The Content of a Seeing-As Experience
Undoubtedly, an ambiguous figure prompts a phenomenal difference in one’s experience of it insofar as the figure undergoes a Gestalt switch, i.e., it is seen now one way, under a certain aspect, now another way, under another aspect. For sure, moreover, this phenomenal difference is matched by a different description of such experiences: we say that the figure is now seen as a certain thing, now seen as another thing. Now the question is, does this phenomenal difference really correspond to an intentional difference, in the sense that there really is a difference in intentional content for the different seeing-as experiences involved in alternatively grasping one and the same ambiguous figure, in a nutshell: are such experiences different experiences of?
So-called anti-intentionalists negatively answer this question: at least in some cases, no intentional difference match the phenomenological difference at stake (cf. Peacocke ; Macpherson ; Nickel ). On the other hand, intentionalists give a positive answer to that question. Weak intentionalists limit themselves to say that there just is a correspondence, a mere co-variation, between the phenomenology and the intentionality of a seeing-as experience: no difference in the phenomenology without a difference in the intentionality of such an experience. Strong intentionalists rather say that such a correspondence is an identity: the difference in phenomenal character between two seeing-as experiences of the same figure is utterly captured by their having a different intentional content. Yet independently of the distinction between weak and strong intentionalism, the intentionalist approach has been modulated in two further ways. According to conceptual intentionalists, the intentional content of a seeing-as experience is conceptual: different concepts constitute the different intentional contents mobilized by different seeing-as experiences of an ambiguous figure. Non-conceptual intentionalists rather think that the content of a seeing-as experience whose phenomenal character either supervenes on or is identical with it is non-conceptual, it is not made of concepts (cf. Jagnow ; Orlandi ; Peacocke ; Raftopoulos [2009; 2011]; Tye ).
Yet intentionalists may develop a further theoretical possibility that combines the above two ones. Some seeing-as experiences have a phenomenal character that is merely matched by a non-conceptual intentional content, while some other seeing-as experiences have a phenomenal character that is also matched by a conceptual intentional content. In what follows, I will positively explore this possibility, by attempting to show that it is grounded in the fact that, unlike the former seeing-as experiences, the latter seeing-as experiences are pictorial experiences. In other terms, the experiences that concern an ambiguous figure in the latter case are of the same kind as the experiences we entertain when we perceive pictures, both simple pictorial representations like ordinary snapshots and childish sketches and complex pictorial representations like drawing and paintings; with respect to the involved figure, we simply redouble the experience that in pictorial cases normally, but not always, is just a single one. More precisely, in a normal pictorial experience, in seeing the material object she faces, the perceiver also has an experience of another something that is not there. Whereas in such seeing-as experiences, in again seeing the material object she faces, the perceiver also has different experiences of some different somethings that are not there. In this respect, when we see an ambiguous figure such as the famous Jastrow’s duck-rabbit figure, we are in the very same predicament as when we see an ambiguous painting, like e.g. an Arcimboldo’s painting, in which we reduplicate the kind of experience we entertain when we enjoy Leonardo’s Mona Lisa or Velazques’ Las Meninas; that is, we have two pictorial experiences of one and the same figure. Thus, we see the Jastrow figure either as a duck or as a rabbit pretty much in the same sense as we see an Arcimboldo’s painting either as a human being or a set of fruits and vegetables. Let me therefore consider the second kind of seeing-as experiences as experiences of ambiguous pictures.
The architecture of this paper will be the following. Firstly (section 2), I will hold that there are some seeing-as experiences that merely have a non-conceptual intentional content; namely, those experiences involving merely two-dimensional ambiguous figures: experiences of organizational seeing-as. Secondly (section 3), I will focus on other seeing-as experiences of two-dimensional ambiguous figures that however also prompt an experiential grasping of the third dimension. I will treat those experiences as pictorial experiences of ambiguous pictorial representations. As to pictorial experiences in general, I will share Wollheim’s (19802) idea that such experiences have a complex phenomenology of a sui generis twofold experience that is matched by a two-tiered intentional content. Yet I will further analyze what the folds of such an experience are in terms of a fold of organizational seeing-as and a fold of knowingly illusory seeing-as, by also ascribing them a non-conceptual and a conceptual content respectively. This will enable me to treat those other overall seeing-as experiences as pictorial experiences of ambiguous pictorial representations that are endowed with the above different two-tiered intentional contents. Furthermore, I will briefly investigate the consequences of such a position as to the debate on the so-called cognitive penetrability of perception (section 4). After having considered an objection to this treatment of seeing-as experiences of ambiguous pictures (section 5), I will finally try to show (section 6) how this overall treatment of seeing-as experiences has been anticipated in the later Ludwig Wittgenstein’s reflections on this issue.
Before starting, a caveat. By defending the intentionalist idea that seeing-as experiences have an intentional content, either conceptual or non-conceptual, matching their phenomenology, I do not want to defend intentionalism in general. So in the end I am neither a strong nor a weak intentionalist. I indeed believe that as a general thesis on qualitative states intentionalism fails, for there are qualitative states – notably, intero- or proprio-ceptive sensations and moods – whose qualitative properties are not matched at all by intentional ones (cf. Voltolini ).
2. Organizational Seeing-As Experiences
Let me start by considering different seeing-as experiences of an ambiguous merely two-dimensional figure, that is, a figure that merely involves different two-dimensional perceptual interpretations. The prototype of such experiences is the one concerning the so-called Mach figure, namely the figure that can be seen either as a diamond or as a tilted square.
To my mind, there is no doubt that the two experiences here involved have a different phenomenology. (Some admit that, at least in some cases, this is questionable. This sounds rather incredible to me – I clearly entertain a phenomenal switch in grasping the different aspects of the figure. Yet even if this were really the case, it would merely involve a problem with the example – just change the example and consider a case in which one and the same array of dots can be seen either conforming to a vertical or a to a horizontal organization).
