By Shalom Chalson
What leads to our perception of mind-independent objects? Perhaps, looking at things in front of us stimulates our neurones to fire in specific ways. And the image we “see” is a result of that firing. Or maybe, we see things as they are — the red tomato is seen “redly” or “roundly”. The Causal Theory of Perception (CTP) is an account of perception which had a significant following over the 20th Century. It consists of the (i) causal thesis, (ii) effect thesis, and (iii) conceptual thesis. P. F. Strawson claims that the causal thesis entails the effect thesis. But I don’t believe this to be the case. A subject being causally affected by an object is not a sufficient condition for a visual experience of it. To argue this, I will first explicate the CTP and explain the relationship between (i) and (ii). I will then argue that there is a distinction between personal and subpersonal levels of seeing. This is because one may have visual experiences, but not ‘see’ the object of causation. To illustrate this, I will suggest that one can have a visual experience that does not match the object of causation. I will then suggest that one can have a visual experience that does not contain the object of causation. And last, I will raise and address two objections: the personal and subpersonal distinction does not necessarily contradict entailment, and the awareness of objects of causation must appeal to transparency (I’ll explain these terms in a bit).
According to Paul Snowdon, the CTP aims to establish a conceptual truth about perception (1980: 176). A conceptual truth is something that anyone can acknowledge by referring to their own experiences. It is constituted by the (i) causal thesis, (ii) effect thesis, and (iii) conceptual thesis. The causal thesis involves the claim that if a subject (S) sees a mind-independent object (O), then O causally affects S. Grice notes that it is both ‘necessary’ and ‘sufficient’ that if S sees O, S’s experience is causally dependent on O. The phenomenal character of our experiences also seem to be counterfactually dependent on mind-independent objects in the world. Counterfactual dependency is the notion that had O not been present, or in a position to affect S in a similar way, S would not have experienced the same phenomenal character. (ii) The effect thesis, on the other hand, involves the claim that O must cause a visual experience (E) in both good and bad cases. Good cases are veridical, or successful perceptions, and bad cases are non-veridical, or unsuccessful perceptions. It seems to follow that if our experiences are causally dependent on objects, our visual experiences are constituted by those objects. And the idea that the causal and effect theses are conceptual truths is outlined by (iii) the conceptual thesis.
The conceptual thesis is such that anyone who ‘sees’ can ‘endorse’ both (i) and (ii). It is plausible, however, that an instance of perception can satisfy all the requirements of (i) and (ii) but not appeal to a perceiver. Take, for example, Michael Tye’s robot-thought-experiment as outlined in “A Causal Analysis of Seeing”. To the perceiver in the example, there is doubt as to whether he sees ‘Tim’ or ‘Tom’. Tim causally affects the perceiver and the phenomenal character of the perceiver’s experience is counterfactually dependent on Tim. However, the perceiver would attest to seeing Tom. And William Fish notes that our ‘intuitions remain’ that he sees Tom and not Tim. This suggests that (i) and (ii) are not conceptual truths. The causal affectation of a subject is not reason enough for anyone to endorse having a visual experience of that object. While the conceptual thesis is dubitable, Tye argues for the empirical truth of the causal thesis. He believes that an ‘adequate’ causal analysis can be formulated. Strawson further argues that the truth of the causal thesis entails the effect thesis, or “I see that P entails P.” Here, (i) is a sufficient condition for (ii). If S is causally affected by an object — on which S’s phenomenal character is shown to be counterfactually dependent — there must be a resulting visual experience.
But I am partial to the view that a distinction between the personal and subpersonal levels of seeing is plausible. Daniel Dennett introduces this distinction by arguing that mental entities cannot be reduced to physical entities. On the subpersonal level, external stimuli, such as the perception of an object, cause neurological activity in the brain. David Lewis notes that the scenes we are presented with cause a “pattern of stimulation,” which enables our neurones to fire in particular ways. So, one’s mental interaction with the world is “mediated by sensory and motor transducers” in the central nervous system. On the personal level, the neurological experience is supplemented with a personal vocabulary. At this level, the subject (S) has an awareness of her own experiences. S can now attribute feelings, beliefs, and judgements to her perceptual experience.
