King’s College, London
In this paper, I discuss Jenefer Robinson’s personalist account of pictorial expression.  According to personalism, a picture possesses the expressive properties we attribute to it because we take it that someone expresses E in the work. Robinson’s particular strategy exploits the concept of an implied persona who ‘unifies’ and ‘specifies’ what is expressed.  Dominic Lopes challenges this view by attacking what he takes to be a flawed assumption motivating the personalist account: the priority of figure expression.  Once we acknowledge this flaw, he argues, there is no good reason to prefer personalism to an impersonalist theory. I will argue that Robinson qua personalist can pre-empt this strike by clarifying the relation between (a) what a picture expresses and (b) what is depicted as expressing in the picture. Rather than leading with the idea that personalism unifies or determines the meaning, I argue that impersonalism diminishes it and that, should it be accepted, we would be left with an impoverished view of what specific art pieces can mean.
‘‘NightHawks’ is melancholy’ (J1)
‘Hopper is melancholy’ (J2)
If Hopper feels sad, angry or nervous, Josephine might pick up on how he feels simply by looking at the expression on his face. When Hopper’s mental states are revealed in this way, his expression is said to be ‘transitive’. Some behaviours are intransitive, that is they have the outward appearance of a transitive expression (a smile) but there are no mental states expressed. An example would be a smiling zombie. Josephine sees the zombie with the look of a smile, thinks of something mentalese, but nothing is in fact conveyed.
Whether we can map this distinction onto instances of pictorial expression is controversial. Intuitions run in both directions. Some think that we see the “mind, sensibility and skills” of the painter in the work. Others think that the look of (J1) is parasitic on the public criteria or look of (J2), and so expression can be successfully tokened intransitively. 
Personalists argue that pictorial expression is always transitive, although they do not claim that all pictures are expressive in this sense. To give an indication of the kind of pictures that are considered expressive, ‘The Scream’, ‘The Raft of the Medusa’ and ‘Guernica’ are typically discussed as promising candidates. Meanwhile, scientific illustrations and maps seem to lack these expressive qualities. There are of course difficult cases, like children’s drawings or the work of mental patients, but I will just put these to one side for now.
Robinson, qua personalist, constructs her strong version of the view by conjoining two claims,
(Pa) viewers view or should view the expressive content as transitive, and
(Pb) the creation of expressive content should be understood as a transitive act of expression. 
Even if a personalist did commit to (Pa) and (Pb), (Pa) does not imply (Pb), and vice versa. However, as we shall see, Robinson’s account endorses both claims.
According to Robinson, pictorial expression is transitive. A picture expresses only if the artist expresses, by articulating her mental states (M) through the work: (Pb). For the sufficient conditions of expression to be met, a competent viewer must be able to pick up the expression: (Pa).  The artist articulates (M) through an implied persona, whose (i) expression is picked up by viewers and (ii) who is a psychological extension of the actual artist. This two-step process makes her view weaker than so-called transmission theories, according to which the actual artist’s emotions are said to be directly transmitted through the vehicle of the painting to a receiver.  The implied persona, functioning as an expresser, unifies and determines what is expressed.  In a discussion of Spiegelman’s ‘Maus’ , Robinson says:
Barring any reasons for scepticism about Spiegelman’s sympathies, it seems clear that the horror and dread expressed by the implied author of the picture reflect traits of compassion and distress that can safely be attributed to the artist himself. 
In this way, it is suggested that a psychological link ties implied artist to actual artists. Because implied artists are a construct of the actual artist’s mind, they are constrained by the actual artist’s own psychological profile and so contain “traces” of them when extended into the pictorial world.  She says ‘The Scream’ is,
full of repressed desires, melancholy, and angst, and this is of course an important side of Munch’s own personality. 
