July 16, 2012

What you see is what you grasp

In my previous blog I promised to get back to the question of what tempts us to accept accounts of other people’s experience in the style of Gopnik’s account of what it is like to be autistic. What I had in mind was something like the following idea: “Suppose I can see that those beings over there are human beings and you can’t see it. That means that it must be given to me in what I see that they are human, whereas it is not given to you. Hence they must look different to me than to you.”
                      I want to try out the suggestion that this is connected with a general inclination, in speaking about seeing, to focus on locutions of the form “seeing an X” (”seeing a lion / a fire / clouds”) – what we might call the “direct-object form”. Less attention is usually paid to locutions of the form ”seeing that ---” (”seeing that a lion is hiding in the bush / that something is burning in the fireplace / that the sky is clouding over”). One reason for this may be that we don’t think very much about the difference between these forms, we just assume that it does not really matter which form we focus on, and find it simpler to speak about direct-object locutions (“After all, in either case, we’re talking about the same thing: seeing”). 
Another, slightly more sophisticated reason may be that we are inclined to think of the direct object form as simpler or more basic than the seeing-that form, in either a logical or an epistemological sense, or both: 
- logically, in the sense that all cases of seeing-that are taken to involve the corresponding forms of direct-object seeing (“seeing that the cat is on the mat consists in seeing the cat and the mat and the relation between them”); 
- epistemologically, in that cases of seeing-that are taken to be based on the corresponding forms of direct-object seeing: “it is through seeing the cat and the mat and the relation between them that I am able to see, hence to judge, that the cat is on the mat.”
In short, on this view, the direct-object form gives the most concrete account of what seeing involves.
Differently put, whenever I see that ---, there must be some state or entity which mediates my seeing: there must be a look of things, a set of percepts or sense-data or visual appearances, or a way in which I am appeared to. 
                      (On the other hand, it will be agreed that the seeing-that locutions generally give a more direct expression to how my seeing bears on my activities. My seeing the red traffic light explains my stopping the car only provided I saw that there was a red traffic light. For I may see the red light without seeing that the light is red.)
                      Here I want to set out briefly what I think about this approach to seeing (I realize that I am here trying to cover enormous ground in a few steps; possibly I’m getting in over my head, and I’m ready to be corrected):
(1) Actually, what the “see-that” locutions capture is usually more specific and concrete than what the direct-object locutions capture. The use of direct-object locutions is more varied and harder to survey than seeing-that locutions. 
On some uses, seeing an object is compatible with not realizing what object one is seeing, on other uses it is not.  We sometimes use a locution like “She saw the red traffic light” together with qualifications such as “but she doesn’t know what traffic lights are”, or “but she couldn’t make out what it was”, or “but she is colour-blind and thought it was green”, or “but she was thinking of something else”. On other occasions, we imply that the seer does grasp what she is seeing. (On those occasions, we could mostly have used a see-that locution instead.) 
The context will make it clear in which way the word “see” is being used; however, when doing philosophy we tend to gloss over those distinctions, and to ignore the role of the context. The consequence is that when philosophers discuss seeing, they apparently have only a hazy idea of what precisely they are discussing. (I plan to give examples of this in a later blog.)
(2) On the other hand, it is usually quite easy to grasp how what will count as seeing that --- depends on the particular context of speaking. Thus, we can easily imagine how “She saw that there was food on the table” would fit into a context in which she was starved and looking for something to eat or where she had been asked to change the table-cloth or where it meant that dinner was being served; on the other hand, without some such story we would not know what to make of the ascription of seeing.
(3) Focusing on the direct-object locution tempts us to concentrate on the first person case: we contemplate “what it is like” for us to see things, at the cost of asking in what circumstances a person will qualify as seeing that ---. (On the whole, it is often healthy to shift one’s attention to the third person in discussing issues like those of perception, memory, as well as many other psychological phenomena.)
(4) In the ordinary sort of case, seeing that --- requires no mediator: it does not usually consist in, nor is it based on, any specific state or entity. To decide that someone sees that the traffic light is red all we need to do, roughly speaking, is to establish that, with the help of her eyes, and the light being red, she judges that it is red. (In many cases of seeing-that, it isn’t even clear what the object might be: you can see that it’s windy from the way people in the street clutch their clothes, you can see from her eyes and her expression that she has been crying, you can see that the suitcase is heavy from the way that man carries it, you can see that there’s been bread in this basket from the crumbs left in it.)
(5) A great many (maybe most) tangles in the philosophy of perception are due to the idea of a mediating state or entity, and to the effort to decide what that state or entity must be like.

