Understanding Resemblance in Depiction:

What Can We Learn From Wittgenstein?

Elisa Caldarola

Depiction theories deal with two-dimensional surfaces, covered in marks and colours, whose configurations are meant to stand for something else. Not all surfaces of this kind are depictions, though. Depictions visually resemble, to a certain extent, what they stand for, or so it seems: portraits, landscapes, genre-pictures are typical examples of depiction[1]. A key topic for theories of depiction concerns the explanatory value of this intuition: 1) What does it mean to say that depictions resemble what they stand for? 2) Does resemblance explain depiction? The kind of representation peculiar to depiction is usually called “pictorial representation” (more on this below).

Accounts of depiction can be distinguished between those attributing an explanatory role to resemblance – see, among others, Budd (1993); Peacocke (1994); Hopkins (1998); Hyman (2006); Abell (2009); Newall (2011) – and those focussing on other aspects of pictures, be they conventional, experiential or cognitive – see especially Goodman (1968); Schier (1986); Wollheim (1987); Lopes (1996). Among resemblance-based accounts we can distinguish between those concentrating on objective resemblances between pictorial surface and aspects of other objects in the visual world – Hyman (2006) – and “experiential” approaches, i.e. accounts claiming that certain aspects of the pictorial surface of a painting arouse in the viewers a peculiar experience of resemblance between the pictorial surface and depicted objects/scenes – see especially Hopkins (1998).

Wittgenstein’s remarks on “seeing-as” have influenced several scholars working on depiction. They have especially inspired those, such as Gombrich (1960) and Wollheim (1968; 1987), who think that in order to understand depiction we should understand the specific kind of visual experience depictions arouse in the viewer. In this paper I would like to go a different way. My hypothesis is that certain of Wittgenstein’s claims both in the Tractatus and in his later writings resonate well within the context of an objective resemblance account of depiction.

1. Objective resemblances

Depiction is a practice we find developed in a variety of human cultures across time and space. This practice, as I understand it, is based on three assumptions: 1) the assumption that we can render a three-dimensional visual world on two dimensions; 2) the assumption that the physical support of depiction is to be looked at frontally (this admits for some variations, which are, however, no more than variations from a widespread standard)[2]; 3) the assumption that we can exploit certain objective resemblances between aspects of a two-dimensional surface and aspects of three-dimensional objects in order to make the surface stand for such objects. These three assumptions need not be explicitly stipulated, but they are embedded in the practice of depiction. 3) is a special way of making 1) work and it is at the core of Hyman’s proposal on depiction (see Hyman [2006]; [2013]). 1) holds for both depictions and other kinds of visual representation on two dimensions (e.g. graphs). 3), instead, only holds for depictions and for all depictions. In the practice of depiction we need to move from the assumption that we use certain specific objective resemblances in order to depict objects, because that there are objective resemblances between two objects does not per se mean that one depicts the other, and because there can be more objective resemblances between a picture and the object it depicts than those that are exploited by depiction (see Goodman [1968]: chap. 1).

Styles and conventions governing depictive practices vary enormously across cultures. Basic resemblance aspects, however, are relevant to all depictions. First, as Hyman claims, «a picture represents an object by defining its form» (Hyman [2006]: 81). Depicted objects are always depicted relative to a line of sight, which does not need to be explicitly marked on the surface of the picture, but is always identifiable (certain pictures, e.g. cubist paintings, present objects as depicted from multiple lines of sight). The real contour of an object, depicted relative to a line of sight, is its “occlusion shape”.

The concept of occlusion shape goes back to Euclid’s Optics, which begins by postulating a thick cone of rays connecting the eye with an object that is seen. The apex of the cone is in the eye and its base is on the visible surface of the object. Of course this cone of rays is an idealization, since the visible surface of an object may have any shape and the direction in which light travels is affected by refraction. But an object’s occlusion shape is the shape of a perpendicular cross section of this so-called cone or solid angle. It is therefore possible to identify the occlusion shape of an object, relative to a given line of sight, by measurement and calculation. But it is also possible to see it […]. Since few occlusion shapes have names, discursive treatments of the topic tend to concentrate on a few simple geometrical examples. But we can see, recognize, and record an unlimited variety of occlusion shapes and not merely these few. (ivi: 76)

It is crucial to stress that the occlusion shape of an object is defined as a real aspect of the object as seen from a certain point of view, and not as something that is “in the eye of the observer”. Therefore, resemblance in occlusion shape is resemblance between an objective aspect of a pictorial surface and a kind of two-dimensional visual aspect that many objects can present (e.g. a circular shape resembles the shape of a coin seen frontally, but also that of a ball seen frontally, of the full moon seen from Earth, etc.).

