘not very plausible’,41 i.e., given our epistemic limitations, beliefs about protons are not the sorts of things human beings can have non-inferentially justified beliefs about. However—as will be suggested later—in the cases of cognitive penetration under consideration there may not only an etiological dependence on background states, but also an epistemological dependence. Hence, even if perceptual experiences non-inferentially justify, this may be epistemically dependent non-inferential justification. Arguably, Sturgeon's worry is best interpreted as really being about epistemically independent justification for beliefs about protons, not non-inferential justification per se.42

Consider finally an objection which challenges the claim that ethical veridical phenomenal representation can indeed confer non-inferential knowledge and justification (I am assuming that justification is necessary for knowledge). This objection starts with the observation that perceptual knowledge that F plausibly requires the ability to discriminate Fs from not-Fs, e.g., my perceptual knowledge that that is a red cup depends on my being able to distinguish red cups from blue cups. One way of capturing this is to say that a necessary condition for perceptual knowledge of the presence of a feature, F, is that the perceiver is counterfactually sensitive to the presence of F, i.e., there are a suitably large number43 of counterfactual cases where if not-F then the subject would not represent the presence of F. However, recall from §2, that it is apparently a mark of presence as absence representation that human perceivers are not counterfactually sensitive to the presence of these properties, i.e., if the property, F, were altered whilst holding everything else constant, subjects would continue to represent F. Given this, we might be led to think that ethical perceptual experience is unable to ground perceptual knowledge (even when veridical). Call this the Sensitivity Objection (SO) which says:

P1: If a perceiver, S, has perceptual knowledge that F then S must be perceptually counterfactually sensitive to the presence of F.

P2: No ethical perceiver, S, is perceptually counterfactually sensitive to the presence of an ethical property, F.

C: No perceiver, S, has perceptual knowledge of the presence of an ethical property, F.

Despite initial appearances, the SO equivocates on different senses of ‘counterfactual sensitivity’. As will become clear, the sense of ‘counterfactual sensitivity’ that plausibly makes P1 true, is a sort of sensitivity which is indeed sometimes (perhaps often) possessed by ethical perceivers. So, on this understanding of ‘counterfactual sensitivity’, P2 is false. On the other hand, if we assume the sense of ‘counterfactual sensitivity’, the lack of which makes P2 true, then P1 comes out as false. This is because it is not always the case that perceivers need to be in possession of this sort of sensitivity in order to have perceptual knowledge. Let me explain.

Consider the sort of counterfactual sensitivity that plausibly makes P1 true. On one simple account, the sort of counterfactual sensitivity required for perceptual knowledge is the following: for a subject to perceptually-know that something is F requires that the subject can perceptually discriminate Fs from not-Fs, but where the not-Fs are limited to relevant alternatives.44 We can think of relevant alternatives to the actual case as alternative scenarios where not-F is instantiated which can be found in nearby possible worlds. What makes an alternative scenario nearby (and hence relevant) is determined—at least partly—by what alternative cases one might expect to find in the environment in which one is situated, where ‘environment’ is being understood in a relatively fine-grained way45, i.e., it includes the laws of nature as well as historically contingent factors such as social conventions, trends, climatic factors, etc. To illustrate: if I am looking at a sheep in a farmer's field in Central Scotland, then given the prevalence of cattle farming in Central Scotland, in some nearby worlds where I am not looking at a sheep I would have been looking at a cow instead. However, given that there are no sheep holograms in the fields of Central Scotland, it is not the case that in some nearby worlds where I am not looking at a sheep I would have been looking at sheep hologram, i.e., these are not relevant alternatives.

Now recall the sort of counterfactual sensitivity which renders P2 true, and which is plausibly true for all properties that are represented as phenomenally present as absent. P2 says that, for an ethical property, F, subjects are not counterfactually sensitive to F in the following sort of way: if F had not been present and the rest of the visual scene was held constant (as far as is possible), then ethical perceivers would not have represented F. It should hopefully be clear that this is a different sort of counterfactual sensitivity than that which renders P1 true. If these notions of counterfactual sensitivity come apart in at least some ethical cases, then the Sensitivity Objection is unsound.

