On Particularity in Painting

I. Introduction

It is widely believed – by tourists and connoisseurs alike – that proper appreciation of paintings benefits from, and perhaps requires, experiencing the original, that being in the presence of precisely this entity matters. This assumes that paintings are particular or unique in a way underlying whatever value or interest we may attribute them. This article aims to examine that assumption by considering how particularity is realized within and by paintings. Focusing on the constitution and function of their fine-grained marks, I argue that the way paintings realize particularity distinguishes them from other entities, including other kinds of picture, and more specifically that their particularity is more complex than that of mechanically generated pictures.

I begin by considering how to approach this question. Having reviewed some initially plausible but insufficient strategies, I suggest using a contrast with mechanical depictions – that is, projection-based pictures – to understand paintings’ distinctively visual particularity. The third section therefore examines painting’s fine-grained marks, showing how these are ‘complex traces’ whose compound and multifactorial constitution distinguishes them from the ‘simple traces’ underlying mechanical depiction. The fourth assesses how the differences between each kind of mark bear on their representational function, arguing that painting’s complex traces cannot – unlike simple traces – generally be considered contentful. Based on this twofold contrast, the final section concludes that paintings’ particularity consists in a unique configuration of complex traces and the unique way these marks bear content, before showing how this goes beyond the particularity of mechanical depictions and generically conceived pictures.

II. The Question of Particularity

How do paintings realize particularity? What might require being in the presence of such an entity to experience it properly? This question is unorthodox, so I begin by indicating why it matters, before reviewing two initially plausible but insufficient approaches – treating paintings as objects or pictures – and introducing my own approach.

It is sometimes suggested – by Immanuel Kant and Theodor Adorno, for example – that aesthetic judgement or experience is singular, in attending to its object as a particular, and involves experiencing resistance to our cognitive powers. Although with Kant this resistance might be taken as a (‘subjective’) fact about us and transient (during reflective judgement), for Adorno an artwork’s particularity is precisely what reveals the limitations of our concepts and cognitive powers.¹ Without endorsing further details of these authors’ views, I think the distinctive interest – both cognitive and aesthetic – of paintings is at least partly due to their presence, and our experience of them, as particulars.² This dual interest motivates this paper’s question. Nevertheless, my concern is not why their particularity is valuable or pleasurable but the more limited and prior question of what such evaluations are directed to: what is it that makes paintings unique artefacts capable of sustaining attention to their details while both challenging and resisting our cognitive powers?

Although one might worry that anything particular will ipso facto be ineffable, it remains possible to consider which factors underlie such ineffable particularity and its resistance to description. To begin with, a painting shares two ways of being unique with other objects in being a spatio-temporal particular and having a unique causal history. These are the basis of familiar distinctions between an original/authentic work and forgeries, as well as Nelson Goodman’s related autographic/allographic distinction, which might initially seem to account for paintings’ particularity.³ They also play a key role – especially its causal history – in determining a painting’s cultural significance and financial value. These factors converge in Walter Benjamin’s notion of ‘aura’, which refers to an artwork’s ‘here and now’, ‘its unique existence’ as the material bearer of a determinate history, and is linked with its ‘ritual function’ or ‘cultic value’ in the context of social practices.⁴ However, while the fact that Johannes Vermeer executed The Art of Painting may explain its attracting museum visitors, it does not help in understanding this painting’s uniqueness as a visual artefact. Indeed, despite lacking the all-important causal link with Vermeer’s hand, any forgery of this work will likewise be unique. So, while paintings may be unique as originals, as autographic, and even as having an ‘aura’ in Benjamin’s sense, these aspects of uniqueness are insufficient for this paper’s purposes because they fail to take account of paintings’ distinctively visual character.

A more promising approach is to treat paintings generically as pictures, which immediately suggests several ways of conceiving their visual uniqueness. Clearly, each picture has a particular form, a certain arrangement of fine-grained marks that makes it the picture it is. Similarly, each picture plausibly has a particular content – namely, the unique visual scene it represents. It is also commonly assumed that the form and content of pictures are related by systematic covariation, such that any change in form entails a corresponding change in a picture’s content.⁵ Plausible as these claims may initially appear, I want to suggest that treating them as pictures again risks occluding important aspects of paintings’ distinctive visual character. More specifically, I argue that talking simply of ‘form’ or ‘content’ fails to capture adequately paintings’ particularity and that the assumption of systematic covariation does not hold – and ultimately proves unintelligible – for paintings.