There are two more objective resemblance aspects which are key features of depiction, according to Hyman: resemblances between the reciprocal dimensions of the shapes of objects on the pictorial surface and the reciprocal dimensions of the shapes of certain kinds of objects considered as parts of the same scene, i.e. resemblances in “relative occlusion size” (ivi: 98-9); resemblances in colour between portions of the pictorial surface and the kind of objects that have the same colour (ivi: 99-104). Often, the portion of a depicted surface representing a certain uniformly coloured object is actually made of several brushstrokes of different colours, which are supposed to render the way light hits the represented object. The objective resemblance, then, holds between the “aperture colour” of a minimal portion of the pictorial surface (the colour of a minimal uniformly coloured part of the pictorial surface as seen through e.g. a cardboard tube, in order to avoid any change in illumination) and the kind of identically coloured objects.

These three aspects of objective resemblance, the basic resemblance properties, determine the kind of objects we can say a picture depicts, i.e. the kind of objects that present a certain aperture colours and/or occlusion shape and/or occlusion size (relative to that of other objects), when seen from a certain standpoint. Each basic resemblance aspect is necessary in order to depict the aspect it pinpoints (shape, reciprocal sizes of objects, colour). However, basic resemblance aspects are not sufficient for depiction, neither conjointly nor disjointedly. Basic resemblance aspects are not sufficient to understand that a certain surface counts as a depiction, because when we notice that an object resembles other objects in a certain respect we are not automatically entitled to claim that it is a depiction of such objects. According to Hyman, authorial intentions (intentions to depict such and such thing) do not provide a standard of correctness for depiction, although an intention to depict something is necessary for there being depiction (ivi: 136-138). He also claims that there are «two basic kinds of rules […] involved in making and perceiving pictures» (ivi: 171), iconographic conventions and technical rules: «whereas the spectator needs to be aware of an iconographic convention in order to arrive at the specific depiction it licenses, her ability to identify bamboo in an ink painting need not depend on her knowledge of the techniques that guided its design and execution» (ivi: 173).

Hyman distinguishes between the pictorial content of a picture’s design (see ivi: 63-64, 68, 71, 105) and the object a picture may portray. A picture’s design is the marks and colours on its surface and the pictorial content of a picture’s design is determined by the three basic resemblance aspects. According to Hyman, the content delivered by a picture’s design is like the sense of a name without reference, whereas the portrayed subject of a picture is like the reference of a name. Only portraits and, more generally, pictures depicting particulars that exist in the actual world have portrayed subjects, i.e. reference and sense, whereas pictures of objects of a certain kind only have sense (see Hyman [2013]).

How does this theory answer the key questions for a theory of depiction, i.e. 1) What does it mean to say that depictions resemble what they stand for? 2) Does resemblance explain depiction? As for 2), the theory allows to claim that it is the objective resemblance between configurations on the pictorial surface and aspects of visual objects that explains the peculiar form of representation displayed by depictions. As for 1), depictions always resemble what they depict, but what a picture depicts is not always what it stands for. However, depiction is necessary in order to make a pictorial surface stand for (i.e. represent) a certain object in a peculiarly pictorial way, i.e. to represent pictorially. Hyman claims:

The resemblance theory […] presupposes that whatever a picture represents, it represents by representing visible objects, their relations to each other, and their parts; and it is the representation of these visible objects – the ones that are not represented by representing something else – that the theory is intended to explain. (ivi: 64)

Although the theory is intended to explain the representation of visible objects that are not represented by representing something else, is there anything it can help us to say about the objects that are represented by representing something else?

