1. Introduction

Conceptions of pictorial realism abound, and perhaps only a philosopher would fret over such abundance as is to be found in this symposium. Saying that Abell, Armstrong, and McMahon offer us three accounts of three different phenomena rather than three competing accounts of one or two phenomena seems to give up the search for systematic and unifying explanations. Art is a messy business, however, and we should not seek more order in our explanations than there is in the world. So let's first look at the many apparently distinct notions of pictorial realism, and then weigh the prospects for house cleaning.

2. Special Realism

There may be as many conceptions of pictorial realism as there are attributions of it. Just think of the many contexts in which pictorial realism is attributed, which run from informal chat about pictures to the formal discourses of art criticism, art history, and art theory. Some of these attributions are obviously idiosyncratic. (I was recently asked, as a member of a jury, to award a special prize for "realism," which turned out to mean "landscape.") Setting aside the idiosyncrasies, here are several theoretically robust conceptions of realism.

One Realism. McMahon thinks of all pictures as realistic. Their realism distinguishes them from other kinds of representations, notably linguistic descriptions. Thus Brueghel's painting of the fall of Icarus is realistic in a way that William Carlos William's description of the very same scene is not, just because the former but not the latter is a picture. Of course, not every property of pictures that we are interested in has to do with representation: the view is that pictures are realistic as representations. Moreover, pictures may represent by a mixture of pictorial and other means (even linguistic means), so the view is that pictures are realistic as pictorial representations. On this conception, pictorial realism is not a matter of degree. "Realistic" modifies a kind of representation - depiction.

An account of realism so conceived is an account of depiction, and that is what McMahon proposes. P depicts O as F only if its surface is seen as an F configuration. For example, Rembrandt's drawing depicts something as a boat only if its surface is seen as a boat configuration. What configuration is seen depends on sub-personal mental processing that uses an object-centered (rather than view-pointed) description in iconic memory. The surface is seen to have a different configuration if a different object-centered description is used. So there may be many ways to depict O as F: many differently marked surfaces may be seen as F configurations, for it is an empirical matter what object-centered descriptions are stored in iconic memory. Moreover, we learn to see new configurations when new object-centered descriptions are added to memory.

Slow-Dawning Realism. McMahon also observes that realism may dawn slowly. A picture may look unrealistic for quite some time until its realism comes into sight. McMahon proposes that her account explains this phenomenon: we are let in on a picture's realism as we acquire the object-centered description that enables us to see the required configuration. However, we need not assume that only One Realism dawns slowly. The reception of Impressionism illustrates the phenomenon of Slow-Dawning Realism, if we are to believe the usual story. According to that story, Impressionist pictures looked at first quite unrealistic but now they look highly realistic. It is implausible, however, that their initial audience failed to see the configurations of boulevards, umbrellas, dancers, and flowers that they present. Slow-Dawning Realism is independent of One Realism.

One Gradual Realism. One Realism may be amended to yield a conception of realism that comes in degrees. Distinguish pictorial content - the properties a picture depicts a scene as having - from design - the properties of a picture's surface (partly) in virtue of which it has its pictorial content. Now distinguish design from surface tout court. Some surface properties of some pictures are not design properties: these properties of a picture might vary as pictorial content remains the same. Some are not at all representational (e.g. the scratches across some of Rembrandt's etchings), while others represent non-pictorially (e.g. the text in a cartoon thought balloon). Given this apparatus, we may say that some pictures are the more realistic the fewer non-design surface properties they have or the larger the ratio of design surface to non-design surface. The idea is that in more realistic pictures the resources of the medium are more completely directed to depiction.

Lifelike Realism. One Realism, whether gradual or not, is a feature of all pictures, in so far as they depict. Armstrong outlines a conception of realism according to which only a very few pictures are highly realistic. These pictures depict things as having features that give them a lifelike appearance, "features which are central to our encounters with real objects." Think of these as the features that distinguish the appearance of the real from the appearance of the unreal. Armstrong proposes that the features are visual detail, weight and solidity, and vitality and motion. Of course, one might accept the general characterization of Lifelike Realism as the depiction of features which are central to encounters with the real and either reject Armstrong's list or accept it as one of several equally good lists of features which support Lifelike Realism.

Uncanny Realism is an interesting phenomenon made available by tweaking the parameters of Lifelike Realism. Some pictures achieve Lifelike Realism by depicting things as having some but not all of the features characteristic of our encounters with real objects. When taken to an extreme, as when enormous gaps in visual detail are combined with the depiction of weight, solidity, vitality, and motion, the apparently real is apparently unreal, and that is uncanny.

Illusionistic Realism. A picture that is realistic in this sense can be viewed in conditions where it is indistinguishable from the scene it depicts. However, Lifelike Realism does not imply Illusionistic Realism. First, Lifelike Realism is achieved by depicting things as having some properties salient in our encounters with things, and that is consistent with not depicting things as having all the properties they are seen to have in the flesh. Lifelike Realism, especially in its Uncanny variety, may result from a few deft strokes of the pen. Second, Lifelike Realism is a matter of what features scenes are depicted as having, not of how they are depicted. One might depict weight and solidity by marks that always betray the process of depiction.