To show that this is indeed the case, consider firstly the pine tree expert's putative perceptual knowledge that that is a pine. If the scene remained the same but the pine tree was replaced with a clever model, it is plausible that the perceiver would not be sensitive to this change. However, in many (probably most) environments in which human perceivers are likely to find themselves in, the inability to discriminate an actual pine tree from a pine tree model surely doesn't undermine their knowledge. This is because in many (probably most) environments, cleverly constructed pine tree models are not relevant alternatives. Instead, in many environments, a relevant alternative where the tree isn't a pine would likely be one where it is some other sort of tree, e.g., an oak, and it seems that a pine tree expert would not represent the presence of a pine in that case (and, ceteris paribus, wouldn't form the belief that that is a pine tree). In such cases, their failure to be able to perceptually discriminate between pine tree models and the real thing does not undermine their knowledge in the actual case.

With this in mind, consider Cat. Plausibly, an ethical perceiver could not discriminate cases which look like that and instantiate all-things-considered wrongness, from cases where the rest of the visual scene was identical and wrongness isn't instantiated. It is also plausible that they will have a similar insensitivity to prima facie wrongness. Whether this feature undermines perceptual knowledge depends upon whether such alternative scenarios are relevant alternatives. Presumably alternative scenarios which (non-ethically) look the same as Cat but don't instantiate all-things-considered wrongness would include cases where the youths are being forced to carry out this act, or where they justifiably believe that the cat is evil and must be stopped. Alternative scenarios where prima facie wrongness isn't instantiated would presumably include cases where the cat is a robot and the hoodlums are actors.

Although I find it plausible that the alternatives associated with all-things-considered wrongness in Cat are not relevant alternatives (in many environments in which humans are likely to find themselves in), and that ethical perceivers could have perceptual knowledge of the instantiation of this property, some may doubt this.46 However, even if we are uncertain about whether ethical perceivers could have perceptual knowledge of all-things-considered wrongness in Cat, reflection on the alternative cases associated with prima facie wrongness, e.g., where the cat is a robot, should hopefully reveal that in many (probably most) environments in which human perceivers are likely to find themselves, these alternatives do not obtain in nearby possible worlds and are therefore not relevant alternatives (these seem analogous to the model pine tree alternatives).47 In such cases, our specific sort of insensitivity to prima facie wrongness is not epistemologically problematic. Hence, despite appearances, the SO fails to falsify the claim that Ethical Perception could underwrite perceptual knowledge. To clarify: the SO trades upon two different senses of ‘counterfactual’ sensitivity’ to ethical properties. I have argued that there are at least some cases where they come apart, and hence the SO is unsound.

Despite this, there are plausibly instances of ethical perceptual experience which are straightforwardly epistemically dubious. These are cases where one is in an environement where it is common for scenarios which look like that (non-ethically) to fail to instantiate the relevant ethical property, e.g.,

Apparent Theft: Lauren observes David nonchalantly take money from Bob's wallet. Due to cognitive penetration—e.g., this matches a memorial representation of a previous case judged to be wrong—Lauren represents this as all-things-considered wrong.

Even if David was doing something wrong by taking money from Bob's wallet, Lauren is unable to discriminate this from cases which look like that but where David is not doing anything wrong, e.g., in a case where Bob told David to take the money. However, unlike Cat, this insensitivity is epistemologically deleterious because, in the environment in which Lauren finds herself in (where people often lend one another money and are usually comfortable with friends having access to their wallet), scenes which look like that often instantiate no ethical property. This plausibly entails that cases which look like that, but where wrongness isn't instantiated, are relevant alternatives. Given his insensitivity, it is unlikely that Lauren could get perceptual knowledge on this basis.

