In representational art, form and content intertwine, in ways important to the value of the work. So much is more or less truistic. But how should we think of form, content or their interrelations, so as to make sense of this? In particular, how should we, given the following complication: prima facie, form is a feature of the work itself, of the representation really before us; whereas content, what the work represents, is often fictional, and always absent?

Though he concentrates on what he calls ‘medium’, rather than ‘form’ – a choice that is certainly not merely terminological – Michael Morris devotes this absorbing book to something very like these questions. For each of the arts indicated in the title, his discussion is structured around the need to accommodate the

And in each case, he attempts to accommodate the thesis so as simultaneously to solve a paradox. We can get some sense of that paradox from the following question, framed for the case of painting (and combining two questions Morris presents separately) (p. 8):

To accommodate the Non-Distraction Thesis, we need to explain how it is possible for attending to the medium to be a way of attending to content, and vice versa. Morris's way to do that is to find an object of appreciative attention that combines features both of medium and of content – here, something like a face composed of paint. But what could that be, if as (Q.a) suggests, what is worked in paint cannot be a face? The answer lies in the idea of a real likeness.

What are real likenesses? In a slogan (not Morris's), they are real, likenesses, but not mere likenesses. They are real entities, before us when we appreciate representational works, and, at least in the case of paintings and photographs, things we really see. They are likenesses, in that they resemble the real world entities we mention when describing what the work represents. The face worked in paint, for example, resembles a face. But they are not mere likenesses. For the role of these resemblances is to underpin kind membership. In virtue of resembling a face, the face worked in paint is, in a way, a face. It is not a real face, of course: real faces are not made of paint. But it is a face nonetheless. As a toy soldier is not a real soldier, but is nonetheless a soldier, in a way; so the painted face is, in a way, a face (Ch.2, §5).

It is a familiar idea that representations, at least of some kinds, resemble what they represent. But the familiar move is to appeal to resemblance to bridge the divide between the real and the absent, between the representation before us and the thing, perhaps merely represented, that provides its content. That forces a distinction, between resembler and resembled, that at best promises little help with accommodating the Non-Distraction Thesis, and at worst threatens to make accommodation impossible. In using resemblance to underpin kind membership, Morris puts it to very different ends. Moreover, on the familiar idea the resembler is the representation itself: e.g. it is the painted canvas that resembles what the painting depicts. On Morris's view, in contrast, what does the resembling is not the work (after all, the painting is not, not even ‘in a way’, a face), but something ‘in’ the work, something grasped only by those who understand that work. As Morris puts it, the resemblance he appeals to is ‘deep’, not ‘surface’, resemblance.

What then, is this resembler; this real thing, distinct from the work, that, in virtue of resembling a face, counts as a face? Morris describes it as a set of elements in the medium, arranged by the artist, in such a way as to exhibit a certain kind of coherence, or unity. He captures that unity by saying that the elements exhibit ‘normative interdependence’:

In sum, the artist's job is to work the medium (and for Morris a medium is first and foremost something artists use to their ends), so as to create real, normatively governed unities. These unities resemble real things (such as faces), and, in virtue of that likeness, count, in a way, as instances of the kinds to which those real things belong. The appreciator's job is to direct her attention onto the real likenesses the artist creates. Those real likenesses combine medium and content: e.g. a painted face is composed out of paint, but counts as a face. Since they are the objects of appreciation, attending to the medium need not distract from attending to the content, or vice versa: we can accommodate Non-Distraction. We can also answer (Q.a). The real faces are not the only faces, and what we see in the painting is both a face and composed of paint. And if, mindful of my ‘complication’, we ask where in all this we pass from the real and present to the absent and possibly fictional, the answer is: nowhere. A real likeness of a face, though not a real face, is a face, and is real. Content, properly understood, does not reach out to the absent world of real faces and the like. As a result, it is as present to us when we engage with the representation as is the representation itself.

While this structure is common to painting, photography and the novel, working through the details requires careful attention to the distinctive features of each. Of particular interest are Morris's accounts of the medium in photography and novel writing. For photography, he draws on predictable elements such as framing, focus and depth of field. But their application, to give an account of the real likeness the photographer creates, a likeness that, for all its created character also reflects the influence of the photographic subject in determining the result, proves both original and insightful. When he turns to novels, his account of the medium is provocative and distinctive. It involves both the idea that the novel is, in some sense, made to be performed (read out – perhaps to oneself) and that such a performance will itself create ‘verbal mimes’ – oblique likenesses in words, of its characters and scenes. Throughout, Morris implements an ‘art first’ methodology, on which a proper understanding of the representational arts must begin with uses of the relevant forms of representation in art, not their deployment outside it. This is just one respect in which the book constitutes a refreshing rejection of orthodoxy in philosophical aesthetics.