The distinction between the personal and subpersonal levels of seeing could show that the causal thesis does not entail the effect thesis. But in order for this to be argued convincingly, a working definition of ‘seeing’ must be formulated. This would include the concepts of ‘matching’ and ‘awareness’. Strawson notes that what we see in visual experiences are not ‘resemblances’ of an object, but the object itself. Therefore, when considering entailment, the question is not if any visual experience occurs. We must instead ask if S has a visual experience of the object that causal affects her. From a third-person perspective, this would be exemplified by matching. A visual experience can be said to match ‘the scene before the eyes’ if the properties of O match the properties of S’s experience. If (i) entails (ii), any resulting visual experience should match the object in (i). And without a suitable account of awareness in visual experiences, S can be satisfactorily causally affected by any object. S could then still have a visual experience, of a good case/bad case common factor. But this would not be an adequate case for entailment. Awareness counts as first-person proof of entailment. ‘Seeing’ can then be formulated as follows: if S ‘sees’ the objects of causation in her visual experience, the objects would (1) match her visual experience, and (2), there would be no distinction between what causes her neurones to fire, and what she is aware of. Failure to fulfil both requirements would show that the causal thesis is not a sufficient condition for the effect thesis. If argued successfully, the distinction between the personal and subpersonal level would suggest that the truth of both the causal and effect thesis in an instance is fortuitous.
In the case of (1), it is conceivable that even if S sees an object (O), and O causally affects S, S’s visual experience may not match O. Take, for example, that S, is confronted with an image on an old macintosh computer screen. The image is of a tree. As this image is digital, it is entirely composed of pixels. Each pixel contains a particular colour which may differ in shade from the next pixel. Upon seeing the computer screen, S has the visual experience of a tree. The phenomenal character of S’s experience is counterfactually dependent on only the pixels. If the pixels were saturated with different colours, S’s phenomenal character would change accordingly. In other words, had the pixels not been on the screen, S would not have her visual experience. Thus, the objects that causally affect S are the pixels on the computer screen and not the tree itself. This is worth considering because S’s visual experience has a ‘suitable pattern’ of counterfactual dependence but no matching. Even if S’s experience meets all of the requirements of the causal thesis, the effect thesis does not hold. S’s experience does not seem to count as ‘seeing’ because her visual experience does not match the objects causing her experience.
It would not be adequate, however, to rely on the failure in one criterion to argue against entailment. As previously mentioned, a test for seeing for the purposes of this essay includes two criteria: (1) matching, and (2) awareness. A failure in (2) can be illustrated by a pathology of perception. Appecerctive agnosia is the condition in which sufferers are unable to recognise an object. While they can perceive on a basic visual level, they suffer from a defect in early perceptual processing. As Eric Ghadiali notes, sufferers of apperceptive agnosia are “unable to access the structure or spatial properties of a visual stimuli and the object is not seen as a whole or in a meaningful way” (2004). Nevertheless, sufferers are stimulated on a subpersonal basis. The phenomenal character of a subject’s visual experience is still counterfactually dependent on an object. However, there is a failure to recognise or match objects. This suggests that though neurones fire, information is not successfully conveyed into a visual experience at the personal level. A sufferer of apperceptive agnosia may have knowledge of ‘x’ and access to beliefs about x, but merely fails to recognise x when it is front of her. The sufferer, therefore, lacks conscious visual perception of which she may be aware. Her ventral stream, or the pathway of the brain that indicates ‘what’ an object is, is “severely compromised”. However, her dorsal stream, or the ‘where’ pathway, remains intact. Thus, she can recognise the spatio-temporal properties of an object, but not be aware of what the object is. As a result, she has a visual experience that (1) matches the object of causation. Yet, S’s visual experience (2) diverges from the neurological or subpersonal level of seeing. Thus, S would have a visual experience that fails the second criteria in a test for seeing. And this further suggests that (i) may apply, but (ii) need not.
