by Sophia Decherney

Abstract: In this paper, I will discuss each of the primary solutions to the character question and argue for the External Reductionist Representationalism view. I will then defend the view against the objection that it does not provide an adequate account of pain.

1. Introduction to the Character Question

A central question in the field of Perception, referred to as the character question, asks, “What constitutes the character of experience?”.[1] In other words, what does it mean to experience something? There are many answers to this question, each with its own problems and advantages. In this paper, I will discuss each of the primary solutions to the character question and argue for External Reductive Representationalism. First, I will explain the Sense Datum View, largely regarded as an inevitable conclusion from the Act/Object Assumption, an assumption that states that there must be an object of perception. I will explain how the Sense Datum View presents significant problems, such as perceptual imprecision and failure to explain major metaphysical questions. I will then shift to the Representationalist View, currently the predominant view amongst philosophers of perception, and explain how it answers some of the questions that the Sense Datum View faces. Lastly, I will explain two different options within Representationalism–Internal Non-Reductive Representationalism and External Reductive Representationalism–and prove that while the External Reductivist Representationalism view leaves some questions unanswered, it provides the most compelling answer to the character question.[2]

2. Answer One: The Sense Datum View

2.1. The Sense Datum Answer to the Character Question

The Sense Datum view answers the character question by saying that to have a sensory encounter with a particular character is simply to experience Sense Data. Sense Data are non-physical objects of perception created by the brain, the spatial-temporal properties of which roughly match those of physical objects in the world. Imagine you and your sibling are both looking at an apple. According to the Sense Datum View, when you look at the apple, you do not see the apple. Instead, your brain creates an image of an apple, a Sense Datum of an apple, and projects it, similarly to how an Oculus Rift or a projector produces an image and makes it seem as if the image exists in the world. Having explained the Sense Datum View, I will now explain the arguments in favor of the view before moving onto the arguments against it.

2.2. The Argument for Sense Datum

The argument for the Sense Datum View, the Argument from Hallucination, starts with an assumption called the Act/Object Assumption. This assumption states that having an experience with a particular character is to experience that object, and further, that all objects of perception must be either physical or non-physical. The assumption is apparent on an intuitive level: it is natural to think that what you perceive is an object of your perception. In fact, when people speak about perceptual experiences they use objects as well, saying, “I saw it” or “I heard that.” For example, even in a hallucination, we say that we “saw a purple flower,” whether or not the object (purple flower) physically is there or just a mental picture. The next step in the Argument from Hallucination is to say that since the objects of hallucinations clearly exist, but do not exist in the external environment, they must be non-physical. To generalize from that statement, if hallucinations exist as non-physical objects of perception, all perceptions must be non-physical like hallucinations because there is no reason to believe that hallucination would be different from any other perceptual experience. These non-physical objects are named sense data.

2.3. The Arguments Against Sense Datum

There are two significant issues with the Sense Datum View. The first is that sense data are perceptually imprecise. When you look at an object, other objects also exist in the periphery of your vision. According to the Sense Datum View, these objects are sense data as well. Yet, unlike the sense data in the central part of your vision, the periphery objects are inexact, maybe fuzzy. Certainly, you would not be able to represent them precisely if asked to draw or describe them.

So what kind of objects are these sense data that your brain creates imprecisely? The first option is that they are, by nature, metaphysically imprecise non-physical objects. Yet, it is intuitive that there could not be such an object, since an object, by definition, should have defined features even if it is non-physical. A second option on the nature of the above object is that peripheral sense data is metaphysically precise, so your brain makes incomplete sense data based on the information it has. It is as if the brain, lacking information about the object, therefore represents it as imprecise.

This explanation, however, makes little sense, because it asks us to believe that our brains, having gone through the process of creating precise sense data, then fail to pick an accurate color, shape, etc., for the object. If the brain is able to create sense data without reliance on the outside world, there is no reason for it to create sense data that are physically imprecise. In this way, neither answer to the problem of perceptual imprecision succeeds in resolving the issue of peripheral impression of sense data.

The second problem with the Sense Datum View is the problem of the phenomenology of how a sense datum is created and projected. Sense data theorists claim that the object of experience is irreducible beyond sense data. Yet, they cannot explain how the sense data are made, how they exist non-physically, and how their brains can project them with such accuracy. Do sense data exist outside of physical space in their own mental realm? If so, how do they come to be projected in the physical realm?