©Copyright Steve Bell 2011- All Rights Reserved

Abell, in particular, observes that Hyman cannot explain pictorial misrepresentation (Abell [2009]: 189). A typical case of misrepresentation are highly anti-naturalistic caricatures such as Steve Bell’s drawings of George Bush Jr. with a cannon-mouth or, more recently, of David Cameron with a condom-head. Neither Cameron has a condom-head nor Bush Jr. has a cannon-mouth, but we easily identify them as the subjects Bell’s caricatures represent. How so? According to Abell, in order to explain caricature Hyman should adopt a counterfactual solution: he should claim that Bush Jr.’s caricature represents Bush Jr. as if his mouth were shaped like a cannon. But this, Abell argues, would not explain why we understand the caricature as depicting Bush Jr. rather than any other person or thing.

I believe this objection overlooks the point that for Hyman a picture can represent X while depicting Y. In other words, according to Hyman visual representation through pictures is not exhausted by basic depiction. The Bush Jr. caricature represents a figure that has the outline shape of a man (among other things), except for the part that is usually occupied by the mouth, which is replaced by a form that has the outline shape of a cannon (among other things). Our ability to recognize Bush Jr. as the subject of the picture – I take it – depends crucially on the following fact: we can guess some reasons why Bell might have wanted to represent Bush Jr. with a cannon-mouth. To formulate a hypothesis it is sufficient to know that the caricature was made in 2003, at the beginning of the Second Iraq War. This information allows us to trace links between what we identify as the depictive content of the picture and the subject it portrays. Therefore, that the picture has a certain depictive content is relevant for us to understand its subject. More precisely, we understand that a cannon is used to represent Bush Jr.’s mouth pictorially because the picture is intended to work as a visual metaphor: in 2003 Bush Jr.’s mouth was like a cannon, because he was verbally expressing his will to attack Iraq. Moreover, in this specific case, we know what Bush Jr. looks like and identify certain of his features in the caricature: in other words, there are certain genuine aspects of resemblance that the depictive content of the caricature shares with Bush Jr., namely those depicting his hair, forehead and, to a certain extent, his eyes. This gives us further reasons to identify with Bush Jr. the man-with-the-cannon-mouth. However, that there are some aspects of objective resemblance with the subject is not necessary for a depiction to work as a caricature-portrait. We can conceive of a picture of a cannon-with-a-mouth, which does not visually resemble Bush Jr. in any objective respect, and which, in the appropriate context, can be understood as a pictorial representation of Bush Jr. For instance, on the background of the cannon we could put an American flag and give the picture the title Presidential address, February 2003.

The theory, however, is exposed to another objection: it might seem that it cannot explain the difference between a pictorial representation of X and a picture that non-pictorially represents X. I believe, however, that also this objection fails. If a picture pictorially represents X it depicts either X or some other object Y whose visual features can be exploited in order to represent X. On the other hand, a picture that non-pictorially represents X might be a picture that depicts Y, where Y is used to represent X by means of a convention for which Y’s visual features are not relevant. The case of a non-pictorial visual representation of X differs from certain highly anti-naturalistic pictures, such as certain so-called “hermetic” cubist pictures, because there is a difference between conventions tout-court and depictive conventions. In a visual work that is a depiction, the depictive content is the means (or, at least, one of the means) to convey meaning. If basic resemblance aspects are not used to establish a link between depictive content and portrayed subject then they have to be used to make other points that are relevant to understanding the respect under which the portrayed subject is shown. In other words, among the variety of styles of visual representation, there are depictive styles, which exploit depictive content in order to convey meaning, even if they do not exploit a depictive content that resembles their portrayed subject. In Picasso’s portrait of Daniel-Henri Kahnweiler (1911), for instance, the depictive content is given by the intersections among several triangles, rectangles, squares and parallelograms. Only a few of those are used to establish a link between the picture’s depictive content and its portrayed subject. Two eyes and a nose are suggested by a certain conglomerate of geometric figures and we also see a few wavy signs, reminding of hair. All the rest of the depictive content is just a great articulation of geometric figures, which are in their amalgam capable of evoking various parts of a human body, as of many other objects, as seen from multiple viewpoints. This makes for the very distinctive depictive style of the pictorial representation: by means of fracturing visual objects into geometric forms and making the portrayed subject of the picture almost ungraspable through a simple look at it, Picasso made a very bold point on portraiture. Portraiture here is not conceived as a matter of copying a certain subject, because a pictorial rendition of the world can be a lot more expressive, can tell a lot more, if it exploits visual features peculiar to paintings, such as the fact that with a painting we can give the impression of a world of ever-changing objects, where the features of a human face can barely be grasped because they are part of the constant stream of mutable forms. We are required to observe how most of the depictive content is manipulated deliberately in order to make it impossible for us to link it to a certain subject. This, then, is a way of using a great part of the depictive content not in order to “embody” a certain subject, but in order to tell us, with peculiar pictorial means, something about the reasons why we see so little of the subject in the picture. In a non-pictorial visual representation, such as a graph, instead, the specific outlines, colours and reciprocal occlusion shapes of the marks and colours on the pictorial surface are not necessarily relevant in order for the viewer to grasp the information conveyed by the picture. Moreover, the image does not need to exploit basic resemblance aspects at any level in order to work as a representation.