Idolic Realism, Realism in Looking. E. H. Gombrich characterizes realism not as a property of pictures but rather as a transaction between pictures and their spectators.[1] The idol was made to stand in for the god and is therefore to be treated as the god. Put another way, the idol mandates that spectators imagine that the space they occupy is one where the god also sits. That is a kind of realism - a taking for real. Its complement is a kind of realism in which the depicted scene is placed in a space apart, available for scrutiny. One no more scrutinizes the idol than one would scrutinize the god himself, but the realism of much painting after the Renaissance requires and builds upon an intense looking that is inappropriate except in pictures.

True Realism. Some pictures serve as sources of information. A picture is realistic to the extent that it accurately informs us about the appearance of the scene it depicts: its content is accurate. That is True Realism.

Informative Realism. I may speak the truth and fail to inform, as we learned from Paul Grice; and, likewise, an accurate picture may fail to inform because it does not convey information that is relevant in the circumstances. I am right to deem as uninformative the perfectly accurate stick figure you drew when I asked you what David Lewis looks like. Abell proposes an account of Informative Realism. Pictures are realistic to the extent that they are informative about the appearance of scenes, where informativeness implies relevance. Information is relevant to a thinker to the extent that it yields new information when added to information she already has. An informative picture of Lewis added to some perceptual information yields the new information that Lewis is in the corner arguing with Kripke. The stick figure does not yield this new information; it is less informative, hence less realistic.

Revelatory Realism. Informative realism begets another realism.[2] Some pictures convey information that cannot be gathered visually, for example, pictures of tiny, gigantic, and inaccessible objects, and also pictures that depict scenes as having combinations of properties they could not be seen to have. These pictures may add to our store of information in eye-opening ways. They are revelatory, informative in relevant and also surprising ways.

Suppose that each of these conceptions of realism has a place in some theorizing about pictures. We then have at least ten viable conceptions of special realism, and there may be more. Whatever the number, it is multiplied if we accept certain theses.

Stylistic Realism. Is realism attributable only to individual pictures? Maybe not. According to Stylistic Realism, realism is a property of pictorial styles, as well as individual pictures. A style is a picture kind whose instances share a mode of representation or a kind of content. Abell accepts Stylistic Realism. Any realism in our list can be construed stylistically, with the exception of One Realism. Take One Gradual Realism and construe styles as kinds of pictures individuated by surface determinables, for example, monochromatic and polychromatic pictures are two styles. Some styles are the more realistic the larger the average ratio of design surface to non-design surface features of their members. The polychromatic style is more realistic than the monochromatic style.

Relative Realism. Are attributions of any special realism uniform across contexts? Relative Realism holds that realism is relative to context of use. Abell accepts this also, for what is relevant depends on the state of the spectator - information in P is relevant to S. It is also predicted by McMahon's account of One Realism if users differ in picture-reading skills - that is, if they have different object-centered descriptions in iconic memory and so see different configurations in picture surfaces. Lifelike Realism requires a specification of the features that are central in our encounters with real objects. Do those features vary historically or culturally? If they do, then incompatible attributions of Lifelike Realism are true in different contexts. Thus Armstrong's account is also consistent with Relative Realism.

Normative Realism. Does the realism of a picture or a style of picture number among its merits or demerits? According to Normative Realism, to attribute realism to P is sometimes to attribute a (de)merit to P. It is implausible that calling a picture realistic is always to praise it; realism is a failing in some pictures. Moreover, it is implausible that judgments of realism are always normative. Normative Realism is thus a weak thesis and it is easy to think of cases suggesting its truth.

The special realisms multiply if some apply to styles as well pictures, relativize to context of application, or come in normative as well as descriptive versions. Many realisms indeed!

3. General Realism

The more we multiply the special realisms, the worse appear to be the prospects for a general conception of pictorial realism. Such a conception can be achieved in two ways: the way of contention or the way of abduction.

The way of contention seeks reasons to reject one conception of realism if it is inconsistent with another. One conception of pictorial realism is inconsistent with another just in case attributing the one to any picture precludes attributing the other to the same picture. Theoretical considerations must decide between them. However, Rembrandt's Canal with a Rowing Boat,below, discussed by Armstrong, is a good candidate for realism on many counts. Granted, one might attribute more of one kind of realism to the Rembrandt than another (little Idolic Realism, for example, and much Realism in Looking) but that falls short of inconsistency. Many realisms peacefully co-exist; the way of contention is a dead end.

Rembrandt, A Canal with a Rowing Boat, pen and ink drawing.: This image is used with the permission of Chatsworth Photo Library.  Any form of reproduction, transmission, performance, display, rental, lending or storage in any retrieval system without the written consent of the copyright holders is prohibited.
Rembrandt, A Canal with a Rowing Boat, pen and ink drawing.
This image is used with the permission of Chatsworth Photo Library. Any form of reproduction, transmission, performance, display, rental, lending or storage in any retrieval system without the written consent of the copyright holders is prohibited.