This mysteriously irreducible process suggests that the Sense Datum View is not useful in answering the character question. Although, according to the Sense Datum View, to have an experience with a specific character is simply to experience Sense Data, the failure to explain how these mysterious sense data operate continually brings us back to the original question: what constitutes the character of experience? The Representationalist View will answer these questions by explaining perceptual imprecision and by positing a more precise description of the metaphysics of perception.

3. Answer Two: Representationalism

3.1. The Representationalist Answer to the Character Question

The Representationalist View answers the character question by arguing that to engage with an experience is to represent an array of perceptual properties experimentally. But what does this seemingly circular definition really mean? To understand this view, take the case of hallucination. The Representationalist View also utilizes the “seems gambit” in the case of a hallucination. The “seems gambit” means that if you hallucinate an apple, you do not actually perceive an apple object – physical or nonphysical; instead, it just seems as if there is such an object. Having no need for an object of perception is a significant departure from the Act/Object Assumption, which states that such an object must exist in actuality. In experience, according to the Representationalist View, it seems that an object is there, but in actuality, it is merely a representation that your brain creates.

3.2. Argument for Representationalism

The main argument for Representationalism is that it does a better job than the Sense Datum View in explaining major problems in perception. First of all, it currently offers the most plausible explanation of the external directedness and internal dependence problem. The external directedness and internal dependence problem is the intuition that perception relates both to things out in the world and to our own internal processing of perceptions. Take the example of noise: it is intuitive that noise is simultaneously something that we process internally, and something we believe is also out there in space. Representationalism incorporates this intuition nicely by saying that although there are physical noises out in the world, your experience of hearing is just for your brain to make it seem to you (represent the noise) that there is noise in the space around you.

Representation also avoids the problem of perceptual imprecision. If your brain merely represents sensory experiences, it makes sense that, when the level of detection varies, so will the level of precision and completeness. To comprehend this view, it is helpful to picture representational experience as language. Language, when heard, conveys the understanding and meaning of what the speaker intends to say. Yet it is not in itself the meaning. For example, the word “apple” represents an apple, but while it may trigger the idea of an apple, it is not itself an apple.

Experiential representation also explains the problem of the object’s location with which, as we saw, the Sense Datum View struggled. In Representationalism there is no need for a non-physical or mental realm because these representations are existence-neutral, meaning that in order for the brain to represent something, there need not exist anything for it to represent. While the actual nature of these mental representations may still be mysterious, in these ways, Representationalism far exceeds the former views in answering critical questions with precision.

3.3. Questions for Representationalism

The main question for Representationalism is how much representations rely on internal and external factors. In other words, to what extent does representationalism require there to be something in the world for the brain to detect? How much of representationalism is reliant on internal directedness as opposed to the external environment? For example, is it that colors are real and that our brains are detecting and representing them, or is it that objects are colorless and the brain merely represents them with color properties? How reductive is experienced representation?[3]

These questions will be answered differently by different views within representationalism. I will consider two options. The first option is the Internal Non-Reductionist View of Representation, which will favor internal dependence. According to this view, color does not exist in the world but rather is only represented by our brains. This view faces many of the same problems as the Sense Datum View. The second option is the Reductionist Externalist View, which grounds representations in the external world. For example, this view would have us believe that our brain is representing real colors that exist in the world. but the image we see only “seems” to be those colors. I will explain the Internalist Non-Reductive View first, and compare it to the Sense Datum View, explaining the similar problems it faces.

4. Non-Reductive Representationalism

4.1. Non-Reductive Representationalism’s Answer to the Character Question

If you remember the Sense Datum View mentioned above,you have already grasped the basic intuition about Internalist Non Reductive Representationalism. Internalist Non-Reductive Representationalism holds that sensible qualities like color or smell are only internally dependent on experientially represented experiences; they are not “out there” in the world. However, Internalist Non-Reductive Representationalism does not hold that experience can be further reduced to any physical object or process. To experience the color red is simply to experientially represent the color red, and further, there is no property of red attached to any object in the real world parallel to this experience. This understanding of the location of sensible properties contrasts the reductive internal views that claim to further describe the process of experientially representing or the external views that rely on the physical external world to influence experiential representation.

Although the proponents of Externalist Non-Reductive Representationalism agree that representing red depends on neural processes in the brain, they also believe that this representation cannot be reduced to those properties.