So far I have considered anti-naturalistic pictures and caricatures that have sense and reference. What about non-naturalistic pictures that only have sense, i.e. depictions of objects of a certain kind, whose features we can at least in part describe thanks to our acquaintance with certain particulars? Think of another analytic cubist painting, Picasso’s The Glass (1911). Thanks to our acquaintance with the visual world, we are in the position to describe the picture not just as the depiction of an object of a certain kind, where the kind is the kind of objects that have such and such shapes and such and such colours (objects that we are unlikely to encounter in the real world, for that matters), but where the kind of objects that is depicted is also the kind of objects that are concave, since we see a number of concave forms, and, possibly, the kind of objects that can stand on a surface (since we see something that looks like an object standing on a surface), etc. Depictive aspects are conjured up in such a way that not only we can recognize aspects of particulars we are acquainted with but we also feel that it is appropriate to describe the picture, in part, in terms of such recognized aspects. On the opposite, we do not feel entitled to give such a description of those abstract pictures that do not prompt us to recognize aspects of objects we are acquainted with as depicted by them.

It seems to me, then, that Hyman’s theory of depiction can tell us something about how depictions come to have subjects that do not necessarily resemble the configurations of marks and colours on their surface. The theory, then, has a considerable explanatory power. I shall now turn to Wittgenstein and link Hyman’s proposal to some of his remarks on pictures.

2. Tractarian Bilder and pictorial representation

In the Tractatus Wittgenstein famously claims that propositions are Bilder, pictures: «A proposition is a picture of reality. A proposition is a model of reality as we imagine it» (TLP, § 4.01). “Reality” means the totality of the states of affairs (TLP, §§ 1; 2), both those that exist and those that do not (TLP 2.06), i.e. all the states of affairs that are logically possible, which are all the states of affairs we can talk about. The representational character of language is explained by means of a concept of picture, Bild, that is better exemplified by conventional representations, such as representations of music by means of notes on pentagrams (TLP, §§ 4.011; 4.014-41), hieroglyphs (TLP, § 4.016), or abstract mathematical models – as Wittgenstein told Waissmann (WWK, 185/173-4) – rather than by pictorial representations, which generally strike us because of their resemblance to other objects (be it objective or not)[3]. However, according to the “first” Wittgenstein, the Tractarian concept of picture explains a feature of all forms of representation: «Every picture is at the same time a logical one» (TLP, § 2.182). This is because a logical picture is nothing but thought, which is essential to any representation of the world that we might produce (TLP, § 3). «“A state of affairs is thinkable”: what this means is that we can picture it to ourselves» (TLP, § 3.001). It seems, then, that the Tractatus cannot help us understanding the specific features of a particular kind of pictures, pictorial representations, given that Wittgenstein claims in this work concern the fact that all representations (and, among them, all pictures) are logical pictures of the states of affairs they represent.