To develop this argument, I will distinguish two modes of depiction. The first, mechanical depiction, is defined by a projective relation holding between the pictorial medium and whatever is depicted. Familiar examples of such depiction are provided by everyday photography in which a camera produces a projective or systematic rendering of a source scene in some target medium. This typically involves an optical lens projecting a scene onto some photosensitive surface, generating a record of the scene that is used to reproduce it in a photographic image.⁶ Mechanical depiction is clearly not limited to this familiar case and allows for different kinds of recording or rendering device (digital/analogue, mechanical/electronic) and different kinds of projective relation. Nonetheless, what defines mechanical depiction as a mode of depiction is not any specific technology or production processes but a two-term relation, between the picture and a scene, with features of the picture standing in a lawlike projective relation to features of what is depicted.⁷ With the second mode of depiction, freehand depiction, there is no strict correlation between the pictorial medium and what is depicted. I take this to be exemplified by painting, although it naturally includes all handmade pictures, such as drawings. Freehand depiction differs from mechanical depiction in that further factors (such as the painter’s body) are constitutively involved in the depiction process, so that depiction cannot be conceived simply as a two-term, projection-based relation.

This distinction, and my intended use of it, requires some clarification. First, one might worry that it relies on an overly simplistic ‘orthodox’ or ‘single stage’ conception of photography, whereas the ‘new theory’ of photography has shown that photographs are the end product of a multistage process.⁸ On this view, capturing visual information from a scene – in the ‘photographic event’,⁹ yielding a ‘register’ of ‘encoded information’¹⁰ – remains essential, but this is considered insufficient to explain the multiple ways such information might be rendered in a photographic image. However, whatever virtues a multistage conception of photography has, it does not affect my intended use of the mechanical/freehand distinction. First, as defined here, ‘mechanical depiction’ makes no claim about how many stages are involved in producing a picture (photograph), focusing instead on the relation – of lawlike projection – between a certain kind of picture and the depicted scene. Second, the new theory does not deny that such a relation can hold between photographs and scenes. In Wilson’s words, a ‘dependency relation’ between the photograph and depicted objects must be recognized as holding ‘for at least some photographs’,¹¹ while for Lopes images that ‘depict by belief-independent feature-tracking’ remain an ‘important subset of photographs’.¹² However, third, this concession suffices for my intended use of mechanical depictions as a contrast case that highlights important features of paintings. Even if only a subset – relatively few or aesthetically uninteresting – of photographs qualify as mechanical depictions, these can serve as examples for this purpose. In the following my attention is limited to this subset, which, to avoid misunderstanding, I will call ‘mundane’ photographs.¹³ Finally, to be clear, my aim is not to make general claims about photography, let alone photography as art: I assume simply that considering mechanical depiction, of which mundane photography provides some examples, provides a helpful contrast for identifying certain features of painting.¹⁴

Indeed, the aim of the freehand/mechanical distinction is not to define either painting or photography. It is therefore not undermined by counterexamples such as paintings that look like photographs (a trompe l’oeil, say) or photographs that look like paintings (any photograph of a painting). Rather, the distinction identifies two contrasting ways of relating marks and meaning, of which standard cases of painting and mundane photography provide clear examples. Nor is it problematic that hybrid cases are possible – such as ad hoc modification of photos (‘photoshopping’) or the use of ‘mechanical’ methods (such as central perspective) in painting. Rather, contrasting these two modes of depiction projects a space of possibilities in which depiction – including hybrid cases – occurs. Finally, this distinction is not simply being invoked to explain how paintings realize particularity. Rather, considering this question simultaneously involves – as a by-product – exploring and more fully explicating this distinction between two modes of depiction.

To what extent do claims about freehand depiction generalize to all paintings? After all, hybrid cases are not only possible but historically common, because painting has often used mechanical devices or methods to efface or overcome its freehand character. However, that this requires directed effort is significant in two ways. First, it highlights that painting is by default freehand depiction. While it is conceivable that a painting might be produced in a pervasively mechanical way, this would be an exceptional or idealized case. More typically, insofar as mechanical order is not deliberately imposed, paintings are characterized by the freedom of the hand – the absence of rule-governed projection. This is why paintings typically exemplify, but cannot be defined as, freehand depiction. Second, the need to impose mechanical order suggests that below some level of detail paintings typically retain their freehand character, so that the hallmarks of freehand depiction will be most clearly in evidence at the level of fine detail.

The following two sections therefore (literally) look more closely at these contrasting modes of depiction by first characterizing the kinds of marks basic to each and then considering how these marks bear content. Section V then shows how this contrast allows more adequate understanding of the way paintings realize particularity than is possible by treating them either as objects or as pictures.¹⁵

III. Complex and Simple Traces

Paintings are often designed to be viewed from a certain distance, rather than being observed too closely, on pain of missing their magic and failing to see what they represent. Similarly, depiction theories typically centre on the objective or representational content of pictures and higher-level formal configurations – such as outline or occlusion shapes – that allow such content to be seen.¹⁶ However, this focus on representational content misses much that is distinctive of paintings’ appearance, as highlighted by the practice of copying paintings: even if a painted copy reproduces every identifiable object in the original, each painting retains a distinctive look – visible differences that elude description in terms of ‘seeing in’ or ‘seeing as’.¹⁷ Against this top-down approach, particularity in paintings is therefore better understood through a bottom-up approach that starts with their fine-grained structure.