Let’s get back, then, to the Gopnik example. I would suggest that she makes the following assumption: the way human beings appear to people with autism has to differ from how they appear to others, since it is through their appearance that it is given to those others, and is not given to people with autism, that what they all see are human beings. (Here I grant the notion that we do not normally infer that people we see are people. I am not, on the other hand, committing myself one way or the other on whether her assumption that people with autism are “mindblind” is well-grounded or how it is to be understood.) 
My criticism of the passage by Alison Gopnik, then, has two parts. On the one hand, I would question the idea that there is a context-independent way of articulating how things appear to a person, and on the other hand, I would suggest that there is no reason for accepting the way Gopnik chooses to articulate how things appear to an autistic person, in whatever context.


  1. Someone might argue – I am not really sure I am one of them – that “seeing-direct object”-constructions deserves the primary attention, if we want to analyze the perceptual act of seeing, because ”seeing-that”-constructions doesn’t really denote the *perceptual* act of seeing, but rather the *reflective* act of realizing something. It could be argued that it is exactly because “seeing-that”-constructions do not really denote mere perceptions, that seeing-that locutions generally give a more direct expression to how things bear on my activities.

    My seeing the red traffic light explains my stopping the car only provided I *realized* there was a red traffic light. An account which argues that “seeing-that” is really just synonymous with “realizing” also has the benefit of being consistent with such uses as “I can see that non-Euclidian geometry is really useful” and “After going to Mekka, you will see that God is great” – instances which cannot paraphrased by “seeing-direct object”-constructions, exactly because they are not instances of the perceptual act of seeing, but of the reflective act of realizing something.

    1. I’m glad you bring this up. To those who argue the way you suggest I wish to make two comments:

      (1) It is not immediately clear what kind of project “analyzing the perceptual act of seeing” is. This was part of what I wanted to suggest in my blog.
      One thing the phrase might be taken to refer to is what happens in the sensory organs. Neurologists and perception psychologists might study the conditions for making things out, the sources of visual illusions, etc. Their findings might be of great interest, although they would of course be empirical, by which I mean that they would not directly be addressing the philosophical issue of how perception vocabulary is used; what kinds of point people make in using the word “see”. Furthermore, I would contend that the empirical study of vision is dependent on establishing facts about seeing-that: it is ultimately by studying people’s ability to get by in the world by the help of their sense organs (their ability to see / hear / sense that ---) that we are able to detect the conditions of perception and the sources of misperception.
      On the other hand, someone might suggest that “the perceptual act of seeing” refers to the *experience* of seeing: to *what it is like* to see that something is the case. But again, I am not sure what this expression means outside of particular contexts. If someone asked me, What is it like to see that the light is red?, I might think he wanted me to tell him what a traffic like looks like, or what you have to look out for when approaching a traffic light, or maybe he is wondering what it feels like when the light suddenly turns red and you realize you won’t have time to react. Or the question might be addressed to someone who is losing his eye-sight, or someone who is red-green colour blind, etc. What I wish to claim here is, first, that there is no *one* thing that the experience of seeing is like, and, second, that the whole idea of its being “like” something is only intelligible in a certain context of speaking.