In the Tractatus, however, Wittgenstein makes a central claim that, I believe, may turn out to be fruitful if applied to the analysis of pictorial representation. He distinguishes between proposition (or picture), which, in its entirety, does not need to be correlated to an occurring state of affairs in order to make sense – which has meaning even if it is false, in other words –, and elements of the proposition (or picture), which are correlated to the simple objects that are the basic constituents of the represented state of affairs and are intertwined to each other by means of relations of the same abstract type of those holding among the basic constituents of the represented state of affairs (TLP, §§ 2.12; 2.17). This means that between a representation and the state of affairs it represents there must be a structural isomorphism in order for the first to be a representation of the second. A good example is provided by musical notation (TLP, § 4.014): in a music sheet notes are spatially ordered from left to right, whereas the same notes, when played, are temporally ordered. The music sheet is a representation of music because the relations among its constituents (the spatially ordered notes on the pentagram) correspond to (are of the same abstract type of) the relations among the constituents of the piece of music that they represent (the temporally ordered sounds), although they are not identical to each other (being, respectively, spatial and temporal in character).

According to the Tractatus, then, each simple name in a proposition (each basic element of a representation, Bild) is related to a simple object; now, going back to Hyman’s theory of depiction, we have seen that basic elements of the pictorial surface depict visual aspects of the extra-pictorial world in that they resemble them. In Wittgenstein’s Tractatus, then, we have names (basic elements of the proposition/Bild) that stand for simple objects (i.e. logical objects). Analogously, in Hyman’s theory of depiction we have basic resemblance aspects that stand for aspects of visual objects or kinds of visual objects, which are the basic constituents of depiction, the basic elements by means of which a depiction works as a representation. Moreover, in the Tractatus the Bild/proposition does not have to represent a state of affairs that occurs, but can also represent states of affairs that do not occur, and the same, according to Hyman, is true of a depiction, which can have either only sense (and therefore represent state of affairs that do not occur) or reference and sense. It is also relevant to notice that the distinction between pictures with sense and pictures with sense and reference was put forward by Wittgenstein in the following passage from Philosophical Investigations:

522. If we compare a proposition to a picture, we must think whether we are comparing it to a portrait (a historical representation) or to a genre-picture. And both comparisons have point.

When I look at a genre-picture, it “tells” me something, even though I don’t believe (imagine) for a moment that the people I see in it really exist, or that there have really been people in that situation. […]

A portrait is a picture that has reference and sense, whereas a “genre-picture” is a picture that only has sense. I shall come back to this below.

3. Pictures after the Tractatus

As it is well known, after the publication of the Tractatus Wittgenstein came to criticize the view defended in the book, especially because of the ontological commitment to the existence of simple objects it implied. The post-Tractarian Wittgenstein focuses on how we use propositions to assert and deny, how the same proposition can be used to express different meanings in different contexts (“good job!”, for instance, can express both a positive and a negative evaluation), and the fact that there are propositions which are meaningful even though they do not seem to refer to any clearly determinable state of affairs (e.g. “stay more or less there!”). Moreover, he stresses that in order to understand how we can express our thoughts and how this happens (the main explanatory goal of the Tractatus) we need to take into consideration also context-dependent aspects of language, such as indexicals (“I”, “here”, “now”, etc., see PI, I, § 264), the force of our assertions, orders, prayers, etc. (see PI, I, § 27), as well as the fact that certain thoughts are not expressed by means of verbal languages but, for instance, through gestures (see Malcom [1958]: 69).

Wittgenstein put forward these remarks on thought and language in Philosophical Investigations: thinking amounts to expressing oneself through language or some other form of representation, all forms of representation are activities, representations are acts performed in a context, games played while following certain shared rules. It might seem that in the Investigations not much room is left to an analogy between language and images, not even if images are conceived as abstract visual models, because the depictive conception of language, as a complete and satisfactory conception of how language works, has been rejected by Wittgenstein. However, we see that in the Investigations as well as in other writings of the “later” Wittgenstein a parallelism between images and language is used to explain how language games work: I believe this can be useful also for an analysis of pictorial representation. Let us examine a few passages.