Yet a more definitive case against entailment would be one in which neither criterion applies. This would be a case in which a perceiver does not even have a visual experience of the object of causation. For the purpose of this essay, I will use Alex Byrne and David Hilbert’s take on John Hyman’s account of causation. It is outlined as follows: on Monday, Hyman looks for his keys and sees them on the kitchen table. The next day, Hyman looks for his keys again. They are on the kitchen table, but he does not see them. Though the conditions around his perceptual experience are the same, his state of mind prevents him from visually experiencing the object.
Hyman’s visual experience of his visual field is constituted by numerous objects. He may not, however, instantly recognise all the shapes. The causal theorist might then argue that Hyman is not, at this point, causally affected by his keys but by some other object on which he is focused. However, Hyman’s prior knowledge of his keys being on the table primes him to ‘see’ the keys. Furthermore, Hyman’s phenomenal character is still counterfactually dependent on his keys being on the kitchen table. Dennett outlines an argument for this in Consciousness Explained. He notes that when we see a crowd, we can visually tell that the crowd consists of human beings because we recognise features like movement. All at once, we see numerous particular details captured within the visual field. The same can be said of Hyman. If, for example, there were a few more keys scattered within Hyman’s visual field, his visual experience would consist of numerous keys. In such a case, it only becomes more obvious that Hyman is stimulated by details in his visual field on a subpersonal level. Hyman is merely unable to notice that the keys are on the table. This suggests that (1) in a test for ‘seeing’ is not met. His visual experience does not match the objects causally affecting him.
But in order to affirm that Hyman does not ‘see’ the object of causation, the second criteria in a test for ‘seeing’ should be considered. A test for awareness would determine if what Hyman is aware of on the personal level is that which causally affects him on the subpersonal level. Fred Dretske notes that a test for judgement can constitute a test for awareness. when Hyman is aware of the keys, he would “embody” judgements about perceiving the keys. One’s judgements about awareness automatically count as awareness of the particular experience. Judgement occurs when a perceiver predicates a property of an object, of which the perceiver is also aware. In addition to judgement, reportability can also test for awareness. David Chalmers argues that reportability is a “common bridging principle” for experiences. He argues that reportable information is consciously accessible. And therefore, for this essay’s purpose, it should augment judgement in a test for awareness. Reportability is a subject’s conscious verbal attribution of properties to an experience. If Hyman cannot judge or report a visual experience of his keys, there is reason to believe that his “mental entities” are not identical to “physical entities.”
Hyman does not judge or report an awareness of the keys. His behaviour (a continued search for the keys) shows that he believes that they are not on the table. His failure to both judge and report his keys as being on the table in the original case suggests that subpersonal seeing does not entail personal seeing. Firstly, Hyman’s visual experience not match the object he is causally affected by. Secondly, Hyman does not even have a visual experience of the object to qualify for a case of seeing. In passing neither criterion, (1) or (2), for the test of awareness, (i) holds, but (ii) does not.
One could argue, however, that the personal and subpersonal distinction does not necessarily contradict entailment. Mark Johnston argues that “seeing” cannot simply count as matching and awareness. His argument from truthmakers suggests that in the cases presented, the perceivers are aware of and have visual experiences of the objects causally affecting them:
“The immediate judgement is true if the item isolated by attention has the predicated feature isolated from the state by selective attention. So, if one senses the item’s having the feature, and then abstracts the feature, and then predicates it of the item, one will have rearranged the elements of a state into a proposition which one is immediately judging…. The connection is so intimate that it has been mistaken for identity.” (Johnston, 2006)
Johnston argues that what a perceiver predicates of an item in a visual experience is, by nature, an abstracted property of the item. An abstracted property is whatever is “immediately judged on the basis of sensing.” He claims the nature of perception is that it is directed at “spatio-temporal particulars” and not the factual properties of an object. To think of reportability as an indication of the actual properties of an object would be to misunderstand perception. This is because perceptual judgements are “made out of” sensible properties that do not have to be facts. Johnston calls these sensible properties truthmakers. In other words, an object can causally affect a subject, on which the subject’s phenomenal character is counterfactually dependent. The subject is aware of some property related to the object, which is naturally abstracted, but immediately judged. And this property does not have to match the object of causation. So even if there is a distinction between the personal and subpersonal, it would not contradict entailment.