What are the ‘fine-grained’ marks that make up paintings? In the following I use the term ‘trace’ to describe painted marks for two reasons. First, in contrast to ‘brushstrokes’, it allows for various methods of applying or working paint (such as dripping, scraping, rubbing) by remaining neutral regarding the process or tool used. Second, more importantly, the term ‘trace’ implies that a mark owes its structure and significance to the process in which it was produced – and of which it is a trace. Accordingly, traces are inherently meaningful rather than mere forms or shapes, and understanding how they are meaningful involves considering the process they result from. This section describes how the traces making up paintings are ‘complex’ in both their form and their underlying constitution, before showing how they contrast with mechanical depiction’s ‘simple’ traces. Much of this section’s discussion comprises phenomenological observations or reminders, which draw attention to features that seem obvious based on our experience of paintings but whose implications and significance – as explored in the following two sections – are often overlooked when doing philosophy.¹⁸

Before continuing, I want to clarify the sense in which the following discussion is phenomenological, given two possible concerns. First, one might think it lacks the right relation to experience, because it is not presented, say, as first-personal, as based on experiential reports (for example, artists’ testimonies), or as presupposing phenomenological reduction. Nonetheless, this section’s characterization of painted marks is intended to be phenomenological in the minimal sense of being descriptively accountable to our pre-theoretical or ‘lived’ experience of paintings, and thus in identifying features that we recognize as producers or viewers of paintings. Further, in relying on this characterization, the following two sections’ discussions remain similarly grounded in lived experience. However, second, one might still suspect that this discussion lacks the right relation to the phenomenological tradition, given that it does not rely on terminology or authors from this tradition. This is deliberate, because it is intended to be independent of such debts and grounded in pre-theoretical experience. Moreover, the appeal to such experience is intended as a minimal commitment that is both shared by all positions properly described as ‘phenomenological’ and enables dialogue with positions that might be labelled ‘analytic’. Nonetheless, my proximity to some phenomenological authors is also deliberate and should be recognizable. Thus my minimal characterization of descriptive accountability parallels Martin Heidegger’s ‘formal’ definition of phenomenological method, while my talk of ‘lived experience’ echoes Maurice Merleau-Ponty.¹⁹ Indeed, I take the latter’s discussions of painting, embodiment and style to provide a theoretical framework open to the kind of particularity under consideration here.²⁰ However, Merleau-Ponty arguably lacks a phenomenology of the painted surface and its constituent marks, and in this respect the following discussion complements his views and can form part of a phenomenological conception of painting.

Individuating painted marks faces two significant difficulties. First, a finished painting’s visible structure is often not a legible record of the individual brushstrokes (and so on) that produced it. Not only are single brushstrokes often invisible but any given passage of a painting is typically reworked multiple times, so that merging or superposition yields compound marks comprising more than one ‘brushstroke’. This is a first sense in which painting’s traces/marks are ‘complex’. The second difficulty – particularly relevant in contrasting with mechanical depiction – is that painted marks are inscribed in an otherwise unstructured space rather than a pre-articulated spatial ‘grid’. That is, a painting is generated not by filling in a matrix of predetermined geometry and resolution with colour values but by introducing often highly irregular forms into an initially blank space. These two facts – the absence of simple marks and an underlying spatial articulation – together have the important implication that there is typically no single or privileged way of individuating the basic marks comprising a given painting.²¹ Thus a painting must in general be understood as consisting of a differentiated field of marks but having no basic articulation of forms. I will call this (common) feature of fine-grained painted marks their ‘ambiguity of form’.

Painted traces are also complex in that they result from and attest to a multifactorial constitution process. A complex trace – and its inherent meaning – thus remains indebted to several factors that generated it rather than being explained by a single factor. Here I discuss three factors – embodiment, materiality and processuality – that the production of painted forms depends on.

1. Embodiment. The production of painted forms depends first on the painter as an embodied agent and hence on various capacities linked with the human body’s physical constitution and a background of training and knowledge. These capacities include the painter’s mental states, particularly their understanding of what painting aims to achieve generally and in each given case. Of more immediate significance, however, is that painted marks are produced by, and are literally the traces of, physical gestures and bodily movements through which the paint is applied and worked. As James Elkins puts it, ‘[a] brushstroke is an exquisite record of the speed and force of the hand that made it, and if I think of the hand moving across the canvas – or better, if I just retrace it without thinking – I learn a great deal about what I see’.²² Elkins speaks as an experienced painter for whom such responses are second nature. But even the layperson can to some extent recognize painted marks as such traces and vicariously imagine the movements that generated them.