Now, as many have maintained, this different phenomenology is matched by a difference in their intentional content. Criticisms to this claim (cf. Peacocke ; Macpherson ; Nickel ) merely work as criticisms to improper formulations of what such a content really amounts to. To my mind, there indeed is a proposal as to what this intentional content consists in that resists all such criticisms: such experiences mobilize in their content different grouping properties of the involved figure. These properties are the different ways for the figure’s elements of being arranged according to different orientations. Let me therefore say that the two different seeing-as experiences that concern a merely two-dimensional figure are different experiences of organizational seeing-as. In the Mach figure, one can group its array in a certain way, along a certain orientational axis – the one actually following the angles’ bisection – so that the organizational seeing of the figure as a diamond arises. Yet one can group its array also in another way, along another orientational axis – the one actually following the sides’ bisection – so that the organizational seeing of the figure as a square arises.
Anti-intentionalists would immediately wonder which kind of properties these grouping properties are (cf. Nickel : 286). Well, the different ways of grouping one and the same perceived figure are different objective properties, i.e., different properties of such a figure. Yet they are orientation-dependent properties. Change the external frame of reference that allows a certain orientation of the figure’s elements, and you will change the way of grouping the figure. This dependence makes it the case that, although they are compatible properties – the figure possesses all such ways for its elements to be arranged – no experience mobilizes both properties at one and the same time. To be sure, one might suppose that orientation-dependence makes grouping properties subjective properties, i.e., properties of the experience rather than of what is experienced. Yet orientation-dependence occurs in a geometrical, not in an egological space, characterized by subjective perspectives to be described in a language using perspectival locutions (“on the left of / on the right of”, “on the top of / on the bottom of” …). So, even though they are dependence-involving properties as much as subjective properties are, grouping properties are not subjective properties.
In order to grasp this difference, suppose for instance to draw an ambiguous figure such as the distorted square – kite figure on a transparent vehicle such as a window pane and imagine to see it from both sides, i.e., both from a front and from a back side. Now, if the different ways of seeing the figure, the “kitish” and the “squarish” ones, occurred basically by grouping its dots either in a upleft-to-downright sense or in a downleft-to-upright sense when one frontally sees the figure, one would have to say that when one saw the figure from the other side, further different ways would emerge by grouping the figure’s dots in a downright-to-upleft sense and in a downright-to-upleft sense when one saw the figure from the other side. Yet in point of fact from both sides there simply are two same ways of grouping the figure, directionally arranged along a cardinal frame of reference (say, an east-to-west way rather than a west-to-east way) rather than an egological frame of reference. The different perspectival characterizations simply are different approximate descriptions of one and the same cardinal way of grouping the figure. So, the different perspectives notwithstanding, there are just two ways of grouping the figure prompting exactly the very same different seeing-as experiences. Likewise, although they depend on different cardinal orientations, the different grouping properties of the Mach figure do not depend on perspectival graspings of the figure.
In this respect, it is correct to remark that as to the Mach figure, what prompts the phenomenal switch from seeing the figure as a diamond to seeing it as a square it is not passing to see the figure as tilted, i.e., as perspectively oriented in a certain way with respect to the perceiver.
If grouping properties must be objective, however, one might wonder whether they have to depend on external frames of reference, as I have just maintained, or rather on object-centered frames, coinciding with some intrinsic axes or other of the figure (Jagnow : 336-338). Yet I believe that object-reference frames must yield way to external frames.
On behalf of the objectualist, one might remark that, if instead of taking the Mach figure, which we see either as a diamond or as a tilted square, we took a normal square, which we can however well see not only as a square but also as a tilted diamond, the very same cardinal orientation (say, a north-to-south orientation) would prompt different seeing-as experiences of the two figures respectively (say, a “diamondish” experience in the case of the Mach figure and a “squarish” experience in the case of a normal square). So, according to this remark, external frames of reference cannot account for the difference in phenomenal character among seeing-as experiences (ibid.).
Truly enough, in such a predicament one and the same cardinal order would corrispond, say, both to the “diamondish” experience of the Mach figure and to the “squarish” experience of the normal square. Yet one must recall that grouping operations always occur after that one has grasped more basic visual features of a figure. These are the features of a figure that remain constant in a seeing-as switch that concerns it: namely, its colours and shapes. In our case, there is a difference in such more basic visual features of the two figures involved that makes it the case that there is a phenomenological difference between such figures which is preliminary to their also being grouped in certain ways. This phenomenological difference indeed depends on the fact that before their being somehow grouped, the Mach figure and the normal square are differently shaped, so that they differently parse the respective region of space. In other terms, what makes these figures phenomenologically different are their respective features that remain perceptually constant in the respective phenomenal switches that concern such figures, namely their different shapes. Because of this prior phenomenal difference between the two figures, we may still have that what accounts for the seeing-as differences that affect the Mach figure are a certain cardinal orientation of its shapes, prompting a “diamondish” seeing-as experience of it, and another cardinal orientation of its shapes, prompting a “squarish” experience of it, whereas what accounts for the seeing-as differences that affect the second figure, the normal square, are the former cardinal orientation of its shapes, yet prompting this time a “squarish” seeing-as experience of it, and the latter cardinal orientation of its shapes, yet prompting this time a “diamondish” seeing-as experience of it. Or so I claim.
Moreover, clearly enough grasping grouping properties is a matter of attention, as many say. As to the Mach figure, focusing on its vertices favors grasping the “diamond”- aspect, while focusing on its sides favors grasping the “square”- aspect. Yet it would be wrong to say that such a focusing is a form of spatial attention spotting light on some points of the figure (cf. Raftopoulos ; : 498-507). Even though one fixes a point in that figure, attentional focusing amounts to a certain overall grasping of the figure. Consider an ambiguous figure that we still take as a merely two-dimensional figure in which now a certain contour, now another contour appears visually relevant, as in the “tiles”- case below in which either the “cross-organized” 2, 4, 6, 8 tiles or the “X-organized” 1, 3, 5, 7, 9 tiles visually emerges. Since such contours overlap, the switch between them cannot refer to the fact that a certain area of the figure rather than another one is spotted. Rather, focusing on certain points of the figure immediately favors a certain holistic grasping of it, while focusing on other points of the figure favors another holistic grasping of it. Likewise, once you fix a vertex of the Mach figure, the whole figure itself will immediately appear in its “diamond”- aspect. Whereas once you fix one of its sides, the whole figure itself will immediately appear in its “square”- aspect.