For each art, the positive phase of Morris's discussion is preceded by a critical assessment of existing views. Generally these views have not been designed to make sense of Non-Distraction. However, Morris is able to connect them to his inquiry via the theme of paradox. Existing accounts of pictorial representation or picture perception can be read as offering various answers to (Q.a). There's a parallel question for photography (see below), which Morris uses as a lens through which to view accounts of the role in photography of causation, and the possible consequences for its aspirations as an art. And when he turns to the novel, Morris makes an exhilarating attempt to argue that the traditional Paradox of Fiction (how can we respond emotionally to characters we know not to exist?) is in effect a version of his more general paradox, in its manifestation for literature (how can we find in novels characters who seem to be people, when what we find there is essentially something written?).

Much of this is original and fascinating, some is richly suggestive, and its best ideas may be profound. The book would repay careful study and long digestion. Though my own engagement with it is only beginning, here are some initial questions and observations.

First, Morris's ‘art-first’ approach to these issues opens up new perspectives, but at what cost? If we define real likenesses in terms of the successful deployment of a medium to artistic effect, what are we to say about non-art pictures, photographs and prose? Seeing Hendrickje Stoffels in Rembrandt's portrait of her is in many ways continuous with seeing a house in a child's painting of one. Has the child managed to create a real likeness too? If not, whence the continuity? Of course, there are many, many differences between the two, together forming a chasm in the possibilities for appreciation on offer. But do any of these differences amount to the presence in the one case of a normatively governed, visible real entity, the like of which is wholly absent from the other? To say No invites the question of what normative interdependence really amounts to, if even a child can create work meeting the relevant norms. To say Yes is to postulate a difference in the objects the pictures put before us that experience does not clearly confirm. (Here it is worth noting that others who appeal to the idea of deep resemblance in analysing pictorial representations (e.g. Husserl, Lambert Wiesing and M.G.F. Martin) construe the resembler in perceptual, not normative, terms.) While at various points Morris wrestles with something like this problem, especially for photography and the novel, it is not clear his exertions succeed.

Second, while Morris's efforts to make contact with existing literature are admirable and often illuminating, sometimes they misfire. Consider his discussion of Walton's idea that photographs are transparent – that when we see a photograph, we literally see the object photographed. That discussion is motivated by a paradox lurking in a question analogous to (Q.a):

Walton's view threatens to dissolve any paradox here: since to see the photo is literally to see the originating object, the thing seen ‘in’ the photo is a real face, after all. If transparency solves the problem, we need not appeal to real likenesses to do so. Eager to avoid this result, Morris sets himself against transparency. But the dialectic here is more complex than he acknowledges. For one thing, since the object photographed is not worked in the medium of photography, justice has yet to be done to (ii). For another, Morris's own view dovetails neatly enough with Walton's, properly understood. Walton's claim is that in photography we see the subject indirectly by directly seeing something else. For Walton, that other thing is the photograph. But why should it not be Morris's real likeness? A combined view thus beckons: by directly seeing real likenesses in photographs, we see indirectly the objects that were before the lens. The real likeness isn't redundant, since only it meets both conditions (i) and (ii). And the appeal to transparency allows us to recognize an important continuity between art and non-art photos (thereby addressing part of my first question). For Walton makes the transparency claim as a way of accounting for ‘intimacy’ – our sense of closeness to real world objects and scenes when we look at photographs of them. Intimacy seems to hold for photographs of all kinds, be they art or not. The combined view can make sense of this continuity. All photos are transparent, and so all involve intimacy. It's just that for art photos we indirectly see the object before the lens by directly seeing a real likeness of it, whereas for non-art photos we do so by directly seeing the photo itself.

Finally, consider the role of resemblance in Morris's theory. It is supposed to underpin kind membership. But is resembling an F really a way of being an F? To decide, it would help to have examples of such grounding elsewhere, examples not too close to the very cases at issue. The only examples Morris offers are toy soldiers and their ilk (Ch. 2). These are certainly not works of art. And they are in one respect simpler than the art cases. For while paintings, photographs and novels contain real likenesses, toy soldiers and wooden horses are themselves such likenesses (p. 76). Even so, these examples remain uncomfortably close to the phenomena they are intended to illuminate. In all these cases we have something like representation, and that makes it hard not to feel that they all raise the same basic questions. All are artefacts designed to put before us entities that somehow relate to real world kinds, such as faces; without themselves instantiating those kinds in the usual way. Given this, it's unclear whether the analogy between painted faces and toy soldiers bolsters Morris's treatment of the former, or merely spreads the mystery.

There may be good answers to these questions. Even if there are not, Real Likenesses is an original and important contribution. It contains many interesting discussions and ideas there has not been space even to mention here. Its provocative, sometimes eccentric, but fascinating and carefully thought through proposals should help refocus debate in the philosophy of the representational arts onto what are perhaps its most challenging issues.