It might seem tempting to suggest that painted marks’ gestural character has deeper expressive significance. For Elkins it allows paintings to convey some ‘mood’ or ‘emotions’.²³ Nigel Wentworth sees painted gestures as expressing the painter’s holistic being-in-the-world,²⁴ while Paul Crowther attributes them a primordial evolutionary significance.²⁵ I do not endorse these more ambitious claims. The first ignores the fact that different moods and emotions may be experienced over the time a painting is produced, and that paintings result from a deliberate creative project rather than simply emanating from an underlying mood. The second similarly neglects the distinctiveness of painting as a practice – a delimited domain with its own determinants – that is no more necessarily explicable in terms of one’s life ‘as a whole’ than by supposedly underlying emotional states.²⁶ Finally, there is no reason to assume that gestures have the same significance for our differently enculturated bodies as for those of past or even prehistoric agents.

Nonetheless, it seems right that recognizing the traces of characteristic human movements can ground a certain kind of sensorimotor response to a painting.²⁷ Thus the style and arrangement of marks can give us some sense of the painter’s movements – as dynamic/frenzied or static/calm, say – and in this (limited) respect of how the painter was ‘feeling’. Although this sense is no doubt fallible,²⁸ the fact that we discern traces of a human body’s action – the intimation of movements our own bodies might mirror – partly explains why paintings often strike us as more expressive than mechanically produced pictures. Finally, it also seems right that this kind of sensorimotoric expressiveness varies in degree and is more characteristic of works in which the brushstrokes are salient and freely formed – works that are in Wölfflin’s sense ‘painterly’ and ‘open’ rather than ‘linear’ and ‘closed’.²⁹

2. Materiality. A second factor determining paintings’ complex traces are material constraints – the available colours, brushes, canvas, and so on. These are important to paintings’ particularity because they function as specific constraints on their production that are indefinitely fine-grained (perhaps continuous). Together with the imperfect control we have over our own bodies, this makes exact reproduction of any given constellation of painted forms a practical impossibility, guaranteeing that every painting is, at some level of resolution, a unique configuration of marks.³⁰ The materiality of painted marks also serves as a reminder that painting takes place ‘in the world’ rather than being the outward projection of mental acts occurring in the head. Moreover, as a highly deliberate act that typically aspires to produce a visually meaningful artefact, painting serves as a model for thought processes that are in direct engagement with and conditioned by the surrounding world. Thus it provides an example of thinking in action, essentially embedded or ‘scaffolded’ cognitive processes, and of mental activity as a messy, hands-on business. As Elkins nicely puts it, ‘[t]hinking in paint is thinking as paint’.³¹

Painted traces’ materiality grounds certain possibilities for realizing meaning in painting, for example Wollheim’s ‘twofoldness’, such that viewing a painting requires simultaneous attention to its (material) surface and to what we see ‘in’ its forms.³² However, seeing something ‘in’ a painting’s marks is not always possible, say when we look closely or at isolated parts of a painting, or when the marks’ form or texture renders them opaque. In such cases, the opacity of painted marks draws attention to the medium itself, a self-referential possibility realized more easily in painting than mundane photography – where the medium’s materiality is effaced to allow ‘seeing through’.³³ Nonetheless, its materiality constrains painting only in a loose or minimal sense. Although requiring that painterly forms be materially realized, it does little to define or restrict the space of formal possibilities painters operate in. Rather, as the history of art demonstrates, an indefinite and unforeseeable multitude of forms and styles remains possible – with each painter remaining free to develop an individual style and formal ‘vocabulary’.

3. Processuality. Because a trace is the trace of something, the residue of some process, it implicitly refers to a temporal horizon. Accordingly, a work’s manifold of painted marks testifies – as a third factor – to the fact that painting takes time and reveals something of its production process. For my purposes here, painting’s temporally extended production has four interesting implications.

First, it draws attention to the pervasiveness of the compositional intentions guiding a painting’s production and the various levels at which such intentions operate. The production of a painting is not explained by a single intention governing the overall process – although art historians and critics often assume such intentions. Rather, the production of every feature is guided by more concrete – localized and fleeting – intentions on the painter’s part.³⁴ However, that paintings are pervasively subject to compositional intentions – intentions to produce each feature in a certain way – does not mean that all the marks/traces are equally or uniformly intelligible. Some individual marks, taken in isolation, have a determinate representational function (stand for some visible feature). Other marks function only as part of a larger configuration – a potentially referring configuration – that can be seen ‘as’ something (for example, representing an object). Alternatively, they might be contributing to stylistic or atmospheric effects over the whole painting (patterning, colour balance). Finally, painted marks may have no representational function and be simply a decorative (harmony, balance) or contingent feature. Moreover, they may simultaneously have more than one of these roles. In sum, the compositional intentions at work in producing a painting vary in their focus or localization, from the holistic level to fine detail, so that individual painted marks can have different (representational) functions. I will call this feature their ambiguity of function.