One might object that in such cases Lauren wouldn't represent the wrongness, e.g., McBrayer (2010a) (a defender of EP) suggests the following as a necessary condition for perceptual experience:

in order to have a perceptual experience as if X is F, things that look like this (where the ‘this’ picks out a certain phenomenology) are normally (but not always) F.48

This apparently entails that perceptual experience doesn't represent ethical properties in cases like Apparent Theft. However, while something along these lines may be a plausible condition on perceptual knowledge, it is surely not a condition on perceptual experience. Two reasons for this: firstly, it is plausible that we can represent impossibilities in perceptual experience, e.g., the Penrose triangle. However it is never the case that such impossibilities are instantiated. Adopting McBrayer's condition would implausibly rule out this sort of representation. Secondly, reflection on the models of cognitive penetration discussed in §3 should reveal that these are compatible with counterexamples. Indeed, in the case of Apparent Theft Lauren represents wrongness because it just so happens that a visually similar previous experience was judged to be wrong. There seems little reason to rule out such cases a priori.49

5. Epistemic Dependency

  1. Top of page
  2. Abstract
  3. 1. Preliminaries
  4. 2. The Phenomenal Representation of Ethical Properties
  5. 3. Ethical Perception and Cognitive Penetration
  6. 4. Objections and Responses
  7. 5. Epistemic Dependency
  8. References

The following claims will now be taken for granted: (i) Ethical Perception is true according to the cognitive penetration model(s), (ii) at least some ethical perceptual experiences confer non-inferential justification and knowledge. Despite this it will now be suggested that the proponent of Perceptual Intuitionism may struggle to show that the following argument against their position isn't sound:

P1: If Perceptual Intuitionism constitutes an independently adequate intuitionism then the non-inferential justification that it posits must be epistemically independent.

P2: The justification conferred by ethical perceptual experiences is epistemically dependent.

C: Perceptual Intuitionism does not constitute an independently adequate intuitionism.

Call this the Epistemic Dependence Objection (EDO).50 Before reiterating the points made in the Introduction about epistemic dependency, and arguing for P2, it is worth stressing why the conclusion of the EDO is significant. Recall that an important motivation for Perceptual Intuitionism is that it could constitute an independent a posteriori intuitionism, i.e., an empiricist account which is independent from the denial of other accounts (including a priori ones). However, if Perceptual Intuitionism is dependent on other varieties of intuitionism being true, this would mean that proponents will have to defend another epistemically more basic form of intuitionism (perhaps an a priori account).

In order to understand P1 we need to recall the concept of epistemic dependency:

Epistemic Dependency (ED): a state, d, epistemically depends on another state, e, with respect to content c iff e must be justified or justification-conferring in order for d to be justified or justification-conferring with respect to content c.

As was suggested, a paradigm example of an epistemically dependent state is arguably memory. This is because it is plausible that memory can only confer justification (non-inferential or otherwise) for believing that p if one previously had some non-memorial justification for believing that p. It is arguably because of this feature that Memory Intuitionism is, by itself, an inadequate epistemology. Whether or not this view of the epistemology of memory is correct need not detain us. What is important is the general lesson: in order to constitute an adequate epistemology, an intuitionism must ultimately account for non-inferential justified belief in terms of epistemically independent sources. P1 is simply the application of this to Perceptual Intuitionism.

The question of whether P2 is true will be the focus for the remainder of the paper. The strategy will be as follows: an example will be offered in support of P2 and then three objections will be discussed. Although limitations of space mean that the final verdict will be inconclusive, the discussion provides a useful starting point for future debate on the prospects for Perceptual Intuitionism.

Consider the following example (note that due to space-constraints only the Non-Matching Model of cognitive penetration is discussed):

Dog: One day Margaret rounds a corner and is presented with some hoodlums kicking a dog to death. Upon doing so, her emotional response of outrage with content that is wrong penetrates her perceptual experience such that she perceptually represents the wrongness of what the hoodlums are doing.

Two important points related to this case: the first is that, with respect to her belief that that is wrong, it seems that it is Margaret's emotional experience that is really doing the epistemological work. Indeed, one might say that the perceptual experience that that is wrong isn't adding anything to Margaret's justification for her belief. The second regards a counterfactual scenario where Margaret's emotional response is unjustified, e.g., suppose that Margaret's sense of outrage is caused by a malevolent super-scientist who randomly puts Margaret in intensive emotional states. In this sort of case her outrage is unjustified and it is plausible that her perceptual state doesn't justify her either. If this is correct, then Margaret's perceptual state is epistemically dependent.