Let us stick to the Mach figure. As to it, it is correct to say that one does not need to master the concepts of being a square and being a diamond, or any other concept at all, to perform the relevant phenomenal switch. There is no principled reason as to why no such switch should occur for a perceiver if she did not possess such concepts. This is particularly evident in the aforementioned “dots”-figure; the different vertical and horizontal perceptual arrangements corresponding to the phenomenal switch occurring there can be grasped and even described without referring to any concept at all. Hence, one may well say that the intentional content of the two organizational seeing-as experiences of one and the same ambiguous figure is utterly non-conceptual.
This is definitely not a new idea: all the intentionalists hitherto quoted have precisely maintained this position. What is possibly new is simply that the relevant element in this non-conceptual intentional content is made by grouping properties, properties dependent on a cardinally-framed orientation. Once you grasp different such properties, a different organizational seeing-as experience of one and the same figure arises.
3. Overall Seeing-As Experiences
Yet experiences that concern ambiguous figures like the Mach figure are not the only seeing-as experiences of ambiguous figures that there are. For such experiences involve only two-dimensional groupings: the third dimension, depth, is not involved. Depth comes in whenever in the relevant experiential groupings of an ambiguous figure a figure-ground segmentation emerges, which is not the case as regards figures like the Mach figure. In these new cases, while contemplating an ambiguous figure which in itself is just a two-dimensional item, its perceiver differently groups its elements not only along height and length, but also in such a way that some elements are more distant than others. This prototypically happens in the case of the “Rubin’s vase”-figure, in which, depending on the ways in which its elements are also three-dimensionally grouped, one sees the figure either as two dark faces on a light background or as a light vase on a dark background. Yet it also happens, though perhaps less evidently, in the case of the famous “duck-rabbit”-Jastrow’s figure, in which one sees the figure either as a duck (on a background) or as a rabbit (on a background).
Once more, there is no doubt that a phenomenal switch also occurs in these cases; it is one thing to see the Jastrow figure as a duck, quite another thing to see it as a rabbit. One should not be led astray by the fact that, before and after the switch, something remains constant in the experience of the figure, namely its colors and shapes. For the switch is phenomenal insofar as the groupings of the figure’s elements change. In the Jastrow figure, one basically passes from a certain “cardinal” grouping that implicitly involves depth as well (hence approximately, but basically improperly, described as a perspectival “left-to-right” grouping) and enables one to grasp the “rabbit”-aspect of the figure to another “cardinal” grouping that again implicitly involves depth (hence approximately, but basically improperly, described as a perspectival “right-to-left” grouping) and enables one to grasp the figure’s “duck”-aspect.
On behalf of the intentionalist, one might think that these three-dimensionally grasped ambiguous figures are just other cases of the same predicament as the one displayed by figures like the Mach one: the different seeing-as experiences trace back to two different non-conceptual intentional contents, captured in terms of (admittedly more complex) different grouping properties (see e.g. Jagnow ; Orlandi ; Raftopoulos [2009; 2011]). Yet this way of putting things neglects the fact that the experiences involved here are definitely more complex than those involved in the previous cases.
This clearly transpires if one notices the fact that, unlike figures like the Mach one, figures like the Jastrow figure are ambiguous pictures, i.e., two-dimensional figures that also involve different three-dimensional interpretations insofar as they pictorially represent different items – a duck and a rabbit, in this case. Accordingly, experiences of these figures are pictorial experiences, more precisely mutiplied pictorial experiences insofar as they concern ambiguous pictures: they replicate in terms of different pictorial experiences what in the case of normal pictures is just a single pictorial experience. The Jastrow figure is both a picture of a duck and a picture of a rabbit, so that the overall experience of it is both a seeing of it as a picture of a duck and a seeing of it as a picture of a rabbit. Now, although many people taking part in this debate have noticed this fact (cf. e.g. Peacocke : 17; Macpherson : 97-98), its consequences have rather gone unnoticed. So let me expand on it.
As anyone knows, pictorial experiences are complex experiences, for in ne such experience at one and the same time a perceiver experiences the typically material object she faces – the picture’s vehicle – and what such an object depicts, which unlike the vehicle is not there – the picture’s subject. Richard Wollheim has tried to capture this complexity by holding that pictorial experiences are sui generis experiences, experiences of seeing a subject in a vehicle, having a specific phenomenology that is expressed in the experiences’ twofoldness. For him, a seeing-in experience indeed has two distinct yet unseparable folds, the configurational fold, in which one grasps the picture’s vehicle, and the recognitional fold, in which one grasps the picture’s subject (cf. Wollheim [19802; 1987; 1998]).
Many people have found themselves dissatisfied with this proposal as it stands, for it is not clear what those folds really amount to and how they can mix together if they are about such different entities and yet the seeing-in experience has a unitary phenomenal character (cf. respectively Lopes : 50; Hopkins : 167-171). Now, the cases of ambiguous pictures may precisely help one to give an answer to both perplexities. For the different seeing-as experiences of an ambiguous picture are precisely two different seeing-in experiences, to be treated as seeing-as experiences more complex than the previously considered ones.
As Wollheim himself recognizes, there would be no seeing-in experience of something if there were no seeing of that very something as a picture (cf. Wollheim : 226). Let me therefore consider a seeing-in experience the same as the overall experience of something as a picture. Since in the case of ambiguous pictures we have two seeing-in experiences with respect to one and the same figure, seeing a figure as a picture here splits itself into two different such seeing-as experiences, to be further unpacked as we will now see.