A second implication of painting’s extended production is that the painter is in dialogue with the medium and the painting being enacted, responding continuously and in situ to the work in progress. Thus, changes in one part of the painting often motivate changes elsewhere, and later stages are conditioned by, reframe, and rework earlier stages of the work. Accordingly, compositional intentions are dynamic rather than static – responsive, not pre-planned – and will evolve throughout the dialogical process of the painting’s production. In addition to its embedded and scaffolded nature, this highlights the enactive or processual character of painting. No painting is fully planned in advance.³⁵ As Elkins explains, ‘[n]o painter knows what the picture will look like [… P]ainting is something that is worked out in the making, and the work and its maker exchange ideas and change one another.’³⁶ Put simply, in Merleau-Ponty’s words, in painting ‘“conception” cannot precede the “execution”’.³⁷

A third implication is that in painting articulated forms are built up gradually, such that detailed or determinate configurations arise from the schematic. Of course, a single brushstroke can have a particular shape that is irregular, irreproducible, and unique. But to depict something – an object, person, or scene – determinately requires more work, a process in which progressively more complex forms are developed. With portraiture, for example, a greater degree of detail is required to capture the physiognomy of a particular person than that of a generic face.³⁸

Finally, the temporally extended production process provides reason to think of a painting’s compositional, formal, or stylistic features as significant or meaningful. This does not mean there is some basic alphabet of inherently meaningful or expressive forms and colours – as Wassily Kandinsky, for example, imagined.³⁹ Rather, because the application of marks is pervasively guided by compositional intentions and intelligence, the overall pattern of painted marks comprises a record of the extended process of reflection/deliberation that produced them. Although this record may be imperfectly legible (owing to the ambiguity of form), we are entitled to ask about the internal ‘logic’ of a painting’s formal arrangement, why its manifold marks have been painted the way they are rather than otherwise.⁴⁰ In other words, that these marks are deliberately and reflectively produced suggests they might make sense – have a kind of meaning – in a way not necessarily connected with or explained by the representational content that we might see or interpret ‘in’ the work. Thus stylistic choices often reflect general assumptions about the way recognizable objects are conceived, how as opposed to what is being represented. For example, in Wölfflin’s ‘linear’ style depicted objects are clearly delimited from their surroundings, intimating their independence and primacy, whereas in his ‘painterly’ style depicted objects transition gradually into their surroundings, suggesting their interdependence and embeddedness within the whole.⁴¹ Similarly, Paul Cézanne’s use of ‘modulation’ – patterns of salient parallel brushstrokes – in portraying objects conveys their essential solidity, as opposed to the impressionists’ emphasis on transient atmospheric effects, without relying on (‘linear’) outlines. Of course, stylistic traits remain open to interpretation: there can be many ‘painterly’ depictions of any object, while similar stylistic choices may have different underlying motivations. However, despite this residual ambiguity, stylistic choices often have determinate motivations and are significant aspects of a painting’s overall composition.

The distinctiveness of painting and complex traces can be better appreciated by contrasting them with the simple traces making up mechanical depictions such as mundane photographs. The simplicity of these fine-grained marks has several aspects. First, their basic structure or form is unambiguous, because they are not overwritten or reworked in the production process, and are inscriptions in a space of predetermined geometry and resolution (determined by the photosensitive medium). A second aspect of their simplicity is that simple traces are due to a single factor – the depicted scene, as a chromatic and spatial configuration, to which they stand in a projective relation – and therefore lack various features due to complex traces’ multifactorial constitution. Thus, mechanical depictions’ simple traces have a uniform grain over the picture rather than being shaped by the human body’s actions, movements, or gestures. This makes them less ‘expressive’ (in the sense described above) than painting’s complex traces, which, as traces of embodied action, can speak to our sensorimotoric responsiveness. Further, simple traces are not a site of embedded thought and action, or direct engagement with the medium’s materiality, as paintings are. Indeed, in mechanical depiction there is no specific site of picture generation, as the structure of the ‘image’ is the same across the projection space (between scene and picture), hence inside and outside the camera. Finally, mechanical depictions are not shaped by a piecemeal, temporally extended process. Instead, the function of their marks – recording the projected scene’s structure – is defined and homogeneous (lacks ambiguity), with their forms being maximally determinate from the start, rather than built up gradually as painted forms are. There is also no dialogue with the medium and the nascent work, so that their structures are not shaped by compositional intentions at various levels that evolve throughout the production process. Consequently, in mechanical depiction – unlike painting – the patterning of fine-grained marks cannot be exploited as a significant design feature to express assumptions about how recognizable objects are conceived.⁴²

IV. Trace Structure and Content

Having described some differences between complex and simple traces, I now consider what these imply for the function of such marks. While paintings may be ‘meaningful’ in various ways, I focus here on their role as bearers of representational ‘content’, the cognitive meaning that can be taken up in or form part of thoughts. Given the difficulties of describing them with language, it is sometimes suggested that pictures have content that is ‘non-conceptual’: although they have content (depicted objects, symbolic or narrative meanings) properly captured by concepts, their fine-grained details lie beyond these concepts.⁴³ One may also suspect that particularity can be understood along these lines.⁴⁴ However, because the label ‘non-conceptual’ is ambiguous, depending on which feature of ‘conceptual’ content is negated, I focus on whether it makes sense to conceive fine-grained complex traces as bearing content.