Rather than present more examples, three objections to this claim will be discussed. Firstly, it may be objected that we can account for why Margaret doesn't get justification for believing that that is wrong when her outrage is unjustified in a way that avoids conceding that the experience is epistemically dependent. The thought is that we can admit that her perceptually-based belief is unjustified, whilst maintaining that the reason for this is that the presence of her unjustified emotion defeats (by undermining) her non-inferential perceptual justification.51 If correct, then her perceptual justification would only be negatively epistemically dependent on the absence of defeaters.52 But this sort of dependency is compatible with positive epistemic independence, and it is this sort of dependency which is at issue.

In response: note this would entail that physicists who have perceptual experiences as of protons can have epistemically independent non-inferential justification for beliefs about protons. This is an implausible conclusion. More generally, it is odd to think that an experience that is cognitively penetrated by epistemically unjustified states could nonetheless justify, only to then have that non-inferential justification defeated by the presence of the very states that cognitively penetrated it, i.e., the same cognitive state would be both facilitating and defeating justification. It is more plausible that the perceptual experience is positively epistemically dependent.

An example might help: suppose that Margaret has some unjustified beliefs that p, q, r, and then forms a belief that s, which is caused by and rationally connected to those beliefs, and that she then forms a further belief that t, which is caused by and rationally connected to p and the other beliefs. It seems incorrect to say that Margaret's belief that t is justified only to have this justification defeated by the presence of beliefs p, q, r. Instead, it makes more sense to say that all of her beliefs are unjustified. The relevance of this to P2 is that it is arguably more-or-less analogous to the perceptual case at issue.53 In response, the opponent of P2 might object that there is an important disanalogy between these cases; perception—unlike belief—is a window or openness to the world. However, it is not at all obvious that there is such a disanalogy with respect to the ethical contents in ethical perception, given that it involves cognitive penetration with strong etiological dependence on penetrating states.

Finally, consider the claim that there is at least one type of ethical cognitive penetration which does not involve epistemic dependency. These are cases of concept penetration, e.g., by a subject's possession of the concept of MORAL WRONGNESS. The objection is that concept possession cannot justify or fail to justify someone in believing a substantive proposition. Concepts simply enable us to be in certain sorts of contentful states. Hence, if a perceptual experience is penetrated by concept possession then, although the perceptual state isn't autonomous (since beliefs are required in order to be in states with that sort of content), the experience is plausibly still epistemically independent.

The most promising line of response involves making the following disjunctive claim: either supposed cases of concept penetration aren't really cases of purely concept penetration, or else they are genuine cases but the penetrated experiences won't be in the business of justifying beliefs, at least not without the presence of other justification-conferring cognitive states, e.g., substantive ethical beliefs. Let me explain. Firstly, call the beliefs required for concept possession C-beliefs.54 My view is that C-beliefs for ethical concepts55 will have very little or no substantive content, i.e., they won't be beliefs issuing in substantive ethical guidance.56 Hence, in cases of ethical perception where we end up with a substantive ethical belief, it is not obvious to me that we should say that these are cases of purely C-belief penetration. Alternatively, if some cases do involve penetration by purely C-beliefs then it is doubtful that the resultant perceptual beliefs are justified unless there are other justification-conferring cognitive states present.

An analogy here might help: it is common to regard analytic propositions as those for which grasp of their content is sufficient for justification,57 and synthetic propositions as requiring more than mere grasp of meaning. The thought is that beliefs formed on the basis of ethical perception by C-belief cognitive penetration will be epistemically akin to believing synthetic propositions on the basis of mere concept possession. More specifically, for those perceptual experiences to justify, it will require the presence of other justification-conferring background states, e.g., substantive ethical beliefs. Someone might object that in cases of ethical perception the agents are doing more than merely entertaining a proposition, they are also getting perceptual evidence for it. In response: it is doubtful that the perceptual state in this sort of case does constitute evidence for believing a substantive ethical proposition. In order to have justification we need “an act of moral thinking”,58 and perception (construed in the way that it has been in this paper) isn't the way in which beings like us do moral thinking.