However, there is more to pictorial expression than merely depicting a figure in an expressive posture or with an expressive facial expression. Some figures may be transitively expressive, such as Munch’s howling figure in ‘The Scream’, but others may be intransitively expressive. Those depictions which merely look sad or happy, such as models in a Calvin Klein advert, or emoticons, are not really cases of pictorial expression. They are examples of technê: skilled depictions of emotion.  Instances of technê map to Zombie ‘expressions’ as they convey nothing, whereas transitive expressions map to (J2) since they convey the artist’s (M). The difference between the two apparently rests on how the artist has used what he paints to “articulate or individuate an emotion” often without knowing “what he will express until he has expressed it”.  For this reason, some depictions are and some are not also pictorial expressions. Pictograms and sentimental technê elements are “typically subordinate to and explained by the overall expression of emotions or emotional attitudes by the artist or his persona in the picture”, where the implied persona acts as a proxy for the mental states of the actual artist. 
4 Lopes’ challenge
Lopes develops one of the few sustained philosophical accounts of intransitive pictorial expression, in which he denies both (Pa) and (Pb). His argument against (Pa) is most germane to the analysis I develop here. 
Lopes identifies three ways a picture expresses emotions.  In figure expression (hereafter FE), such as occurs in Daumier’s ‘Fatherly Discipline’, the depiction of a toddler in a tantrum straightforwardly represents a toddler having a conniption.  In-scene expression (hereafter SE) elements of nature may be depicted expressively to correspond or clash with FE.
The shipwrecked, starving figures aboard Delacroix’s Raft of the Medusa express despair; the roiling sea in which they are set adrift expresses dumb, haughty malignance; and the tiny ship on the horizon that might signify safe harbour instead expresses blind indifference. 
Lopes insists he is not claiming that by FE and SE “expression is depiction”.  His view is that “by depicting a figure or scene as expressing the emotion” pictures express mimetically.  Mimetic expression may be bolstered or undermined by a third mode: design expression (hereafter DE). In DE, some brute fact about the way we correlate colours, shapes and textures with emotions links the “picture’s design or surface” with expressive qualities. For example, Mondrian’s complaint that curves are “too emotional” reveals “the trouble is with the curves themselves, not with anything that they depict”. 
Lopes argues that Robinson (qua personalist) has taken FE to be the paradigm case of pictorial expression and as a result, developed an erroneous genealogy of personalism in order to retrofit SE and DE to FE. He argues that, for personalists
[s]cene expression raises a missing person problem. Unless there can be expression in the absence of a being, to whom the expressed emotion is attributable, then either there is no scene expression or the being in question is one not depicted. 
But, he continues, there is no good independent reason to “attribute the emotion that is putatively expressed by a scene to some person who is not depicted” and so the “expression of emotion does not require that there be anyone to whom the emotion expressed is attributable”. 
Robinson’s crucial error according to Lopes, is to assume that viewers sustain an FE figure expression line of thought in the absence of figures.  But “once this assumption is dropped” he notes we “may adopt an impersonal theory […] a dog can smile when it is not happy (and so can zombies)”. 
5 An Error
Lopes’ aim is to discredit the motivation for personalism. But there is an error in his analysis. By confronting it, Robinson would neutralise an objection to arguments in support of (Pa) and (Pb).
Lopes’ error is to conflate his categories of expression with a picture’s expression of an emotion E. He says FE is, for personalists, the central manifestation of a transitive expression. But if the personalist is not motivated by the missing person problem, as described by Lopes, then his objection is neutralised. And, indeed, the personalist claim is not motivated by the missing person problem. Personalists do not argue that simply seeing depicted figures expressing emotions FE is sufficient for seeing a picture as being an act of expression. The flickr algorithm might compile rows of illustrated smiling faces, each showing those faces expressing happiness, yet I do not see the flickr webpage as an act of expression.
Instead of rejecting Lopes’ analysis, Robinson tacitly accepts the notion of the categories and develops her argument in response to them. In doing so, she accedes to his construal of the ‘missing person’. She argues that the missing person is ‘a persona’ of the artist and also an ‘internal spectator’ in the picture-world. By pointing out that ‘form and content’ are interdependent, Robinson argues that Lopes is wrong to claim there are brute facts about the expressive qualities of DE and that it comes apart from FE and SE.  She says,
Kokoshka’s Self Portrait […] shows the painter looking anxious and insecure (as in ‘figure expression’), but also conveys anxiety and insecurity in the agitated passages of paint, the awkward perspective and the vague, unstable use of space (design expression). 