4.2. Arguments for Internist Non-reductive Representationalism

Unlike the Sense Datum View, the Internist Non-Reductive Representationalism avoids problems of perceptual imprecision. Like all representationalist views, Internalist Non Reductive Representationalism not only explains but also predates such cases by maintaining that the brain needs adequate information for representations. It also avoids non-physical objects such as sense data, therefore circumventing questions concerning location and space. Furthermore, Internalist Non Reductive Representationalism offers a strong account of internal dependence. Lastly, Internist Non-Reductive Representationalism predicts the weak correlation between external stimuli and internal response, as found through numerous neurological studies.

4.3. Problems with Internist Non-Reductive Representationalism

Many of the same phenomenological issues associated with the Sense Datum View are also found in Internalist Non reductive Representationalism. The non-reductive part of the view makes the view phenomenologically unclear. If experience cannot be reduced further, how do those representations come to be? What are they based on? Why are they so similar to other people’s representations? How are they projected on the world? This phenomenological confusion makes the view impractical as an explanation of experience, resulting in more questions than answers. Because of its non-reductive nature, Internist Non-Reductive Representationalism gives very little new information about the nature of experience, maintaining that experience is experientially representing and does not in itself explain what experience consists of. In this way, Internist Non-Reductive Representationalism fails to be useful in answering the character question.

The view also gives a concerning answer to the external directedness part of the external directedness and internal dependence problem by failing to give an account of how the physical world affects perception. The view claims that experientially representing red is distinct from any physical conception of “red”. However, they believe that experience depends on the physical. This complexity leads to an almost dualist account of perception. I will now turn to the second option within Representationalism, External Reductive Representationalism, which will preserve the benefits of Representationalism while at the same time addressing the concerns discussed above.

5. Externalist Reductive Representationalism

5.1. Externalist Reductive Representationalism’s Answer to the Character Question

Externalist Reductive Representationalism believes that representations are grounded in the external. To get an intuitive feel for this view, picture a thermometer. A thermometer represents weather, yet the thermometer is not the weather; it is only representing its own detection measure of the external factor of weather. A thermometer that is designed to measure temperatures up to 90 degrees will represent 100 degrees as 90 just as a brain lacking proper information will experientially represent a hallucination. In this way, Externalist Reductive Representationalism demystifies the relationship between experientially representing the external world and the external world itself. The brain detects sensible properties such as color that exist in the world and experientially represents them with the information it possesses.

5.2. Arguments for Eternalist Reductive Representationalism

Externalist Reductive Representationalism does the perfect job of resolving the internal dependence and external directedness problem. It is essentially a view that believes experience is both internally and externally grounded. It also explains experience in physical terms thereby de-mystifying the nature of experience representation. Like all reductionist views it is able to avoid the Act/Object assumption and the problems that follow with that assumption like perceptual imprecision, location, etc. Moreover, Externalist Reductive Representationalism predicts and explains hallucination very well by utilizing the “seems” gambit and by giving an account of how hallucinations come to be represented.

Lastly, Externalist Reductive Representationalism offers an account of how evolution affects perception by explaining that an evolutionary history is a prerequisite for experience, just as syntax is a prerequisite for language. So it makes evolutionary sense that this perceptual account would be a selected trait. Cognitive American psychologist Donald Hoffman, suggests that, since our limited experiential representation of reality can only represent so much information at a time, the brain chooses the information that has seemed evolutionarily most necessary (The Case Against Reality). Therefore, our perception focuses us on essential experiential representations, allowing the brain to process its environment quickly and decisively.

6. A Major Objection to Externalist Reductive Representationalism

A major problem for Externalist Reductive Representationalism is the problem of how to treat pain. It has been proven that there is a weak correlation between external causes of pain and our experience of it internally. According to Externalist Reductive Representationalism, when the amount of pain increases externally the experience of that pain should increase with relative correlation. This example is similar to that of the thermometer: if the temperature increases outside, the thermometer should show a consistent increase as well. However, studies show that there is a very weak correlation between the levels of external sources of pain and the internal experience of pain. Stevens et al. found in their study on the correlation between the intensity of electric shocks and the internal experience of pain that it “seems to be universally true on prosthetic continua, the category scale is concave downward when plotted against the scale of subjective magnitude” (Stevens 330). In other words, for each of their three methods of determining internal pain intensity (appearance of intensity, neuron response, and self-reporting), the correlation, while positive, was weak between increased external intensity and internal experience of pain (this will be referred to as the weak internal good external correlation problem). However, the correlation between neuron response and reported pain intensity is strong, suggesting that the experience of experientially representing pain is entirely internal (Stevens 335).