On the one hand, a certain overall seeing-as experience of an ambiguous picture is similar to a seeing-as experience of an ambiguous merely two-dimensional figure: it is the grasping of a certain grouping of the pictorial vehicle’s elements that also involves the third dimension. So, such an overall seeing-as experience is, at least in part, an experience of organizational seeing-as that has a certain non-conceptual intentional content, admittedly more complex that the one of a seeing-as experience of an ambiguous figure for it also involves arranging the vehicle’s elements along depth. A different overall seeing-as experience of the same picture will thus mobilize a different organizational seeing-as experience, hence a grasping of a different grouping of the vehicle’s elements still involving depth, hence a different non-intentional content. Consider again the Jastrow figure. One may see the figure in a certain way, by grasping its elements in a “rabbitish” way, yet one may also see the figure in another way, by grasping its elements in a “duckish” way. These different ways are two grouping properties of the figure mobilized in the respective non-conceptual contents of such experiences.
Yet on the other hand, precisely the fact that such graspings involve the third dimension has an import that phenomenal switches concerning merely two-dimensional figures lack. For such grasping display in the overall seeing-as experiences another experiential level, the one in which one sees the ambiguous picture as a certain thing rather than as another one. At this level, within the overall seeing-as experiences of an ambiguous picture, over and above the aforementioned different organizational seeing-as experiences, one has further different seeing-as experiences whose phenomenal difference is also matched by an intentional difference. In the first of these further experiences, one sees the figure as a certain something (as a rabbit, in our example), while in the second of these further experiences, one sees the figure as another something (as a duck, in our example). Now, the different seeing-as experiences that are mobilized at this level are also different in kind from the different organizational seeing-as experiences that are mobilized at the previous level. These further experiences do not indeed mobilize grouping properties. Rather, they are experiences in which utterly knowingly, the picture’s vehicle is taken either as a certain thing or as another one. In the Jastrow figure, for example, one well knows that the picture is neither a rabbit nor a duck. Yet, one is perceptually forced to take it either as a rabbit or as a duck. In this respect, these further seeing-as experiences are like optical illusions recognized as such, as when one goes on seeing an oar as crooked even if one well knows that the oar is not so crooked. Let me find a new label for these further seeing-as experiences: experiences of knowingly illusory seeing-as.
I can now say that it is somehow improper in the pictorial case to take these two seeing-as experiences of a different kind, the organizational seeing-as experiences and the knowingly illusory seeing-as experiences, as experiences per se, insofar as they contribute to constitute one and the same experience. This further experience is what I have labeled the overall seeing-as experience of something as a picture that Wollheim’s seeing-in experience actually amounts to and that comes doubled in experiencing an ambiguous picture. In point of fact, an organizational seeing-as experience and a knowingly illusory seeing-as experience precisely are nothing but the different aspects of one and the same experience, call it indifferently an overall seeing-as experience or a seeing-in experience, encompassing them. They indeed are what Wollheim labels, as we have seen, the configurational and the recognitional fold of a twofold seeing-in experience (see again Voltolini [2012a; 2012b]).
Put alternatively, we have found what those folds amount to: the first one is an experience of organizational seeing-as in which one sees an item, typically a picture’s vehicle, as arranged in a certain way, while the second one is an experience of knowingly illusory seeing-as in which one sees that very item as a certain something. As a result, when an ambiguous picture is at stake, we have two different seeing-in experiences, or overall seeing-as experiences, in which two different folds, the configurational fold or the aspect of organizational seeing-as and the recognitional fold or the aspect of knowingly illusory seeing-as, are respectively involved. When we see a certain something (say, a rabbit) in an ambiguous picture, or we see that picture as a picture of that very something, we both see the pictorial vehicle’s elements as grouped in a certain (“rabbitish”) way and we see that vehicle as that very something, while when we see another something (say, a duck) in that picture, or we see that picture as a picture of this other something, we both see the pictorial vehicle’s elements as grouped in another (“duckish”) way and we see that vehicle as this other something.
Moreover, it is now clear why the two folds of a seeing-in experience are inseparable, as Wollheim claims. For the first fold has an intentional content on which the content of the second fold is grounded. Take again the Jastrow figure. It is because one sees the picture’s vehicle in a “rabbitish” way, by properly grouping its elements three-dimensionally, that one also sees it as a rabbit; likewise, it is because one sees the picture’s vehicle in a “duckish” way, by properly re-grouping its elements three-dimensionally, that one also sees it as a duck. So, taken as an overall seeing-as experience of something as a picture, a seeing-in experience indeed is phenomenologically unitary. For its overall phenomenal content supervenes on its overall intentional content, which is made by the two distinct yet intertwined contents of its folds, the organizational seeing-as experience of the picture’s vehicle and the knowingly illusory seeing-as experience of that vehicle. Hence whenever an ambiguous picture is at stake, one simply has two seeing-in experiences, or overall seeing-as experiences, that respectively supervene on the distinct overall intentional contents of such experiences, which again are respectively made by the two respective distinct yet intertwined contents of those experiences’ folds.
This account of ambiguous pictures thus shows that intentionalism is saved not only with respect to merely two-dimensional figures, but also with respect to two-dimensional pictures leading to three-dimensional experiences. As we have indeed just seen, the two overall seeing-as experiences of one and the same ambiguous picture differ insofar as they have different intentional contents that come out of the interaction between the intentional contents of their respective configurational folds (their respective organizational seeing-as experiences) and the intentional contents of their respective recognitional folds (their respective knowingly illusory seeing-as experiences).
But which kind of content is the one that figures in the recognitional fold partially constituting an overall pictorial experience? Insofar as that fold is a knowingly illusory seeing-as, it involves the so-called report awareness, the kind of awareness needed when one has to recognize that one and the same object is given twice to her. When such an awareness is at play, the content of an experience affected by it can only be conceptual. So, the different intentional contents of two overall seeing-as experiences of one and the same ambiguous picture are partially conceptual. For, while the contents of their configurational folds are still non-conceptual, the contents of their recognitional folds are instead conceptual.