There are possible responses to this line of thought. One would be to adopt a more substantive account of concepts, e.g., a prototype theory of concept possession. However, the ramifications of this for epistemic dependency are somewhat unclear. In any case, a more comprehensive account of the epistemology of different types of ethical cognitive penetration is required. These, and other issues, are for another paper. For now we can say this much: although a promising account of ethical perception, which survives extant objections, can be given, it is unclear whether the Perceptual Intuitionism that emerges is independent of the adequacy of other forms of intuitionism.59

  1. 1

    E.g., Audi 2004, Hernandez 2011, Huemer 2005, Stratton-Lake (ed.) 2002, Roeser 2011, Väyrynen 2008.

  2. 2

    The formulation of EI is deliberately inclusive, and is meant to capture a range of non-inferentialist ethical epistemologies, including those which don't posit the existence or epistemic powers of ethical intuition (see Väyrynen 2008, Zimmerman 2010, and Sinnott-Armstrong 2011 for similar formulations). A more exclusive formulation would reserve the label ‘ethical intuitionism’ for views which posit the existence of ethical beliefs that are non-inferentially justified by intuition (or are themselves intuitions). A further, more idiosyncratic formulation, would emphasise the epistemic role of intuition but neglect mention of non-inferential justification, e.g., Huemer 2005.

  3. 3

    I will be assuming that justification is necessary for knowledge.

  4. 4

    A good deal less consensus surrounds the addition of emotion and testimony to this list. Hard-headed naturalists will reject the inclusion of a priori intuition.

  5. 5

    See Owens 1999, Senor 2007. Note that I am not endorsing Preservationism. See Audi 1998, Schroer 2007 for an alternative view which allows for memory to be an epistemically generative source of justification. Their explanation for the inadequacy of Memory Intuitionism will be different.

  6. 6

    The notion of epistemic dependency comes from Audi 1997.

  7. 7

    See Audi 2010, Dancy 2010, Cullison 2010, McBrayer 2010a, Väyrynen 2008.

  8. 8

    I shift to propositional justification here to leave open the possibility that ethical agents fail to base beliefs upon ethical perceptual experiences, which on some views would entail that they never have non-inferentially justified beliefs which are justified by ethical perceptual experiences.

  9. 9

    Note that one could hold an alternative version of Perceptual Intuitionism, which doesn't make reference to ethical perceptual experiences, e.g., one which counts intuitive or emotional responses (such as a sense of injustice) to a perceptual awareness of non-ethical properties upon which ethical properties supervene as a sort of ethical perception. I focus on this account because it has been much-discussed in the recent literature.

  10. 10

    See Dancy 2004

  11. 11

    I am following Copp 2003 and Shafer-Landau 2003 in characterising naturalism as an epistemological view.

  12. 12

    Väyrynen 2008 is explicit about this.

  13. 13

    As it is formulated, EP is a thesis about the existence of veridical and non-veridical ethical perceptual experiences, but not perceptions. The question of whether the veridical experiences referred to in EP could count as fully-fledged ethical perceptions is addressed in §4. There are, however, related worries in the vicinity that are ignored in this essay, not least the issue of whether perception is necessarily a causal relation, and whether this complicates matters vis-à-vis the perception of ethical properties which are assumed by many to lack causal efficacy. See McBrayer 2010b.

  14. 14

    Harman 1977. In this, it is presumably all-things-considered wrongness that is supposedly perceived.

  15. 15

    E.g., Siegel 2006.

  16. 16

    See Siegel 2006 for contrast arguments in favour of the high-level view of perceptual content. This involves taking the putative difference in phenomenal character—the what it is likeness—of a given pair of experiences, e, and e*, with the same (or as similar as possible) low-level content and inferring that the best explanation of the phenomenal disparity between e and e* is that they differ in high-level content (note that the success of this depends on the assumption that there can be no change in the representational content of perceptual experience without a change in phenomenal character, and vice versa). So, for example, the phenomenal character of the perceptual experience of an individual who can identify pine trees by sight is plausibly different from the character of their experience prior to acquiring this ability. High-level theorists will argue that the best explanation is that the pine tree expert can now have perceptual experiences of the property of being a pine tree. There are, of course, counter-hypotheses that opponents will push, e.g., that the phenomenal change can be accounted for by a difference in attentional focus. I am not aware of anyone who has presented ethical contrast cases.