“What makes a portrait a portrait of Mr. N?” The answer which might first suggest itself is: “The similarity between the portrait and Mr. N”. […] It is quite clear, however, that similarity does not constitute our idea of a portrait; for it is in the essence of this idea that it should make sense to talk of a good or a bad portrait. In other words, it is essential that the shadow should be capable of representing things as in fact they are not.

An obvious, and correct, answer to the question  “What makes a portrait the portrait of so-and-so?” is that it is the intention. (BBB: 32)

If we relied on the intuitive idea that a portrait represents its subject in virtue of the fact that it resembles it, Wittgenstein observes, we could not explain how it is that we distinguish between good and bad portraits. There is nothing like “good” or “bad” resemblance: either two objects resemble each other or they do not. In order to understand portraits we need to introduce the concept of intention, Wittgenstein claims. His conception of “intention”, however, is anything but straightforward, since Wittgenstein discards any psychological explanation of intention.

“Only the intended picture reaches up to reality like a yardstick. Looked at from outside, there it is, lifeless and isolated” – It is as if at first we looked at a picture so as to enter into it and the objects in it surrounded us like real ones; and then we stepped back, and were now outside it; we saw the frame, and the picture was a painted surface. In this way, when we intend, we are surrounded by our intention’s pictures, and we are inside them. But when we step outside intention, they are mere patches on a canvas, without life and of no interest to us. (Z, § 233)

The sense in which an image is an image is determined by the way in which it is compared with reality. (BBB: 53)

No image without intention: what we have without intention are mere surfaces covered in marks and colours. An intention is a way of comparing image and world. What does this mean? I think this passage from The Big Typescript might be illuminating:

“This figure in the picture is I” is an agreement.

Fine, but about what are we agreeing? What relation are we establishing between signs and myself? Well, nothing other than the one that exists, say, by pointing with one’s hand or attaching a label. For this relation is only meaningful because of the system to which it belongs. […]

The portrait is only a picture similar to N (or not even that). But it contains nothing (no matter how similar) that would make it a portrait of this person, i.e. the intended portrait. (Indeed, the picture that looks virtually identical to one person can actually be a bad portrait of someone else.) (TBT, 227e /292)

To intend a picture is to endow it with meaning, we read in the passage from Zettel; to endow a picture with meaning consists in comparing it with the visual world, Wittgenstein explains in the passage from the Blue Book. This comparison, we read in the passage from The Big Typescript, takes place within a system of rules. Provided that author and observer share the use of the system of rules adopted in the making of a given picture, then the intention of the author and that of the observer can overlap: they endow the picture with the same meaning. As we have seen, according to Wittgenstein, mastering a language (i.e. being able, among else, to express our thoughts) amounts to being able to apply the rules of such language, but not necessarily to being able to explain such rules or to answer questions such as “why did you apply this rule instead that another one?” For a philosophical analysis, then, it does not make sense to ask what one is thinking about when one has a certain intention, as if thoughts and intentions were only accessible to the subject who has them.  Thoughts and intentions are accessible in that they exist in so far as they are expressed, and are thereby accessible to philosophical scrutiny. It follows that, when Wittgenstein claims that in order to understand what is a picture we should think that it is the product of an intention, which is a way of comparing the picture to the world, in the context of a certain system of representation, what he is saying is that in order to understand the picture all we have to do is looking at the picture itself and trying to grasp how it works as an expression of the “form of life” (see, for instance, PI, I, §§ 19, 23) it has been produced within. We understand a picture only in so far as our form of life has something in common with the form of life of its author. In order to understand a picture we look for guidance in the background of information and practices we share (or believe we share) with its author; a picture is not something we can decode by means of applying a scheme of correspondences between its elements and what it represents.

By talking about pictures Wittgenstein here talks about our use of representations in general, and especially of language. But his remarks are also relevant for an analysis of pictorial representation in terms of the objectivist account put forward by Hyman. As we have seen, a picture objectively resembles what it depicts, but this might not be sufficient for us to grasp its sense or its reference. In order to do this we will have to take into consideration features that are specific to the “form of life” within which the picture has been produced, to say it with Wittgenstein, i.e. certain stylistic features. This point of contact notwithstanding, one might think that Wittgenstein stresses only the conventional aspects of representation, unlike Hyman. I believe, however, that this is not the whole story Wittgenstein has to tell us. Let us look at two passages from the Investigations, one of which I have also quoted in the previous section:

522. If we compare a proposition to a picture, we must think whether we are comparing it to a portrait (a historical representation) or to a genre-picture. And both comparisons have point.