4. A Short Detour on Cognitive Penetrability
If what I have just said is correct, it has some bearings as to the issue of whether experiences are cognitively penetrable. Cognitive penetrability can be meant in two senses, a weaker and a stronger one. On the one hand, weak cognitive penetrability is the thesis according to which either the phenomenal character or the intentional content of an experience are permeable by states of their subjects’ cognitive systems, hence by the concepts that constitute the intentional content of such states (cf. Macpherson ). Seeing-as experiences of the former kind, the mere organizational seeing-as experiences involving merely two-dimensional figures, are cognitively penetrable in this weak sense. As I have said, the intentional content of such experiences is non-conceptual. In such cases, one can indeed entertain a phenomenal switch even if one has no mastery of the relevant concepts. Nevertheless, concepts may well affect the phenomenal characters of such experiences, insofar as they may well activate attention in performing the grouping job that prompts the different experiences involved. On the other hand, strong cognitive penetrability is the thesis according to which the intentional content of an experience is permeable by states of their subjects’ cognitive systems, hence by the concepts that constitute the intentional content of such states. Seeing-as experiences of the second kind, overall seeing-as experiences involving ambiguous pictures, are cognitively penetrable in this strong sense. For, as we have seen, they involve a layer of content which is conceptual, the one corresponding to their recognitional fold.
People who deny that perception of ordinary objects is conceptual standardly distinguish between early vision, which enables a perceiver to individuate an object in her surroundings, and late vision, which enables the perceiver to reidentify the object as one and the same entity at different times. Insofar as this is the case, they say, early vision is non-conceptual while late vision is conceptual. However, they further claim, only the former kind of vision is genuinely perceptual, is a way of perceptually grasping the objects out there; the latter is just a matter of interpretation of the perceptual data. This is why, they comment, perception is cognitively impenetrable (cf. e.g. Pylyshyn ; Raftopoulos ).
Yet if a pictorial experience works in terms of an overall seeing-as experience as I have here maintained, this way of putting things cannot hold of it. For in a pictorial experience so conceived, early vision is completely exhausted by the organizational seeing-as fold of that experience, which admittedly is non-conceptual in its intentional content. If early vision can be unpacked in terms of a primary sketch, in which basically the perceived objects’ contours are grasped, and of a 2½D sketch, in which depth hence distance relations among the perceived objects are grasped, then such an unpacking is precisely what takes place in the organizational seeing-as fold of a pictorial experience conceived as an overall seeing-as experience. In such a fold all such features of the picture’s vehicle, which as such is just an object among many other objects, are grasped qua its grouping properties. Yet if that pictorial experience so conceived is really a perception although a sui generis one as Wollheim claims by treating it as a twofold seeing-in experience, its perceptual character is not absorbed by its organizational seeing-as fold; its knowingly illusory seeing-as fold is genuinely perceptual as well. Yet, as we have seen, this latter fold is conceptual in its intentional content. By being conceptual, this fold is a case of late vision. Yet since it contributes to the perceptual character of the pictorial experience so conceived, at least in such a case late vision utterly contributes to perception; hence, perception in the case of a pictorial experience so conceived is cognitively penetrable in the strong sense, that is, concepts figure in the overall intentional content of such a state.
5. An Objection and a Reply
One might still wonder whether an overall seeing-as experience made by two typologically distinct seeing-as folds, as to my mind a pictorial experience of seeing-in amounts to, has to have an intentional content that is at least partially conceptual. As regards ambiguous merely two-dimensional figures, I have admitted before that one who masters e.g. neither the concept of being a diamond nor the concept of being a square can entertain the relevant phenomenal switch concerning the ambiguous Mach figure. In fact, I have said, the intentional content of the relevant organizational seeing-as experiences there is non-conceptual. Yet, one might wonder, is it not the same in the case of ambiguous pictures: as to e.g. the Jastrow figure, cannot one who masters neither the concept of being a duck nor the concept of being a rabbit entertain the relevant phenomenal switch? And if this is the case, does not this show that the content of either seeing-as experience is utterly non-conceptual, it is at most described or interpreted as conceptual? (Cf. again Raftopoulos [2009; 2011] and Macpherson herself : 95).
First of all, one might retort that empirical evidence goes against this hypothesis: children less than four-years-old who do not possess the relevant concepts are enable to entertain the switches concerning an ambiguous figure (cf. Gopnik, Rosati ; see also Leopold, Logothetis ). Yet even if contrary to such an evidence it turned out that individuals without such a mastery are able to perform the relevant phenomenal switch, this would only prove that the conceptual content an overall seeing-as experience is less fine-grained than we expected: it has to be accounted for in terms of a certain conceptual disjunction rather than in terms of a single concept.
In this respect, it is interesting to note that as regards an ambiguous picture, just some concept prompts the relevant phenomenal switch. Consider once again the Jastrow figure and suppose you see it as a duck. If you saw it as a goose, you would still have the same experience. You have to see it e.g. as a rabbit to entertain the relevant phenomenal switch. On the other hand, suppose you see the figure precisely as a rabbit, by therefore having another seeing-as experience concerning it. If you saw it as a hare, you would still have the same experience. You have to see it e.g. as a duck to entertain the relevant phenomenal switch.
A non-conceptualist may say that this fact shows that a certain overall seeing-as experience has no conceptual intentional content for it is an experience that it is merely conceptually interpreted in a certain way. But this is not what this fact shows. Rather, it simply shows that there are limits to the interpretation: in one way of seeing the ambiguous picture, one can appropriately interpret the picture according to different concepts, yet all these concepts have to be compatible with a certain phenomenal character of the picture’s experience. Other concepts appear to be appropriate only once the experience has a different phenomenal character, it becomes an utterly different overall seeing-as experience. Now, this predicament can be accounted for by saying that in its knowingly illusory seeing-as fold, a certain overall seeing-as experience has a conceptual intentional content whose extension is broader than what one originally supposed; namely, an extension that coincides with that of the disjunction of the concepts that are mobilized in the appropriate interpretations.