  17. 17

    McBrayer 2010a appears to assume that the LO is independent from a denial of the high-level view. However, this seems false given that the LO can arguably be brought against just about any putatively represented H-property.

  18. 18

    As it is being understood, phenomenal presence as absence representation is roughly the same as what some philosophers refer to as amodal representation. The same goes for non-sensorial representation. It is left to the reader to decide on the label they find most appropriate. For discussion of presence as absence see Noë 2009 and Macpherson forthcoming. For discussion of amodal representation, see Nanay 2010.

  19. 19

    Noë 2009, pp. 470–1.

  20. 20

    Note that Noë's view is a bit more nuanced: he thinks that the occluded parts of objects are perceptually present but not represented. See Nanay 2010 and Briscoe 2011 for discussion of how the phenomena of presence as absence/amodal perception may be representational.

  21. 21

    See Peacocke 1992, p. 107.

  22. 22

    The important point about the example is that the only alteration is the presence of the pine tree property.

  23. 23

    Both Peacocke and Macpherson speak of this in terms of scenario content.

  24. 24

    More specifically, she thinks that this involves the addition of visual imagery.

  25. 25

    Thanks to an anonymous referee for PPR for pressing me to discuss this important point, and thanks to Fiona Macpherson for very helpful discussion on these issues.

  26. 26

    Recall that I am taking the phenomena of amodal and phenomenally present as absent representation as equivalent.

  27. 27

    See Kanizsa 1985, Pylyshyn 1999, Nanay 2010, and Briscoe 2011 for discussion and arguments for this conclusion.

  28. 28

    Making this point provides a basis for responding to the Morally Blind Objection; roughly, the claim that in cases of supposed ethical misperception, we are reluctant to say that there is any perceptual error going on. See Cullison 2010 for a response which amounts to the thought that the objection begs the question against the view that, given conceptual or cognitive sophistication, subjects can perceive things over-and-above what are normally taken to be basic objects of perception, e.g. colours and shapes.

  29. 29

    E.g., see Macpherson 2012, Pylyshyn 1999, and Siegel 2012.

  30. 30

    Audi 2010 might be interpreted as claiming that ethical perception can be facilitated by cognitive penetration: ‘suppose, however, that we distinguish between a phenomenal—and, especially a cartographic—representation of injustice and a phenomenal representation constituted by a (richer) response to injustice. The sense of injustice, then, as based on, and phenomenally integrated with, a suitable ordinary perception of the properties on which injustice is consequential, might serve as the experiential element in moral perception.’ (88–89).

  31. 31

    E.g., Delk & Fillenbaum 1965 and Plant & Peruche 2005.

  32. 32

    Siegel 2006.

  33. 33

    Note that there may be similarities with this and the prototype theory of concepts.

  34. 34

    See Prinz 2004, for an account of emotional experiences being triggered by the matching of perceptual content with emotional elicitation files. Also, in her 2012, Siegel suggests that brain architecture (i.e. connections between the brain area V1 and the amygdala) is consistent with there being emotional influences on visual experiences.

  35. 35

    There is another more indirect possible form of cognitive penetration which involves a cognitive state altering or brining into existence some non-perceptual state with phenomenal character, e.g., an imaginative state, which itself modifies the content of perceptual experience. See Macpherson 2012, pp. 30–32.

  36. 36

    It would make no difference to this (or Harman's Cat) if we say that the perceiver is perceptually aware of prima facie wrongness. See W.D. Ross 1930, 2002 for the classic exposition of the distinction between prima facie and all-things-considered rightness. Note that the best way to understand the concept of prima facie rightness (and wrongness) is in terms of pro tanto reasons for (or against) actions. wrongness of this action, but would not have done so had Brian not led him to expect this.

  37. 37

    Prichard 1912, p. 36.

  38. 38

    Väyrynen 2008 and Cullison 2010 discuss a similar worry.

  39. 39

    Siegel 2009, p. 519.

  40. 40

    E.g., see Pryor 2000.

  41. 41

    Sturgeon 2002, p. 202.