When I look at a genre-picture, it “tells” me something, even though I don’t believe (imagine) for a moment that the people I see in it really exist, or that there have really been people in that situation. But suppose I ask: “What does it tell me, then?”

523. I should like to say “What the picture tells me is itself.” That is, its telling me something consists in its own structure, in its own lines and colours.

Here, as I have said before, Wittgenstein contrasts two kinds of pictures: portraits and genre paintings. Portraits are “transitive” pictures: they can be linked to a referent in the real world, they have reference. Genre-paintings (pictorial representations of real life scenes in fictional worlds), instead, are “intransitive” pictures, which cannot be linked to any referent in the real world, they only have sense, as Hyman puts it. Wittgenstein observes that we do not even need to imagine that the objects depicted in genre paintings must exist or have existed in order to understand what such paintings “tell us”: “What the picture tells me is itself”. How can we make sense of this claim? Pictures as such, I believe, are endowed with meaning in that they are intentional artefacts that have been produced within a certain “form of life”. Pictures, as Wittgenstein explains in § 523, tell us something by means of presenting certain forms and colours on their surfaces, which, we could say with Hyman, objectively resemble all those visual objects, or kinds of objects that are characterized by the same forms and colours. The basic resemblance aspects of Hyman’s theory are pictorial conventions that are shared by any “form of life” where pictorial representation takes place, they are the conventions that determine the presence of pictorial representations as a specific kind of artefact within any given culture. The content determined by the “lines and colours” of a certain pictorial representation, according to such basic rules, is its pictorial content, which guides us in understanding the sense, or the reference, of the picture, together with other conventions, which vary according to the “form of life” within which the picture has been produced (these are, in other words, stylistic conventions).

Wittgenstein, then, does not directly appeal to resemblance when talking about pictorial representation, but he does not focus only on their conventional aspects either. Since he defends the claim that being able to use a certain language amounts to grasping its rules and differs from being able to explain how such rules work, in the case of pictorial representations he has to show what it means to grasp the rules that govern this form of representation without being in the position of identifying in advance, with certainty, the sense or the reference of any given pictorial representation. What he says is that we have to look at the lines and colours on a pictorial surface, keeping in mind that they are being used within a certain form of life, but he does not give us an account of how exactly lines and colours can guide us in the process of attribution of sense or reference to pictorial representations. I hope this paper can bring a little contribution to the understanding of how this process works[4].


Abell, C., 2009: Canny Resemblance, “Philosophical Review”, 118, pp. 183-223.

Budd, M., 1993: How Pictures Look, in D. Knowles and J. Skorupski (eds.), Virtue and Taste, Blackwell, Oxford, pp. 154-175.

Frascolla, P., 2000: Il Tractatus Logico-Philosophicus di Wittgenstein, Carocci, Roma.

Gombrich, E., 1960: Art and Illusion, Phaidon, London and New York.

Goodman, N., 1968: Languages of Art, Hackett, Indianapolis and Cambridge.

Hopkins, R., 1998: Picture, Image and Experience, Cambridge University Press, Cambridge.

Hyman, J., 2006: The Objective Eye, The University of Chicago Press, Chicago and London.

Hyman, J., 2013: Depiction, “Royal Institute of Philosophy Supplement”, 71, pp. 129-150.

Kulvicki, J., 2006: On Images, Oxford University Press, Oxford.

Lopes, D., 1996: Understanding Pictures, Clarendon Press, Oxford.

Malcom, N., 1958: Ludwig Wittgenstein: A Memoir, Oxford University Press, Oxford.

Newall, M., 2011: What is a Picture? Depiction, Realism, Abstraction, Palgrave Macmillan, Basingstoke.

Peacocke, C., 1987: Depiction, “Philosophical Review”, 96, pp. 383-410.