So, we should not properly say that a certain overall seeing-as experience of e.g. the Jastrow figure phenomenally is the experience it is independently of whether it is interpreted either as the seeing of the figure as a duck or as the seeing of the figure as a goose. Rather, we should say that such a seeing-as experience phenomenally is the phenomenal experience it is insofar as in its knowingly illusory fold, it is the experience of seeing the figure as an anatid (i.e., either as a duck or as a goose). Likewise, we should not properly say that another overall seeing-as experience of the same figure phenomenally is the experience it is independently of whether it is interpreted either as the seeing of the figure as a rabbit or as the seeing of the figure as a hare. Rather, we should say that such a seeing-as experience phenomenally is the phenomenal experience it is insofar as in its knowingly illusory fold, it is the experience of seeing the figure as a leporid (i.e., either as a rabbit or as a hare).
One must here not be led astray by the fact that the intentional content of the picture – in one of its interpretations, if it is an ambiguous picture – has a narrower extension than the conceptual intentional content of the knowingly illusory fold of the relevant overall seeing-as experience that concerns such a picture. For such a content is what turns out of a negotiation that fixes what, among all things that one can see the picture as, is the content of that picture, what that picture is of – again, if it is an ambiguous picture, in one of its interpretations.
For instance, with respect to one interpretation of the following ambiguous picture, in the knowingly illusory fold of a certain overall seeing-as experience concerning it one can see it either as Lenin or as Gengis Khan or as many other moustached human males, insofar as one can generically see it as a moustached human male. Yet in this interpretation it is only a picture of Lenin, for this is one way the picture has been officially meant. While with respect to another interpretation of that picture, in the knowingly illusory fold of another overall seeing-as experience concerning it one can see it pretty much in the same way as if she looked at it from standing above it; that is, either as Che Guevara or as the Italian ex-comedian Beppe Grillo or as many other bearded human males, insofar as one can generically see it as a bearded human male. Yet in this interpretation it is only a picture of Che, for this is another way the picture has been officially meant (see Voltolini [2012a]: 183-186).
6. Wittgenstein as a Precursor of the Present Ideas
All in all, therefore, we have two different kinds of seeing-as experiences, a mere experience of organizational seeing-as that affects merely two-dimensional figures whose intentional content is non-conceptual and an overall seeing-as experience that affects pictures, i.e., two-dimensional figures endowed with a three-dimensional interpretation, i.e., an experience of seeing a figure as a picture of something. In its turn, this overall seeing-as experience is Wollheim’s analyzed seeing-in experience. It is indeed made of two intertwined yet typologically different experiential folds, a fold again of organizational seeing-as whose intentional content is again non-conceptual and another fold grounded on the previous one, a fold of knowingly illusory seeing-as, whose intentional content however is conceptual.
Now, although Wittgenstein is generally interpreted as the main defender of the conceptualist intentionalist position, his ideas on seeing-as experiences are more articulated. In point of fact, he somehow captured the above distinction between different types of seeing-as experiences when he drew an analogous distinction between optical and conceptual aspects. For Wittgenstein, first of all, optical aspects are characterized by the fact that the phenomenal switch they mobilize between different seeing-as experiences of one and the same figure occurs automatically; while as regards conceptual aspects, one such switch is subject to one’s will. Moreover, an optical aspect is mobilized precisely when one merely sees a two-dimensional figure in a certain way (rather than in another one), while a conceptual aspect is mobilized when, in virtue of seeing a two-dimensional figure, one sees it as a three-dimensional item (or as another one) (cf. Wittgenstein : I, §§ 970, 1017). It is only in the second case that the relevant seeing-as experience has a conceptual intentional content; in the first case, we can conjecture (although Wittgenstein does not put it in these terms) that its intentional content is merely non-conceptual. Finally, one and the same figure can be the object of both kinds of experiences. This is the case of the so-called “double cross”-figure.
One may well pass from seeing the above figure as a white cross on a black background to seeing that figure as a black cross on a white background. Yet one may also pass from merely seeing a two-dimensional black figure (call it a two-dimensional black cross if you like) flanked as it were by a white two-dimensional array of triangles to merely seeing a two-dimensional white figure (call it a two-dimensional white cross if you like) flanked as it were by a black two-dimensional array of triangles. In this “two-dimensional” stance to the figure, we have a switch between different organizational seeing-as experiences (whose intentional content is, we may add, non-conceptual). To be sure, Wittgenstein does not even say that in the previous “three-dimensional” stance to the figure, one has distinct overall seeing-as experiences made by two folds whose content is respectively non-conceptual and conceptual. Yet he clearly means that that only in such a stance concepts are involved in the intentional contents of the two relevant seeing-as experiences. As he puts it:
Those two aspects of the double cross (I shall call them A aspects) might be reported simply by pointing alternately to a free-standing white and a free-standing black cross.
Indeed, one could imagine this as a primitive reaction in a child, even before he could talk. […].
The A aspects are not essentially three-dimensional; a black cross on a white ground is not essentially a cross with a white surface in the background. One could teach someone the idea of the black cross on a ground of different colour without showing him anything other than crosses painted on sheets of paper. Here the “background” is simply the surrounding of the cross. (Wittgenstein : II xi, §§ 215, 218).
Block, N., 2010, Attention and Mental Paint, “Philosophical Issues”, 20, pp. 23-63.
Chalmers, D., 2004, The Representational Character of Experience, in B. Leiter (ed.), The Future of Philosophy, Oxford University Press, Oxford, pp. 153-81.
Chisholm, R., 1993, Act, Content and the Duck-Rabbit, in J.W. Canfield, S.G. Shanker (eds.), Wittgenstein’s Intentions, Garland, New York, pp. 94-95.
Churchland, P., 1989, Perceptual Plasticity and Theoretical Neutrality: A Reply to Jerry Fodor, in A Neurocomputational Perspective, The MIT Press, Cambridge MA.