  42. 42

    It may have been the case that Sturgeon's 2002 worry was indeed about epistemic independence and that the only reason he formulated it in terms of non-inferential justification is because this was (at least until now) the standard way of formulating ethical intuitionism. Thanks to an anonymous reviewer for PPR for suggesting this point to me.

  43. 43

    Perhaps it isn't simply a matter of the number of cases, but the range and kind of cases, which is important here.

  44. 44

    See Goldman 1976. The account discussed here is a simplified version of that found in Pritchard 2010. There he offers a more complex relevant alternatives account which is designed to deal with problems facing the simple account I present here, posed by the Closure Principle. I do not have space to discuss the more complex account, except to say that the amendments he makes are consistent with the conclusion I reach here: that the Sensitivity Objection is unsound.

  45. 45

    How to say precisely what constitutes an ‘environment’ and to what level of specificity, is a tricky issue which I do not have the space to address. Here I am utilising that notion of ‘environment’ appealed to by epistemologists which makes it true that the unlucky individual who is in ‘fake barn country’—i.e., an area where there are lots of fake barns around and few real ones—is an environment where the presence of fake barns are relevant alternatives when assessing the epistemic credentials of their perception of real ones, and which makes it true that most environments that human beings are likely to find themselves in are not like this. See Goldman 1976 for original description of this case.

  46. 46

    Fully addressing this sort of worry would require giving a full-scale account of what it is for an alternative to be ‘relevant’ (and how to spell out what an ‘environment’ is), which is beyond the scope of this paper. Thanks to an anonymous reviewer for PPR for suggesting this sort of objection to me and for pressing me to be clearer about my argument against the Sensitivity Objection.

  47. 47

    The reader may have worries about our perceptual knowledge in the pine tree case, i.e., the mere possibility that that could be a fake pine tree is enough to undermine knowledge that that is a pine. Further, this would cast doubt on our perceptual knowledge of prima facie wrongness in Cat. Note, however, that this would seem to entail a quite wide-ranging scepticism about perceptual knowledge, which many (myself included) will find implausible.

  48. 48

    McBrayer 2010a, pp. 318–9. Presumably, normalcy is being indexed to environments.

  49. 49

    How extensive the scope of ethical perceptual knowledge is may impact on how reliable that process is. Depending on how we individuate processes (a notorious problem for reliabilist accounts of knowledge), this may have a general impact on the epistemic credentials of ethical perception (if one is a reliabilist about knowledge).

  50. 50

    Cullison 2010 gestures towards this objection.

  51. 51

    Given the example, this would be a normative defeater.

  52. 52

    See Audi 1993 for discussion of negative epistemic dependence.

  53. 53

    Siegel 2013 makes a similar point.

  54. 54

    It is not essential that we assume concept possession is constituted by the possession of beliefs. The discussion follows Pryor 2000.

  55. 55

    This claim may only be plausible for thin concepts like WRONGNESS, i.e., it may not apply in the case of thick concepts like CRUELTY. However, if semantic contextualism about thick ethical concepts is true—roughly that possession of thick concepts involves some quite general descriptive component and an evaluative component whose valence is context sensitive, e.g., see Väyrynen 2011 for discussion—then it is not obvious whether they really do have substantive ethical content.

  56. 56

    Even proponents of ethical self-evidence admit this, i.e., that mere grasp of the meaning of a substantive self-evident truth isn't sufficient for being justified in believing it. Deeper understanding is thought to be required, e.g., Audi 2004.

  57. 57

    E.g., Boghossian 1996.

  58. 58

    Prichard 1912, p. 36.

  59. 59

    Thanks to David Bain, Anna Bergqvist, Ben Colburn, Michael Brady, Jennifer Corns, Stuart Crutchfield, Jimmy Lenman, Fiona Macpherson, Martin Smith, and an anonymous reviewer for PPR for helpful comments on earlier drafts of the paper.

References

  1. Top of page
  2. Abstract
  3. 1. Preliminaries
  4. 2. The Phenomenal Representation of Ethical Properties
  5. 3. Ethical Perception and Cognitive Penetration
  6. 4. Objections and Responses
  7. 5. Epistemic Dependency
  8. References