There is an unresolved tension in Robinson’s reply to Lopes. The above quotation suggests that she thinks that DE is solely responsible for giving us the implied persona expression (FE is responsible for the figure ‘looking anxious’, DE for how the picture ‘conveys anxiety’). “Design expression” must refer to DE, since she denies that there “is a fourth species of pictorial expression, in addition to figure, scene and design expression”.  However, this contradicts a different response she makes to Lopes, namely, that: “Figure, scene and design ‘expression’ in and of themselves are only ‘expressive elements,’ not genuine artistic expressions of emotion”.  It may be that the categories are significant in the way we come to see the picture as expressive, but on the basis of her Kokoshka example alone, it is difficult to see how they relate to each other.
My main concern, however, is that Robinson leads with the idea that Kokoshka’s expression specifies and unifies what we see in the picture. I think that there is an alternative way to phrase the motivation for the personalist case, now that Lopes’ challenge to the personalist starting point has been neutralised. This alternative links to Robinson’s observations concerning the formal qualities of the picture as well as the things we see depicted in it, and brings out the different levels at which we understand a picture. For example, at first glance ‘NightHawks’ is a straightforward scene executed in a “light touch” noir-ish style. But given due attention, we can see a more serious or ‘deeper’ meaning in the picture. ‘NightHawks’ exhibits some stylistic features that seem incompatible with each other: for instance, the faces are rendered in a quite crude and mask-like way, while the ambience of the picture overall is sophisticated and self-assured. This can encourage the view that a naïve painterly style is being put to use by a mature, skilled artist. Walton suggests that this kind of multi-level impression of an implied painter operates at a “deeper level” than just taking it that the work is painted by a naïve illustrator and can lead theoretically to the viewer experiencing a ‘naïve illustrator’ in some paintings that “derives from an obscure partial awareness of a multi-level situation of this kind”. 
Since Lopes’ categories of expression can only account for meaning derived from the formal (visible) qualities of the picture surface, meaning is restricted to what goes on inside the represented pictorial world. Robinson can push the worry that this is insufficient to capture the full extent of expressive meaning. The appeal of (Pa), in contrast, is to be found in how it explains these additional levels of meaning, unavailable to interpreters working solely with Lopes’ categories. If so, then there are good reasons to consider the argument for implied expression.
An obvious counter from the impersonalist could involve denying that implied personae have any place in the interpretation of paintings. However, Robinson would appear to be on firm ground as (1) we need to account for more than the meaning of the visible formal properties of the picture, and (2) implied (narrator) personae are well established in the (relatively mature) philosophy of literature, where they are broadly understood as conveying intentions, beliefs or norms that do not belong to any of the characters in the novel.  Without a principled reason against extending (2) to pictures, Robinson is free to push for (Pa).
There is, however, the following objection to (2). Although some pictures, like some novels, give us the impression of the kind of person who created them, we should be cautious about how we extend the concept from narrative literary works to pictorial expression.
Walton notes that for those novels that have implied narrators, the narrator plays a crucial role because,
He mediates the reader’s access to the rest of the fictional world; we know what happens in the fictional world only from his reports about it. 
But since the way viewers represent the pictorial world can be much more direct, the mediating gaze of the implied painter does not seem as crucial. I judge that a (fictional) tragedy has occurred by being told by an (implied) narrator, “baby shoes for sale; never worn”. In contrast, I can judge (in the pictorial world) that there is a listless dynamic between the people in the bar from by how they appear directly to me. Although this deserves much more explanation and unpacking, it seems, from this initial observation, that the expression of an implied painter is going to play an incidental, rather than crucial, role in how the viewer gleans the expressive meaning of the picture. These considerations provide further support for my view that implied expression, wherever it may be found, is an additional level of expressive meaning that enriches viewer understanding, rather than a type of meaning that specifies and coheres what is expressed in the picture.
For reasons of space, I have simply noted this qualification on the arguments for (Pa). I now put the issue to one side since even if one can make out the case for this multi-level experience of (Pa), there is a more pressing issue that needs addressing, concerning the move from (Pa) to (Pb).
6 Moving from (Pa) to (Pb)
Having dealt with Lopes’ objection to the motivation for personalism, I would like to now turn to the way Robinson’s argument moves from (Pa) to (Pb). I note some difficulties with the move and also some reasons to be cautious about the way (2) extends to pictures.