Another problem for Externalist Reductive Representationalism relating to pain has to do with exceptions to the traditional model of pain. For example, phantom pains (pains that appear where a missing body part should be) appear to have no external source yet it is hard to deny that they are pain experiences. In these cases the Internal Physical Model or the Internal Non Reductionist Representational View works better because not only does it predict a weak correlation between external and internal experiences of pain, but it also accounts for pain based on mental states that have nothing at all to do with the outside world (Hill 80).

The folk intuition about pain also supports this theory. Almost all statements about pain locate it in the body rather than in something “out there” that we experience. One would say “I have a headache” or “I am in pain” rather than “I am experiencing a pain that is in the door stubbing my toe.” To understand this difference take the example of color. When talking about color our intuition is that it is located in the object we are experiencing, i.e., “I see that brown door.” By contrast, pain appears to us to be only within our own experience of it, i.e., we say “my toe is in pain” not “that door that stubbed my toe is painful.” This intuition goes against the Externalist Reductive Representationalism view of pain, which always seeks to locate pain in the external world.

7. Addressing the Problem of Pain in Externalist Reductive Representationalism

7.1. Solution One

A simple answer is that pains not grounded in reality (such as phantom pains) are merely hallucinations. This resembles the case of the thermometer which, because it only goes to up to 90 degrees, represents 100 degrees as 90 degrees. While we do not know exactly why the hallucination exists, that is more an issue of the causation of pain than the perception of pain experienced. Thus it is not a requirement of Externalist Reductive Representationalism to answer the question of the cause of pain-related hallucinations.

As explained earlier, Externalist Reductive Representationalism offers a great explanation of hallucination, so it could be objected that phantom pains are not only explained, but also predicted by the Externalist Reductive Representationalist View. It would make no sense that every other sense – sight, sound, etc.– would experience representation and only pain would be required to respond solely to external stimuli. I think that this objection offers a clear defense of Externalist Reductive Representationalism with regard to pains not caused by external stimuli. However, it does not address the weak correlation between internal experience and outside stimuli in cases where there are outside stimuli present. To address this issue I will lay out the three main ways that Externalist Reductive Representationalists explain weak correlation and evaluate each answer based on its compelling evidence and ability to fully set aside the concerns about pain.

7.2. Solution Two: Representing our Own Detection of Pain

One answer to the weak internal good external correlation problem is that when we externally represent pain, we are detecting not external sources of pain, but instead, our own physical response to pain. For example, when we feel a migraine for which there is no external source. what we are actually experientially representing is the physical response of our brain or spinal cord. That would explain, and even predict, the strong correlation observed between neuron response to stimuli and reported pain levels.

One positive result of locating the detection of pain in our body is that the result manages to find a physical source of pain. For those who are attracted to the Externalist Reductive Representationalist View because of its physicality, this is a very compelling resolution. Another positive outcome is that this answer neatly predicts the correlation problem. The negative outcome of this view is that, while the view does locate a physical source of pain, it does not locate an external source of pain as the spinal cord and brain are still internal.

Without an external source of pain, the relationship between hallucination and normal cases remains ambiguous, since phantom pain can still be located in the spinal cords and neurons. Additionally, the answer suggests that pain could not exist independently of human perception of it. It is one of the tenets of the Externalist Reductive Representationalist View that sensory qualities, such as color, shape, etc., pre-dated humanity and continue to exist whether or not they are perceived. In conclusion, this solution, while compelling, may demand a large phenomenological shift in the Externalist

Reductive Representationalist View.

7.3. Solution Three: Normal/Abnormal

A third answer to the problem of pain is that all pain is partially an illusion. That would mean that, while pain exists externally out in the world, our brain always tracks it a little incorrectly and thus all experienced representations of pain are partially illusions. This does fit the evidence, since experience of pain has a positive but not perfect correlation to external sources. The positive of this explanation is that it manages to locate pain entirely externally.

The negative is that Externalist Reductive Representationalism does not want to normalize illusion as an answer for the other senses. Differences in the experiential representation of color, for instance, between color-blind and non-color-blind individuals, is not explained this way. So the explanation forces Externalist Reductive Representationalism into a dualist view in which pain is treated differently than every other perceptive quality. Unless a compelling reason is found why pain should operate differently, it seems unnecessary to resort to a dualist view and makes the Externalist Reductive Representationalist View less compelling overall because it is no longer a good view for simplicity concerns.