In assessing whether some feature of a picture is contentful, two conditions can be used heuristically: (1) the feature at least potentially has a referent: it is configured so that it can stand for something else; (2) it presents its referent in some way or other, attributing it some kind of property. Admittedly, these two conditions may not be independent. To begin with, (1) seems more basic, because attributing any property presupposes identification of what it is being attributed to. Further, (1) arguably entails (2), because any configuration of pictorial marks that identify a referent presents it as having certain properties (such as shape). The role of each condition is also open to debate, because there are differing views of what content is: for instance, an externalist might think (1) sufficient, whereas an internalist requires (2). However, in the following I bracket these complications and consider the two conditions separately, assuming only that at least one of these must hold for the claim that some feature is contentful to make sense.

In describing these conditions as ‘heuristic’, my aim is to consider whether painted marks can be contentful independently of any specific conception of content. But why think this is possible, rather than a sleight of hand? The underlying thought is that these conditions capture what painted marks must do to make seeing something possible, so that the notion of pictorial content is unintelligible unless at least one condition holds. This thought can be illustrated and supported by considering several different kinds of picture. For figurative pictures both conditions clearly hold. In a realistic landscape, for example, configurations of pictorial marks might stand for a real tree, attributing it shape and colour properties that objectively resemble its own. Similarly, in a realistically painted fictive landscape, configurations lack an actual referent but still attribute properties that would objectively resemble features of the landscape (such as a tree) if they existed. Both conditions similarly apply for non-realistic but figurative pictures such as a caricature: the configuration allows its intended referent (a politician, say) to be recognized, even though the properties assigned are now distorted (non-lawlike) and schematic. Accordingly, each of these three figurative cases clearly has pictorial content: we can see something in or through their forms, on the basis of either objective or experienced resemblance.

Both conditions are also met by some abstract paintings. Kandinsky’s Compositions, for example, typically contain recognizably distinct formal configurations that assign colour and shape properties to any would-be referent. These paintings are less obviously bearers of pictorial content: we are not tempted to think they depict real objects and have no way of knowing how they assign properties to would-be referents (objective resemblance or caricature, say). But we might still be tempted to think of such works as contentful, even if it remains unclear what their content is, because, in satisfying (1) and (2), they do something akin to clear examples of pictorial content. In other cases of abstract painting, such as Jackson Pollock’s Lavender Mist, neither (1) nor (2) is clearly satisfied: there are no recognizably distinct formal configurations that might in principle have a referent, and it remains unclear how its marks might assign properties. Although such a painting may be artistically valuable or visually pleasing, to suggest it has pictorial content makes no sense: it allows nothing determinate to be seen and resembles nothing, precisely because it fails to satisfy either (1) or (2). In the following, these two conditions are applied to fine-grained marks rather than whole paintings. However, the challenge remains the same as with Lavender Mist: if we cannot identify potentially referring configurations of marks or discern how the constituent marks assign properties, then it remains unintelligible how these marks could be content-bearing.

To facilitate the contrast with painting, note that with mechanical depiction both conditions are met pervasively – throughout the picture and down to the level of fine-grained simple traces. This is obvious with a mundane photograph, each part of which stands for the corresponding part of a real scene and assigns it certain visual properties (colour, tone). However, fulfilment of these conditions is implied generally by the projective relation between simple traces and the depicted scene. This relation has several implications. First, identifying potentially contentful features presents no difficulty, because a lawlike projection implies that any given segment – no matter what shape or how small – of a mechanical depiction is contentful. Second, mechanical depiction is homogeneous: the representational function of each point in the picture is the same (as defined by the underlying projection). Third, because a projection rule determines marks point by point, each point relates to the source scene independently of surrounding marks, so that fine-grained details are, so to speak, atomic. Fourth, the content of mechanical depictions – just as their form – is maximally determinate from the start. While the projection method might yield regions of low focus (indeterminacy), the degree of detail is determined once and for all by the underlying projection rather than developing gradually during production. Finally, mechanical depiction ensures systematic covariation of form and content, because the pervasiveness and homogeneity of content – both implied by the assumed role of projection – entail that any change in form is equivalent to a change in content. In summary, the content of mechanical depictions is pervasive, homogeneous, atomic, maximally determinate, and entails systematic covariation.