Robinson proceeds in more or less the following way. In real life, we form reliable impressions of people by looking at how they dress, the way they talk and so on.  An extrovert may wear bright colours; a confident character may strut; a wit produces pithy humorous observations and it is these traits that necessarily and inevitably reveal character. Similarly, in reading a novel or looking at a picture we may form an impression of the kind of person who authored or painted it. Sensitive prose and elegant brushstrokes will reveal aspects of the artist’s personality which readers and viewers will pick up through some kind of epistemic seeing.
Even putting to one side the worries about whether we can reliably infer facts about, say, Josephine’s psychology from her preference for ditzy chintz skirts, and simply granting Robinson the inference in real life, it is still not at all clear how this transposes to the case of pictures. This is in part due to her two-step notion of transitive pictorial expression: the claim that viewers can infer facts about the artist’s psychology on the basis of a construct, the implied persona. For example, if it is the case that Hopper expresses his melancholy in ‘NightHawks’ then when the viewer views the melancholy as issuing from his implied persona (a construct) in the work in virtue of the implied persona that Hopper has constructed (hereafter Hopper’), the viewer views Hopper’s melancholy.
Robinson has told us that because Hopper’ is the creation of Hopper, Hopper’ is necessarily shaped by and linked to the personality of Hopper. So if one picks up on an expression of melancholy by Hopper’ we also pick up on Hopper’s melancholy. Robinson relies on this psychological link to move from (Pa) to (Pb).
This is tendentious as it implies painters cannot create implied personas that are psychologically distinct from themselves. A well-known counterexample to this, however, is Tolstoy’s ability to write with compassion while lacking compassion in his personal relationships. For this reason, it expects too much of impressions since it cannot be that viewers go from (Pa) (viewing the painting as an expression by Hopper’) to (Pb) (understanding the painting to be an expression by Hopper), since Hopper’ is merely a construct and so has no psychological reality. I cannot, when engaging with the ventriloquist’s doll, legitimately move from the impression of sentience to believing I am picking up psychological states of the doll that are somehow given reality by an extension of the ventriloquist’s states. So, it is not clear how viewers pick up on the emotions or traits of Hopper by constructing Hopper’. Even if viewers form the impression of an implied persona, since there is no entailment from (Pa) to (Pb), the fact that a viewer has the impression of an implied painter who expresses melancholy is not an argument for (Pb). The concern is that the distinction between the two claims that a personalist can make, (Pa) and (Pb), is being glossed over.
Robinson could argue that viewers infer a sincere connection between artists and their implied personae, by relying on extra-pictorial biographical information. The problem with this attempted solution, however, is that even if the viewer identified Hopper’ as Hopper, it would still be the case that the viewer was connecting with Hopper’ and not Hopper. Put in the counterfactual mood, the viewer would understand the picture even if they did not make the identification. So, the link between Hopper’ and Hopper is not conceptual and, as we have noted, we do not have grounds to allow that it is psychological.
For these reasons, one cannot accept that the expression of an implied persona is also the actual expression of the artist. This should, however, not discourage further reflection on the significance of our impressions of a painting’s origin. Arcing drips, bold palette knife work, and delicate glazes of paint may suggest a sensuous, agitated or careful personality in the style of the work. We can recognise this and seek an explanation for why the viewer chooses to represent the implied persona as if it were a psychological continuation of the actual artist. That is, we may question why viewers make this connection, without supposing that the connection reflects a constitutive tie between two distinct identities. A viewer who sees the painting as an expression by Hopper would be making a harmless transition from viewing the expressive content as transitive to seeing the creation of the expressive content as a transitive act of expression. On this story, Robinson’s move from (Pa) to (Pb) could be understood as a benign further claim, entertained but not entailed by the central issue. Meanwhile, implied expression can still be understood as a significant source of expressive meaning for paintings and, moreover, one that the impersonalist struggles to explain.
I have argued that impersonalism impoverishes our comprehension of some expressive paintings because it fails to fully reveal all the meaning in the work that only implied expression will explain. While this falls short of providing an argument for Robinson’s controversial claims about the actual painter, it provides a reason to accept (Pa). However, there remains a highly complex and unresolved issue: namely, whether, in picking up on implied expression, we pick up on the mental states of painters or their implied personas, or if we merely imagine that we do so.