7.4. Solution Four: Divorcing Externalist Reductive Representationalism from the Folk Perception of Pain

Unlike other types of sensory perception the common or “folk” perception of pain is that to experience pain just is to feel pain (Hill 73). Significantly, we do not believe this about any other sense. For all other senses we distinguish between appearance and reality. For example, an apple might seem green to a regular person and red to a color-blind person. However, we would never say that the apple is red just because the color blind person perceived it as such. Nor would we equate seeing a hallucination of a red flower with actually seeing a red flower. Nor would we say that we failed to see the color of the apple just because there are animals that can see it in more colors than us. Thus, in all cases besides pain, we can successfully utilize the “seems” gambit to understand that appearance versus reality exists in regards to sensory perception.

This solution to the problem of pain can be stated in this way: it is not that our view of perception does not fit with pain, but rather that our view of pain is incorrect. This solution is strongly supported by the failure of any view to completely account for pain. Christopher Hill in his paper “Fault Lines in the Perception of Pain” suggests redefining pain using the “seems” gambit. In his view, all pain is based on some sort of external bodily disturbance. If one feels a pain such as a phantom pain with no apparent cause then one is not actually in pain, they just seem like they are what their brain is detecting and experientially representing. In the same way, if you see a red flower where no flower exists,you are simply hallucinating the red flower.

This solution also addresses opposite examples in which external causes of pain are present but the person does not feel pain. For example, when one has surgery but is under anesthetics, it seems to one as if they are not in pain but in reality they are experiencing external causes of pain. The solution would remain true for cases such as pain asymbolia, a condition where one is unable to feel pain. This answer would be a projectivist account of pain since pain is still found in external bodily disturbance. This account of pain still leaves some questions unanswered, such as the way that these hallucinations of pain work. However, this is a problem best served by neuroscience and essentially does not detract from the view since Externalist Reductive Representationalism already has an account of hallucination.

7.5. Evaluation

In my view, the fourth and final answer is the best answer to the question of pain. It separates accounts of pain described by the hallucination and normal/abnormal answers, while explaining how this view does not create a dualist account and, in fact, works the same way as experientially representing the sense of sight. It also predicts the weak correlation between pain and external sources by eliminating internal experience as the criteria for pain intensity. In these ways, it solves the problem of pain for Externalist Reductive Representationalism while leaving open the possibility of further research on the nature of the relationship between external disturbance and pain.

8. Why is Externalist Reductive Representationalism the Most Compelling View?

Unlike the Sense Datum view, Externalist Reductive Representationalism does not need any assumptions like the Act/Object assumption, does not require any objects that are metaphysically imprecise, and has a clear internal and external relationship. In contrast to the Internalist Non-Reductive Representationalism the Sense Datum View, Exterlist Reductive Representationalism’s phenomenology is not mysterious and it is an externally grounded view. The Externalist Reductive Representationalism successfully reduces the nature of perceptive experiences, answering the character question, and giving an account of the external directedness and internal dependence problem. For these reasons, although there are still research questions, the balance of considerations is strongly in favor of Externalist Reductive Representationalism’s view of perception.

References

Coghill, Robert, et al. “Pain Intensity Processing Within the Human Brain: A Bilateral, Distributed Mechanism.” National Institutes of Health, vol. 82, 1999.

Cutter, Brian, and Michael Tye. “Tracking Representationalism and the Painfulness of Pain.” Philosophical Issues, vol. 21, 2011.

Hills, Christopher S. “Fault Lines in Familiar Concepts of Pain.” Brown University, 2022.

Stevens, S.S., et al. “A Scale of Apparent Intensity of Electric Shock.” Journal of Experimental Psychology, vol. 56, no. 4, 1958.

Hoffman, Donald. Case against Reality: Why Evolution Hides the Truth from Our Eyes. W W Norton, 2021.

________________

[1] Character meaning the ways in which objects appear to you.

[2] There are two other main views of perception not represented in this paper: Naive Realism and the Internal Physical state view. Naive Realism is not represented because it fails to be a unified view. It is reliant on sense data as an explanation of hallucination so therefore any objection to the Sense Datum View also holds for Naive Realism. The Internal Physical View will be briefly mentioned in response to Representationalism but quickly dismissed because it provides no account of external directedness.

[3] Reductive meaning how much can the experience be further broken down in more basic processes.