Peirce, C., 1984: Writings of Charles S. Peirce. A Chronological Edition, ed. by E. Moore, M. Fisch, C. Kloesel, D. Roberts, L. Ziegler, Indiana University Press, Bloomington.

Schier, F., 1986: Deeper into Pictures, Cambridge University Press, Cambridge.

Wittgenstein, L., 2008, 1st edition 1921: Tractatus Logico-Philosophicus, Routledge, Abingdon-New York.

Wittgenstein, L., 1968, 2nd edition: Philosophical Investigations, Basil Blackwell, Oxford.

Wittgenstein, L., 1979, 2nd edition: Notebooks 1914-1916, ed. byG. E. M. Anscombe and G. H. von Wright, Basil Blackwell, Oxford.

Wittgenstein, L., 1980, 2nd edition: Preliminary Studies for the Philosophical Investigations, Generally Known as the Blue and Brown Books, Basil Blackwell, Oxford.

Wittgenstein, L., 1981, 2nd edition: Zettel, ed. by G. E. M. Anscombe and G. H. von Wright, Basil Blackwell, Oxford.

Wittgenstein, L., 1984: Ludwig Wittgenstein und der Wiener Kreis. Gespräche, aufgezeichnet von Friedrich Waismann. Aus dem Nachlaß herausgegeben von B. F. McGuinness, Werkausgabe Bd. 3, Suhrkamp, Frankfurt a.M.

Wittgenstein, L., 2005: The Big Typescript: TS 213, edited and translated by C. Grant Luckhardt and Maximilian A. E. Aue, Blackwell, Oxford.

Wollheim, R., 1968: Art and its Objects,Cambridge University Press, Cambridge.

Wollheim, R., 1987: Painting as an Art, Thames & Hudson, London.

Abbreviations of Titles of Wittgenstein's Works Cited

TLP: Tractatus Logico-Philosophicus, Routledge, Abingdon-New York 2008.

PI: Philosophical Investigations, Basil Blackwell, Oxford 1968.

BBB: Preliminary Studies for the Philosophical Investigations, Generally Known as the Blue and Brown Books, Basil Blackwell, Oxford 1980.

Z: Zettel, ed. byG. E. M. Anscombe and G. H. von Wright, Basil Blackwell, Oxford 1981.

WWK: Ludwig Wittgenstein und der Wiener Kreis. Gespräche, aufgezeichnet von Friedrich Waismann. Aus dem Nachlaß herausgegeben von B. F. McGuinness, Werkausgabe Bd. 3, Suhrkamp, Frankfurt a.M. 1984.

TBT: The Big Typescript: TS 213, edited and translated by C. Grant Luckhardt and Maximilian A. E. Aue, Blackwell, Oxford 2005.

[1] C. S. Peirce (1984) called pictures of this kind “icons” and distinguished them from symbols (which are pictures that represent by means of a convention, such as traffic signs) and indexes, which are pictures that represent what has caused them (e.g. footprints, which stand for the person who has left them). Whereas depiction is generally considered distinctive of icons – see Kulvicki (2006) for a different account – symbolic representation and index-based representation can be accomplished both through pictures and through objects other than pictures (e.g. words conventionally stand for objects and smoke is and index of fire).

[2] See Hyman’s discussion of anamorphic pictures (Hyman [2006]: 94).

[3] In TLP 3.1431 Wittgenstein mentions «spatial arrangement(s)» of objects as a good way to visualize the «essence of a propositional sign», such as three-dimensional models of states of affairs. These models, however, are interesting to Wittgenstein not because they can be realistic visual renditions of the states of affairs they represent, but because their elements can be put in one-to-one correspondence with the elements of the represented state of affairs.

[4] I am very grateful to John Hyman, Enrico Terrone, Gabriele Tomasi and Alberto Voltolini for their precious comments, and to Steve Bell for allowing me to reproduce his work.

DOI: http://dx.doi.org/10.13128/Aisthesis-12849

Creative Commons License

This work is licensed under a Creative Commons Attribution 4.0 International License (CC-BY- 4.0)

Firenze University Press
Via Cittadella, 7 - 50144 Firenze
Tel. (0039) 055 2757700 Fax (0039) 055 2757712
E-mail: info@fupress.com