Crane, T., 2001, Elements of Mind, Oxford University Press, Oxford.
Fodor, J., 1983, The Modularity of the Mind, The MIT Press, Cambridge MA.
Gombrich, E., 1960, Art and Illusion, Phaidon, London.
Gopnik, A., Rosati, A., 2001, Duck or Rabbit? Reversing Ambiguous Figures and Understanding Ambiguous Representations, “Developmental Science”, 4, pp. 175-183.
Hopkins, R., 2010, Inflected Pictorial Experience: its Treatment and Significance, in C. Abell, K. Bantinaki (eds.), Philosophical Perspectives on Depiction, Oxford U.P., Oxford, pp. 151-180.
Jagnow, R., 2011, Ambiguous Figures and the Spatial Contents of Perceptual Experience: A Defense f Representationalism, “Phenomenology and the Cognitive Sciences”, 10, pp. 325-346.
Leopold, D.A., Logothetis, N.K., 1999, Multistable Phenomena: Changing Views in Perception, “Trends in Cognitive Sciences”, 3, pp. 254-264.
Levinson, J., 1998, Wollheim on Pictorial Representation, “The Journal of Aesthetics and Art Criticism”, 56, pp. 227-233.
Lopes, D., 1996, Understanding Pictures, Oxford U.P., Oxford.
Macpherson, F., 2006, Ambiguous Figures and the Content of Experience, “Nous”, 40, pp. 82-117.
Macpherson, F., 2012, Cognitive Penetration of Colour Experience: Rethinking the Issue in Light of an Indirect Mechanism, “Philosophy and Phenomenological Research”, 84, pp. 24-62.
Marr, D., 1982, Vision, Freeman, San Francisco.
Nanay, B., 2010, Attention and Perceptual Content, “Analysis”, 70, pp. 263-20.
Nanay, B., 2011, Ambiguous Figures, Attention, and Perceptual Content: Reply to Jagnow, “Phenomenology and the Cognitive Sciences”, 10, pp. 557-561.
Newall, M., 2011, What is a Picture?, Basingstoke, Palgrave.
Nickel, B., 2007, Against Intentionalism, “Philosophical Studies”, 136, pp. 279-304.
Orlandi, N., 2011, Ambiguous Figures and Representationalism, “Phenomenology and the Cognitive Sciences”, 10, pp. 307-323.
Peacocke, C., 1983, Sense and Content, Clarendon Press, Oxford.
Peacocke, C., 1992, A Study of Concepts, The MIT Press, Cambridge MA.
Pylyshyn, Z., 2003, Seeing and Visualizing, The MIT Press, Cambridge MA.
Raftopoulos, A., 2009, Cognition and Perception, The MIT Press, Cambridge MA.
Raftopoulos, A., 2011, Ambiguous Figures and Representationalism, “Synthese”, 181, pp. 489-514.
Siewert, C., 2004, Is Experience Transparent?, “Philosophical Studies”, 97, pp. 15-41.
Tye, M., 1995, Ten Problems of Consciousness, The MIT Press, Cambridge MA.
Voltolini, A., 2011, How Picture Perception Defies Cognitive Impenetrability, in A. Reboul (ed.), Philosophical Papers Dedicated to Kevin Mulligan, Faculté des Lettres, Université de Genève, Geneva, http://www.philosophie.ch/kevin/festschrift/.
Voltolini, A., 2012a, How to Reconcile Seeing-As with Seeing-In (with Mimetic Purposes in Mind), in G. Currie, P. Kot’atko, M. Pokorny (eds.), Mimesis: Metaphysics, Cognition, Pragmatics, College Publications, London, pp. 383-407.
Voltolini, A., 2012b, Toward a Syncretistic Theory of Depiction, in C. Calabi (ed.), Perceptual Illusions. Philosophical and Psychological Essays, Palgrave, Basingstoke, pp. 164-191.
Voltolini, A., 2013, The Mark of the Mental, “Phenomenology and Mind”, 4, (forthcoming).
Von Ehrenfels, C., 1988, On Gestalt-Qualities, in B. Smith (ed.), Foundations of Gestalt Theory, Philosophia, Munich and Vienna, pp. 82-117.
Wittgenstein, L., 1980, Remarks on the Philosophy of Psychology, I-II, Blackwell, Oxford.
Wittgenstein, L., 20094, Philosophical Investigations, Blackwell, Oxford.
Wollheim, R., 19802, Seeing-as, Seeing-in, and Pictorial Representation, in Art and its Objects, Cambridge U.P., Cambridge, pp. 205-226.
Wollheim, R., 1987, Painting as an Art, Princeton U.P., Princeton.
Wollheim, R., 1998, On Pictorial Representation, “The Journal of Aesthetics and Art Criticism”, 56, pp. 217-226.
 For this difference, cf. Crane (2001): 83-85.
 Traditionally, conceptual intentionalism is ascribed to Wittgenstein (20094). In point of fact, many quotations by Wittgenstein as regards the perception of ambiguous figures go in this direction: e.g., «So we interpret it, and see it as we interpret it» (: § 116), «the lighting up of an aspect seems half visual experience, half thought» (: § 140). Yet as we will see later, Wittgenstein’s actual position is more articulated and closer to the position I will here defend. Conceptual intentionalism, moreover, is defended by those who espouse a theory-laden conception of perception or, which actually amounts to the same thing, believe in strong cognitive penetrability with respect to perception in general, the idea that perception has a conceptual content. For a recent defense of this idea, cf. Churchland (1989).
 Although Nanay (2010; 2011) does not explicitly label his position as non-conceptualist, it naturally falls under non-conceptualism.
 As both Jagnow (: 333) and Raftopoulos (: 507, 511) seem to allow.
 On behalf of the anti-intentionalist, one such further example is put forward by Peacocke (1983): 25-26.
 In the tradition following von Ehrenfels (1988), these properties are labelled Gestalt qualities.