About the Author Vanessa Brassey is a Ph.D. candidate at King’s College, London, where she works under the supervision of Sacha Golob and Derek Matravers, pursuing an investigation into the phenomenon of artistic expression. In addition, she collaborates regularly with the Centre for Philosophy and the Visual Arts, exploring the connections between philosophical theory and the artist’s practice through interviews, public seminars, live events and film.
Booth, Wayne C. (1983). The Rhetoric of Fiction, 2nd edition. Harmondsworth: Penguin.
Collingwood, Robin George. (1938). The Principles of Art. Oxford: Clarendon Press.
Davies, Stephen. (1994). Musical Meaning and Expression. Ithaca, N. Y.: Cornell University Press.
Gayford, Martin. (2010). Man With a Blue Scarf: on Sitting for a Portrait by Lucian Freud. London: Thames & Hudson.
Kivy, Peter. (1980). The Corded Shell: Reflections on Musical Expression. Princeton, NJ: Princeton University Press.
Levinson, Jerrold. (2006). “Musical Expressiveness as Hearability-as-Expression”. In: Contemporary Debates in Aesthetics and the Philosophy of Art, Kieran, Matthew (Ed.). Malden, MA; Oxford: Blackwell Pub, pp. 99-108.
Lopes, Dominic McIver. (2005). Sight and Sensibility: Evaluating Pictures. Oxford: Clarendon.
Robinson, Jenefer. (1985). “Style and Personality in the Literary Work”. The Philosophical Review 94, pp. 227-247.
—. (2005). Deeper Than Reason: Emotion and its Role in Literature, Music, and Art. Oxford: Oxford University Press.
—. (2007). “Expression and Expressiveness in Art”. Postgraduate Journal of Aesthetics 4(2), pp. 19-41.
—. (2017). “The Missing Person Found. Part I: Expressing Emotions in Pictures”. British Journal of Aesthetics 57, pp. 249–267.
Tolstoy, Leo. & Maude, Aylmer. (1962). What Is Art?: Essays on Art. London: Oxford University Press.
Vermazen, Bruce. (1986). “Expression as Expression”. Pacific Philosophical Quarterly 67, pp. 196.
Walton, Kendall. L. (1976). “Points of View in Narrative and Depictive Representation”. Noûs 10, pp. 49-61.
Wittgenstein, Ludwig. (1958). The Blue and Brown Books. Oxford: Basil Blackwell.
Wollheim, Richard. (1987). Painting as an Art. London: Thames & Hudson.
 Also referred to as a ‘persona-theory’ or ‘persona-view’, see Robinson 2005, 2007 and 2017.
 Robinson 2017, p. 260, p. 263.
 Lopes 2005, pp. 50-65.
 The term receives sustained discussion in Part 11 of Wittgenstein 1958. In the literature the terms ‘expressing’ and ‘being expressive of’ are used to express the same idea.
 Gayford 2010.
 Kivy 1980, pp. 67-68; Davies 1994; Lopes 2005.
 Theorists who claim (Pa) include Levinson 2006 and Vermazen 1986. Theorists who claim (Pb) include Robinson 2005 and Wollheim 1987.
 Robinson 2007, p. 36; Robinson 2005, p. 270; Robinson 2017.
 Collingwood 1938; Tolstoy 1962.
 Robinson 2017, p. 260, p. 263.
 Ibid., pp. 255-257.
 Ibid., p. 257.
 Robinson 1985, p .227.
 Robinson 2017, p. 11.
 Robinson 2005, pp. 2005-228.
 Ibid., p. 267.
 Robinson 2017, p. 263.
 Lopes 2005.
 Ibid., p. 57.
 Ibid., p. 53.
 Ibid., p.56.
 Ibid., p. 56.
 Ibid., p. 57.
 Ibid., p. 58 (my italics for emphasis).
 Ibid., p. 59.
 Ibid., p. 70.
 Robinson 2017, p. 261.
 Ibid., p. 262.
 Ibid., p. 265.
 Walton 1976, p. 52.
 Booth 1983.
 Walton 1976, p. 50.
 Robinson 1985.