By contrast, the above conditions for contentfulness are not always met by paintings’ complex traces owing to their ambiguity of form and function. The ambiguity of form is that such fine-grained marks are often difficult to identify or parse. Yet, if these marks cannot be unequivocally identified, it remains correspondingly unclear how to assign potential referents to a painting’s fine-grained structure. Put simply, it becomes impossible to determine what is supposed to stand for what. This difficulty is illustrated well by Elkins’s example of the ‘contorno’, ‘a dance or cascade of contours that work to form a single line […] without ever fully succeeding in erasing their own multiplicity’: contorni are marks that deny the viewer ‘any unambiguous form that can carry a determinate meaning’.⁴⁵

Even where complex traces have a determinate formal identity, the ambiguity of function remains. As highlighted above, while some marks have a determinate representational role (for example, marking some outline or coloured surface), others function only as part of a larger configuration or by contributing to overall style, and may have no representational function. One implication is that – in contrast to mechanical depiction – the function of fine-grained marks cannot be assumed homogeneous over the painting. Further, the function of individual marks is often not independent of surrounding marks but distributed over larger configuration – it is, so to speak, either ‘molecular’ or ‘holistic’ rather than ‘atomic’. A further difference to mechanical depiction is that in painting content determinacy is a late achievement. For, as described above, to represent something particular in painting, fine-grained detail must be built up progressively to attribute determinate properties to the depicted object. This highlights how determinacy of form and determinacy of content can come apart in freehand depiction, as a single brushstroke may be highly particular in shape (beyond geometry and description, say) while completely indeterminate representationally. By contrast, in mechanical depiction, such as mundane photography, fine-grained determinacy of both form and content is inseparable and maximized by default.

Owing to the ambiguity of form and function, many fine-grained marks in paintings fail to meet one or both of the two conditions. Their formal ambiguities mean we often cannot determine which individual marks or configurations are supposed to be reference-/content-bearing (ad (1)). Their functional ambiguities mean we often cannot determine how these marks present their would-be referents (ad (2)). Consequently, it will often be fundamentally unclear how to read the fine-grained marks of a painting as contentful.⁴⁶ This is an important result: their failure (often) to meet one or both of these conditions means that paintings’ fine-grained marks cannot be taken generally to meet the conditions for contentfulness.

This result has two further implications that unpack more fully the contrast with mechanical depiction. First, it no longer makes sense to think of paintings as pervasively contentful. Not only do many marks not meet the conditions for contentfulness but often it is unclear how to apply such conditions. Hence the distribution of content – if we continue using this term – over the fine-grained marks of paintings must be assumed discontinuous and non-uniform. Consequently, second, it no longer makes sense to apply the idea of systematic covariation between form and content to paintings generally. This idea does apply for mechanical depiction owing to the way content is realized (as pervasive, homogeneous, atomic, determinate). In painting, however, the ambiguity of many fine-grained marks, along with the discontinuous distribution of content over a typical painting’s surface, means there is no general way of identifying changes that directly and continuously affect the content of its manifold of marks. Hence, there is no simple correlation between variations in the complex traces of paintings and any content these marks bear.

V. Particularity in Painting

As argued in Section II, the question ‘how do paintings realize particularity?’ can be approached and refined in several complementary ways. Paintings can first be taken as particular objects – as spatio-temporal entities with a determinate causal history. They can also be considered as particular pictures with a one-off configuration of forms and a correspondingly unique content. However, both these approaches, I suggested, fail to reflect paintings’ distinctive visual character. While this seems obvious in the first case, the last two sections have articulated this visual character by contrasting the constitution and function of fine-grained marks in paintings and mechanical depiction. I now consider what this contrast implies for the distinctive way paintings realize particularity and what is missed by conceiving paintings as pictures.

The first distinctive aspect of their particularity is that paintings are made up of a unique configuration of the marks I called complex traces. While this might appear to be captured by talking of unique form, painted marks – as argued in Section III – are often ambiguous or indeterminate, so that paintings typically have no basic articulation of forms. Hence a subtle but significant difference: because the notion of form abstracts from marks’ materiality, a painting’s ‘form’ is not univocally determined by its configuration of complex traces – its unique set of marks does not entail uniqueness of form. The resultant challenge is: how to parse a painted surface; how to identify and group its constituent marks. No such difficulty arises for mechanical depiction: although perhaps difficult to discern in practice (say, too fine-grained), identifying simple traces involves no principled difficulty because such marks are unambiguous, independent (atomic), and uniformly distributed over the picture.

However, painted marks’ singularity goes beyond a unique configuration or arrangement. For complex traces are both the site and record of a unique interaction between an embodied agent, a material artefact, and an extended process of production. As a configuration of such marks, paintings are constitutively and pervasively singular. This contrasts with mechanical depiction, in which the patterning of marks (simple traces) is due to the depicted scene rather than the crafting of these marks. Whereas in painting the particular configuration of marks depends essentially on working the surface, with mechanical depiction the surface becomes inessential – it becomes notionally ‘transparent’, a contingent screen onto which a ‘scene’ is projected.⁴⁷ The particularity of painted marks is compounded because there is no guarantee of – and no need for – complex traces being produced according to the same method or principle over the whole painted surface. This (potential) inhomogeneity, together with complex traces’ multifactorial constitution, ensures that the painted surface cannot be accounted for by an algorithm or a formula. Whereas in mechanical depiction we might in principle specify how a pattern of marks is generated – say, by a defined projection method, viewpoint, and target scene – there is no shorthand way of specifying how to make a painting, no description that abbreviates its step-by-step production.