However, there are difficulties with this thought. According to Alva Noë, what we judge or report of a visual experience has nothing to do with the ways in which our neurones fire, or how our brains cognitively affect our visual perceptions. As long as a distinction between the personal and subpersonal level of seeing is made, entailment is no longer plausible. Johnston commits to such a distinction when he claims that properties are abstracted. The perceptual theorist can either be committed to entailment or a personal-subpersonal divide, but not both. The Causal Theorist is not satisfied as the visual experience no longer has to relate to the objects of causation. If a property can be abstracted, there is no theoretical limit to how the property is changed on the personal level. And this is even if the abstracted property is immediately judged. The proponent of the subpersonal-personal divide will also not be satisfied. Johnston limits the properties of experience to only what is abstracted of an object. But this would still not address the case in which S does not have a visual experience of the pixel (how then does S see the tree?). As a result, Johnston’s analysis would satisfy neither the Causal Theorist’s account of entailment or the personal-subpersonal distinction.
The Causal Theorist might then claim that the account of awareness outlined is not what it is to reflect on an experience. When we report our experiences, we make reference to objects. We do not aim to be accurate with respect to the objects we perceive. Furthermore, our experiences are transparent, not certain. The Causal Theorist would argue that an account of awareness should appeal to the first-person notion of perception. The phenomenology of experiences are, therefore, best explained by transparent accounts of awareness. Transparency is the claim that what constitutes a reflection of ‘what it is like’ to have an experience of something reveals properties of the objects themselves. Tim Crane notes that this produces two claims, that we are aware of objects of experience, and we are not aware of properties of experiences in themselves. The Causal Theorist would argue that when we reflect, we are aware of the properties of objects. As a result, we are directly stimulated by objects. In the example above, S sees the tree. The pixels are arranged such that they produce this image. As a result, the tree is a property of the pixels that causally affect S. This would satisfy entailment. And so, awareness should also appeal to transparency.
Nevertheless, Crane notes that we may not even be able to ‘attend’ to transparent properties. The Causal Theorist can argue that what we are causally affected by are objects that correspond to transparency in awareness. However, there are properties of experience that do not exist within objects themselves. Suppose that S removes her glasses and reports, “everything is blurry”. The property of blurriness does not exist in the objects of experience but in the experience itself. This is a ‘straightforward’ case. There is ‘direct awareness’ of properties of an experience but not properties of the objects of experience. S is aware of blurriness and not the objects around her. The Causal Theorist, therefore, cannot argue away from this without appealing to other theories of perception. But a distinction between the personal and subpersonal levels can just as easily explain such cases. As the objection does not hold, the distinction between the personal and subpersonal levels of seeing is plausible. And as a result, the causal thesis still does not seem to be a sufficient condition for the effect thesis.
To conclude, in order for (i) the causal thesis to entail (ii) the effect thesis, a perceiver must ‘see’ the object of causation. It is plausible that a subject has a visual experience that (1) does not match the object of causation. It is also plausible that a subject is (2) not aware of the object of causation in her visual experience. In these cases, the subject cannot be said to see the object of causation. As a result, a subject being causally affected by an object is not a sufficient condition for a subject’s visual experience of that object. And the causal thesis does not entail the effect thesis.
Image Credits: PC World, N2N, Physics, psychology.wordpress