      (2) You suggest that seeing-that locutions don’t denote the *perceptual* act of seeing but the *reflective* act of realizing something. Why is that? If Jack sees that the light is turning to red, then to be sure that would normally mean that he realizes that it is turning to red, but it also means that he comes to realize that (more or less directly) by using his eyes (rather than, say, by Uncle Herbert shouting out, “Hit the brake, you little moron, didn’t you notice that the light turned red?!” – in which case he might also have realized it turned red, but not by seeing it, nor yet by hearing it, although he came to realize it through something he heard).
      Coming to realize, by seeing it, that something is the case would not in most cases be called reflective. Why do you suggest that it would? Is it because you think of cases like, “I can see that non-Euclidian geometry is really useful” and “After going to Mekka, you will see that God is great”. But those are clearly different from cases of perceptual seeing that ---. For one thing, Jack might say, “The reason I didn’t notice the traffic light was that I had left my glasses at home”, but it would be rather peculiar for someone to say, “At first it was too dark but after the lights had been turned on / After I got new glasses / I could see that non-Euclidian geometry is really useful” (a situation might be imagined in which someone might say that, but then it would be a case of perception). Besides, this use of “see” in the sense of “understanding” or “having an insight” seems a peculiarity of English. In Swedish, for instance, the word might at most be used as a metaphor in this sense.

  2. Some influential theorists of animal cognition view all nonhuman animals as mindblind. They would probably accept Gopnic-like descriptions of animals’ perception of each other.
    Animals use subtle bodily gestures, however. Spontaneously, I’d say that their impressive emotional expressions would seem wasted if they perceived each other as “furry skin bags.”
    Given the assumption that animal perception is mindblind, or “autistic,” it is amusing to note that an individual ape recently was suspected of suffering from autism:
    If all apes are autistic, it is remarkable that an individual ape can differ by exhibiting traits of autism.
    I think you are right that Gopnic-like descriptions of what it is like to be autistic above all reveal what it is like to philosophize about perception. Animals and autistic persons, unfortunately, are often caught in the philosophical crossfire.

    1. The supposition that animals are “mindblind” – in the sense of lacking spontaneous responses to the expressions or gestures of members of their own species, or even, in the case of many animals, to human beings – seems totally gratuitous. It might be interesting to turn the issue around (though this is off topic), and discuss the extent to which human beings may be “mindblind” to animals. Undoubtedly the ability to relate to cats, dogs, or I suppose bonobo, varies from individual or individual. One factor here is probably whether one has grown up with pets. ---- Speaking of this, I think it important to distinguish between the ability to relate to animals, to make sense of their behaviour, on the one hand, and the inclination to attribute human features and abilities to them on the other hand. Indeed, it seems likely that the tendency to anthropomorphize may at times prevent our seeing the animal for what it is.

  3. David Cockburn22 July 2012 at 09:28

    You take to be equivalent (I think) the idea that ‘direct-object seeing’ (in the sense you articulate) is basic and the idea that there must be some entity that mediates my seeing: a look of things. This seems to me a mistake, and one that comes out at a number of points in what you say. For example, you suggest that “focusing on the direct-object locution tempts us to concentrate on the first person case: we contemplate ‘what it is like’ for us to see things”. That seems to me wrong. For, as you note, we may employ this locution in contexts in which the person does not recognise what she is seeing, for example the red traffic light, and there is no temptation in this case to concentrate on the first person or to suppose that there is a ‘what it is like to see the red traffic light’. The crucial idea for the temptation that you are trying to understand is, I think, that of a necessary mediator rather than that of direct-object perception. (Thus, your interesting observation that in many cases it is not clear what the ‘direct-object of perception’ might be does not seem to me to threaten the mediator idea very directly. Of course, the ‘mediator’ conception is often formulated, by philosophers at least, in terms of ‘direct-object’: a fact that presumably contributes to what I am suggesting is a conflation in what you say.)