The second distinctive aspect of paintings’ particularity is the way painted marks bear content. That complex traces do not pervasively meet the conditions for contentfulness (Section IV) does not mean paintings cannot be content-bearing. Rather, it implies that each painting realizes its content in a unique way. Paintings therefore present a second challenge of recognizing how marks are used to realize content. As Section IV showed, no such challenge arises in reading mechanical depiction’s fine-grained marks as – pervasively, homogeneously, and atomically – contentful. However, because its realization of content is more complex (multifactorial) and less pervasive, freehand depiction requires hermeneutic engagement at the fine-grained level: the question arises for each painting, indeed each part of each painting, of whether and how it might be contentful. Consequently, rather than assuming systematic covariation, the relation between form and content must be worked out piecemeal. Paintings thus have a further aspect of particularity that mechanical depictions lack, as the one-offness of form and content is compounded by a one-off interplay between each painting’s constituent complex traces and any content these realize. Each painting is both the site and the record (trace) of such an interplay and the multifactorial process in which it was enacted.

The distinctively visual particularity realized by paintings thus consists in (1) their uniquely configured complex traces and (2) the unique way these marks bear content. This particularity has a more complex constitution than that of mechanical depictions, as shown by the two challenges that paintings present: parsing and interpreting their constituent fine-grained marks and recognizing how these bear content. Moreover, this particularity is not adequately captured by treating paintings generically as pictures with a unique form and content. This is partly because the latter notions gloss over the differences between freehand and mechanical depiction identified in the two preceding sections. But it is also, as this section has shown, because (1) and (2) go beyond uniqueness of form and content: the notion of form obscures the ontologically rooted particularity of complex traces, while each painting not only has a particular content but also a unique way of realizing it. Conceiving paintings in terms of form and content (as ‘pictures’) effectively assimilates them to mechanical depictions.⁴⁸ For in mechanical depiction the notion of form proves adequate because the surface is not the site of image production (it is notionally ‘transparent’), as does that of content because simple traces are pervasively and homogeneously contentful (there is no slippage, so to speak, between marks and content).

There are two reasons to think the view presented here provides an understanding of paintings’ particularity that is not only more revealing than the alternatives but also adequate. First, both aspects of their particularity are familiar – as phenomenological reminders – from our experience with paintings. Often we look closely at painted marks, independently of their content-bearing function, perhaps attending to the painter’s skill in crafting such marks, perhaps wondering how they are made⁴⁹ or (as art historians, say) how to describe them. Similarly, we often find ourselves wondering how these marks make sense – which objects, figures, features, relations, and so on they represent. We are also aware by default that viewing paintings differs qualitatively from both normal perceptual experience and viewing mundane photographs. For, even when this is smoothly negotiated at a pre-personal and pre-reflective level, we face the dual challenge identified here of parsing and interpreting fine-grained painted marks. Second, these two aspects of their particularity give paintings a distinctive kind of interest – and sustain the experience Kant and Adorno identify as aesthetic. Together they account for our experience of paintings’ singularity as visual artefacts, while the need for hermeneutic engagement with the painted surface – parsing and making sense of its configured marks – partly explains the ambivalent cognitive experience paintings afford. Even where the representational function of its fine-grained forms is not immediately obvious, viewing a painting engages our cognitive capacities in an attempt to discern their visual content.⁵⁰

Finally, why does understanding how paintings realize particularity matter philosophically? Although these thoughts cannot be further developed or defended here, I want to finish by noting three ways – implicit in the preceding discussion – that this is potentially of philosophical interest. First, as noted in Section II, it is relevant – as a prior question – to the aesthetic interest of paintings, and perhaps extends to other comparably singular artefacts (such as sculpture). Second, Sections III and IV suggest certain challenges for philosophical theories of depiction. For example, the contrast between mechanical and freehand depiction – particularly the differences in their fine-grained marks’ ontology and function – limits and potentially challenges the idea that pictures are a unified kind. Similarly, the limited applicability of the notion of content to complex traces challenges the idea that a theory of content – including ‘non-conceptual’ content – might apply to the fine-grained marks of all pictures. Third, the way paintings realize particularity suggests a distinctive kind of cognitive feat, bearing the hallmarks – as Section III shows – of embodiment, embedment (materiality), and enaction (processuality). Its complex traces thus reflect a correspondingly complex act of production, a temporally extended process of reflection/deliberation, characterized by nested and evolving intentions and dialogue with the work being produced. Further, the singularity of this act and of paintings as concrete artefacts have no analogue in standard views of cognition, which conceive of concepts and thought as (language-like) syntactic and semantic types. The preceding discussion of how particularity is realized by paintings may therefore also be seen as a step towards understanding and assessing painting’s potential cognitive significance.