 By such orientational axes I don’t mean symmetry axes. As Macpherson (: 103-105) rightly notices, both focusing on a certain symmetry axis rather than on another one prompts no phenomenal switch and phenomenal switches also occur as to non-symmetric figures like the distorted square – kite figure I will immediately talk about below.
 This point faces another criticism by Macpherson (: 103) against ascribing such seeing-as experiences intentional contents respectively made of compatible properties understood in terms of symmetry axes. For a similar point against Macpherson, see Raftopoulos (2011): 508. Moreover, since such grouping properties are compatible properties, they are co-instantiated by an ambiguous figure in all possible worlds in which they are instantiated (provided the figure retains its shape identity). Yet as Nickel (: 285) acknowledges, by itself this is not a problem for intentionalism insofar as intentional content can well be more fine-grained than propositions qua sets of possible worlds.
 For a similar difference between dependence-involving properties, see Newall (2011): 67. Nickel (: 298) criticizes an intentionalist recourse to subjective properties for it allows only a non-reductive intentionalist account, that is, an account that traces back phenomenal properties to an intentional content that still appeals to phenomenal properties. To be sure, some people (e.g. Chalmers ; Siewert ) have no qualms with non-reductive intentionalism. Nevertheless, since grouping properties are not subjective properties, this criticism does not apply to them.
 As Macpherson (: 91, 107-108) also says, insofar as subjective perspectives can always be imposed on any figure, it is not clear why some figures like the Mach figure prompt different seeing-as experiences, respectively that of an untilted diamond and that of a tilted square, while some other figures like that of a tilted A prompt no such difference. For other criticisms to the idea of interpreting the differences in non-conceptual intentional content of seeing-as experiences in terms of viewpoint-centered reference frames à la Raftopoulos (2009; 2011), see Jagnow (2011): 336-337.
 Incidentally, this shows that the reason as to why the different intentional contents of a Mach figure do not amount to different perspectival orientations is not that seeing an untilted figure and seeing a tilted figure, say an A and a tilted A, have no phenomenal difference for the perceiver. (Raftopoulos [(2011): 496] critically puts in this way Macpherson’s (2006) criticism reported in footnote 14.) In this latter case the two figures at stake are indeed differently experienced, for over and above their possibly being somehow grouped, they are differently shaped figures: an A’s shape occupies space in a certain two-dimensional way, while a tilted A’s shape occupies space in another two-dimensional way. So even if there are differences in intentional content for the experiences of such distinct figures, they are irrelevant to account for the difference in content of the different seeing-as experiences of one and the same ambiguous figure having one and the same shape, as in the Mach case.
 From Chisholm (1993), who interprets Wittgenstein in this respect, all the way down to Nanay (2010; 2011), Orlandi (: 317) and Raftopoulos (; : 498-507).
 Many attentional phenomena involving grouping work in this way, as pointed out by Block (: 31-37) in a paper actually defending a weak intentionalist position on the relationship between phenomenal and intentional properties.
 I here assume that the “tiles”- figure can be taken as a merely two-dimensional figure. In point of fact, since the phenomenal switch it prompts is naturally described as involving different figure-ground segmentations, it is more natural to rank this figure with the cases of ambiguous figures involving different three-dimensional interpretations, i.e. ambiguous pictures, I will discuss below.
 Cf. Jagnow (2011): 342, who polemizes against an interpretation à la Raftopoulos (2009; 2011) of Nanay’s (2010) position on this matter – cf. again (2011): 338-40 – that Nanay himself (2011) however rejects.
 As Peacocke says, while adding that in such a case, unlike the case of the so-called Necker cube, the phenomenological constancy is matched by an intentional constancy of the relevant experience. Cf. Peacocke (1983): 16-17. To my mind, the two cases are instead quite similar. In both cases there is a phenomenal difference – seeing the Jastrow figure as a duck, seeing it as a rabbit; seeing the Necker cube as a cube with a certain front face and another back face, seeing it as a cube with another front face and still another back face – which is matched by an intentional, partially conceptual, difference.
 Also Macpherson (: 97-98) is disposed to provide such an intentionalist non-conceptual reading for these cases.
 At least superficially; as Gombrich (: 249) maintained, all pictures are inherently ambiguous.
 On the compatibility between seeing-in and seeing-as experiences, cf. Voltolini (2012a; 2012b).
 The idea of treating the recognitional fold in terms of what I here call a knowingly illusory seeing-as experience originally comes from Levinson (1998).
 As Raftopoulos (: 148, 156) acknowledges.
 I have defended this thesis in Voltolini (2011).
 As Raftopoulos (: 272-274) claims on behalf of Marr (1980).
 As even non-conceptualists would admit if the knowingly illusory seeing-as were not a fold of an overall experience but it completely characterized an experience, as in the case of optical illusions. Optical illusions are actually the examples Fodor (1983) invokes in order to prove the modular character of ordinary perception.
 To be sure, the non-conceptual intentionalist may reply to this that even if possession of such concepts is required to entertain the phenomenal switch, it is not sufficient in order for the intentional content of the relevant seeing-as experiences to be conceptual. Although concepts prompt the switch, they do not figure as constituents of the contents of such experiences, which remain non-conceptual (in the terms of the previous Section, ambiguous pictures as well as ambiguous figures would thus support at most weak, but not strong, cognitive penetrability). Cf. Tye (1995): 140. I however think that this reply is ruled out by what I previously said: pictorial experiences are unlike experiences of merely two-dimensional figures in that they have recognitional folds, hence if they mobilize concepts these figure within the intentional contents of such folds. For other criticisms to Tye’s point of view, cf. Orlandi (2011): 312.
 Perhaps such an evidence is not so decisive, as Macpherson (: 95, note 35) claims.
 As again Raftopoulos (2009; 2011) claims among others.
 As Peacocke (: 25, note 27) originally underlined.
This work is licensed under a Creative Commons Attribution 4.0 International License (CC-BY- 4.0)
Tel. (0039) 055 2757700 Fax (0039) 055 2757712