The way of abduction begins with the assumption that "realism" is a theoretical concept, that attributions of realism are attempts to explain something. With that assumption in place, the next step is to see what needs explaining. The conception of realism that does the best explanatory job is the one to prefer.

It is not hard to read off the description of each special realism what explanatory work it is expected to perform, and it is not hard to see that no single explanandum is a strong attractor for all special realisms. Each of the following facts needs explaining: (1) depiction is a distinct type of representation, (2) pictures afford experiences with a distinctive phenomenology, (3) there are many styles of depiction, (4) styles of depiction change over time in a non-arbitrary way, and (5) some pictures are better than other pictures for certain tasks.

This is progress, at least in so far as the list of five explananda is shorter than the list of special realisms. Moreover, it makes sense to group the explananda by their explanantia. Theories of depiction typically target (1) and (2) together. Some theories give accounts of (1) in terms of the phenomenology of pictures,[3] while others propose accounts of (1) which merely suggest accounts of the phenomenology of pictures.[4] As Abell makes clear, Gombrich's "riddle of style" builds on the conjunction of (3), (4), and (5).[5] Pictures come in many styles and styles change over time. Is that because different styles serve different tasks? But all pictures serve to represent appearances! Gombrich has a non-epistemic solution to the riddle; Abell and Lopes have solutions that stress the different contributions pictures can make to what we can know by representing appearances. Finally, accounts of the value of pictures as mimetic representations address (2) and (5) together.[6]

Where is each special realism on this map of explanations and explananda? One Realism and perhaps One Gradual Realism are concerned with (1) and so with theories of depiction. Lifelike Realism, Uncanny Realism, Realism in Looking, Idolic Realism, Illusionistic Realism, and Revelatory Realism are concerned with different ways of understanding (2) and so with theories of depiction and the mimetic value of pictures. True Realism, Informative Realism, Revelatory Realism, and Stylistic Realism are concerned with (3), (4), and (5) and hence with the informative properties of pictures.

There is more order to the topic of realism than one might have expected just on the basis of listing the special realisms, but there is not enough order to shorten the way of abduction. Each notion of realism has a place in at least one of three explanatory enterprises and is not in direct competition with every other notion. Even when two appear to figure in the same enterprise, they may divide their labour. For example, Lifelike Realism and Illusionistic Realism spotlight different elements of the phenomenology of our experiences of pictures, so maybe we need both to explain (2) and the mimetic value of pictures. The way of abduction is long and winding.

We may draw some lessons that will help see how to go forward. First, any special realism is to be treated as an explanatory concept. An analysis of it should look above all to the explanatory work it does. Second, there may be many explanatory jobs to do, requiring many realisms. Third, some realisms may compete to perform the same job. Then we may hire the better candidate. Fourth, the explanations that call for notions of realism set the job descriptions we should use in evaluating the candidates.

The fourth lesson is especially important. For example, sizing up notions of realism for explanations of the phenomenology of our experiences of pictures means itemizing what we need from an account of (2). A great deal has been written about (2), which may be brought to bear on the choice between, say, Illusionistic and Lifelike Realism. Likewise, sizing up notions of realism that for explanations of (3), (4), and (5) means getting into the epistemic role of pictures - something that philosophers have only just begun to do. Realism is not a topic for study in isolation from theories of pictures.

We need not choose between reading the papers in this symposium as disconnected takes on fundamentally different phenomena or as competitors vying for a place in a unified field theory of realism. Taken together, they raise hard but worthwhile questions about what we need to know about pictures: not just depiction, but also the uses and values of pictures. A good way to begin to address those questions is to attend to the abundant attributions of pictorial realism.


1. E. H. Gombrich, "Meditations on a Hobby Horse or the Roots of Artistic Form," in Meditations on a Hobby Horse (Oxford: Phaidon, 1963).return to text

2. Dominic McIver Lopes, "Pictorial Realism," Journal of Aesthetics and Art Criticism 55 (1995), 277-85.return to text

3. For example, Robert Hopkins, Picture, Image and Experience (Cambridge: Cambridge University Press, 1998) and Richard Wollheim, "Pictorial Representation," Journal of Aesthetics and Art Criticism 56 (1998), 217-26.return to text

4. For example, Flint Schier, Deeper into Pictures (Cambridge: Cambridge University Press, 1986) and Dominic McIver Lopes, Understanding Pictures (Oxford: Oxford University Press, 1996).return to text

5. See also Dominic McIver Lopes, "Pictures, Styles, and Purposes," British Journal of Aesthetics 32 (1992), 330-41.return to text

6. For example, Robert Hopkins, "Pictures and Beauty," Proceedings of the Aristotelian Society 42 (1997), 177-94 and Dominic McIver Lopes, Sight and Sensibility: Evaluating Pictures (Oxford: Oxford University Press, 1995).return to text

Department of Philosophy

University of British Columbia

Vancouver Canada V6T 1Z1


Published January 18, 2006