    The consequence of this is that a good bit of what you say doesn’t seem to be focused quite as is needed. That said, points about there being no context-independent way of articulating the direct-object (in your sense) presumably carry over pretty directly to the articulation of appearances. (I wonder, I’m not sure, if the idea of ‘appearances’ as mediators might, if it is to do the work we expect of it, require the idea of a context-independent way of articulating it.)

    You suggest that Gopnik (and, by implication, all who are persuaded by such descriptions?) makes the assumption that ‘the way human beings appear to people with autism has to differ from how they appear to others, since it is through their appearance that it is given to those others, and is not given to people with autism, that what they all see are human beings’. But while such an assumption may often be at work in such projects it is not clear to me it has to be so. For, first, the notion of a visual appearance does not need to come with any picture of it as a ‘mediator’ (in some, or all, perception); there is a respectable, everyday, use of such locutions (often closely linked with the notion of ‘seeing as’ as Wittgenstein discusses this.) And, second, the urge to wonder what it is like for, say, someone with severe autism is, it seems to me, such a basic and significant form of concern for others I am reluctant to suppose that it rests on a philosophical confusion.

    Now, you don’t say that it must do so. Your conclusions are that there is no context-independent way of articulating how things appear to a person, and that there is no reason for accepting the way Gopnik chooses to articulate how things appear to an autistic person. The first of those points seems to me clearly correct. The second may be too, but if so, I would think it is not on the basis of any considerations in your discussion here, but for reasons of the kind you develop in your previous discussion of this. That is to say: there need be nothing confused in Gopnik’s project; if there is confusion it is in her execution of it.

    Continued below

    1. You may be right. Actually, you seem to agree with one of the two main points I make, I’m not sure what you think about the second, but in any case you disagree with the way I hook them up.
      The first point concerns the temptation to think about perception as mediated by something like percepts, appearances, or the like. I thought of a different way of expressing the idea: my being conscious, e.g., of there being a red traffic light in front of me consists in or is dependent on there being in my consciousness something that represents a red traffic light in virtue of its being red and having the shape of a traffic light. This is the view we are tempted to adopt, and which I wished to criticize.
      (This idea, by the way, seems to me to be an instance of a general tendency within contemporary Western philosophy to think of representation as a relation which obtains between the representation and what is represented by virtue of the relation between the *properties* of the two terms. Thus, it is inherent in my red, traffic-light shaped image that it is an image of a red traffic light. Yet representation is more usefully be thought of in terms of something being assigned *the role* of representing this or that. I should get back to this.)
      The second point is the suggestion that see-that locutions are a more fruitful starting point for discussions about perception than direct-object locutions, the reason being that the use of direct-object locutions is more varied and harder to survey. Also, it is easier to imagine the sort of narrative into which see-that locutions fit, whereas with direct-object locutions the context to be imagined is much more open; hence my suggestion that in cnnn with them it is more tempting to sink into the contemplation of “what it is like to see a red rose”, etc. But I’m not sure whether I am right about this or not.
      On the other hand, as you point out, direct-object locutions are no different from see-that locutions in that they do not enjoy first person authority any more than the other kind. My suggestion was simply that we tend to surround them with a different (hazier) atmosphere.
      I’ll get back to your last point in a separate response.

  4. David Cockburn22 July 2012 at 09:29

    I am inclined now to regard your earlier comments on James and Gopnik as observations of risks in our reading of the descriptions they offer. Any characterisation of how things appear to another person can, in principle, be taken in an inappropriate way; and, with that, may require particular sensitivity on the part of the one offering the description to how it is likely to be taken by the particular person to whom she is offering it. (I am thinking such descriptions may be more prone to such risks than are other forms of description – perhaps because of unclarities about ‘how much of myself’ I need to leave behind when taking on the description. But I’m not sure about that.) Where we are dealing with the experience of people at a considerable distance from ourselves (and even, perhaps, bats) the dangers are particularly acute. But that is not to say that the attempt need be misconceived. (With that, your specific concerns about Gopnik’s articulations might be read as pointers to ways in which they could be improved.)