Klaus Speidel

Can a Single Still Picture Tell a Story?

Definitions of Narrative and the Alleged Problem of Time with Single Still Pictures1

That the same story can be told in different media is one of the fundamental claims of narratology. Claude Bremond (1964) famously listed verbal narrative, novels, theater, movies and ballet among potential vehicles for story. He thus prepared the ground for narratology’s future as a discipline engaged in narrative research across media, in principle including single still pictures. However, narratological research concerned with pictorial narrativity generally proceeds from the assumption that although single pictures may evoke or imply stories, they are unsuitable for storytelling in a strict narratological sense. Focusing on the key issue of temporality, this essay will show that a single still picture may indeed tell “a story proper” (Wolf 2003, 180) and can thus be regarded as a narrative, even according to a narrow definition.

1. Still single pictures: Narrativity without narrative?

Mieke Bal is probably the narratologist who has most consistently analyzed single still pictures as narratives. But the fact is that the single still pictures of which she offered narrative analyses (e.g. Bal 1991, 1994, 1997) are quite incompatible with most definitions of narrative and arguably even her own (cf. Bal 1997, 9). While Bal’s work convincingly shows that the tools of narratology yield rich interpretations of pictures, she focuses mainly on ideology, and does not take up the challenge of demonstrating that pictures can indeed be narratives. We must keep in mind, however, that there are artefacts that are not narratives, but still have a high degree of narrativity (cf. Prince 2012, 26). An artefact which, for example, complied only with five out of six necessary criteria for narrative, would not be a narrative. However, it would still have a much higher degree of narrativity than an artefact which only complied with, say, one criterion. As long as we have not systematically established the compatibility of single still pictures with a plausible definition of narrative, it might always turn out that no picture of this kind complies with all criteria of a minimal narrative.

1.1 The practice of narrating pictures

But how about the tradition of “narrating pictures” (Dieterle 1988)? Doesn’t the ‘narrative reception’ of some pictures show that they are narratives? Dieterle’s work, in which he analyzes different cases where a clear relationship between a narrative and a specific picture can be established, seems to show the opposite. The beholders discussed by Dieterle go far beyond anything that could be called the activation of stories contained in pictures. As Leitch (1986, 40) has pointed out, a “hyperactive audience” can turn any utterance into a tellable story. But when this happens it shows us little about the utterance itself. If Diderot tells a story which is supposed to be motivated by a landscape painting by Vernet where no events can be identified, the motivation for storytelling is external, not internal, to the picture. It tells us more about Diderot than about the picture itself. Commenting on this fact, Dieterle himself compares the kind of narrative readings he is interested in to Rorschach tests (cf. Dieterle 1988, 135).

The fact that a story is sometimes based on a picture does not show that the picture is in and by itself a narrative. But how about pictures which were intended to tell a story and seemed to do so to their creators and their initial audience – for instance, history paintings? Aren’t they narratives? After all they are consistently and regularly received as narratives. The literature on history painting has been growing ever since Alberti first defined painting as a window through which we can see a story (historia or istoria) in his treatise On Painting (1435) and we cannot take it into account here. But I think that we don’t need to do so. The fact is that the historia evoked by a picture was usually mythological, biblical, or, in some cases, linked to famous historical events. These references were known by the intended audience and could be easily identified. If an artist provided the basic story-triggers, beholders easily agreed the story was being (re)told to them when they looked at history paintings. A naked woman, a naked man, an apple and a snake were enough to evoke the story of the Fall of Man. If all elements were there it could easily seem as if the picture told the story. Thus it was rarely noticed that many pictures of Adam and Eve make it look like Eve was feeding the snake, which implicates that no clear narrative is being conveyed by the picture alone. The assumptions of the artists and their initial audience thus structured the perception of the works as narratives. The cultural heritage completed the experience. In any case, artists and scholars only discussed how a story could best be conveyed by a picture, taking for granted the notion that pictures could be narratives until Lessing first criticized this idea in 1766.

1.2 Narrative in art history, Bildwissenschaften and Visual Studies

The treatment of narrative in art history seems to have been strongly influenced by the ideas of Alberti and other early art theorists. In his groundbreaking study on narrative in Caravaggio, published in 2011, the art historian Lorenzo Pericolo thus makes a bold statement:

In art history there is no narratology worthy of the name. It might sound absurd that a humanistic discipline like art history, which has mostly emerged from the early modern debates about the istoria, has stubbornly continued to appraise visual narratives with criteria that are frankly obsolete. (Pericolo 2011, 94)

This strongly resembles the point of view Wendy Steiner expressed twenty years earlier, when she realized that “the general art-historical use of the term ‘narrative’ seems incomprehensible to literary scholars” (1988, 2). While scholars like Felix Thürlemann (1990) and Wolfgang Kemp (1996a, 1996b) offer interesting assessments of the narrative dimension of single still pictures well before the publication of Pericolo’s study, no scholar fundamentally questions the possibility of telling a story with a picture.2

With their more foundational approach to how pictures work and looking for their universal essence the German tradition of visual studies (Bildwissenschaften) might seem to be a better place to look for the kind of argument which is needed here. In a recent text, which introduces Gottfried Boehm’s approach to a French readership, he explains why an episteme of pictures does not yet exist. The fact that there is “a logic of pictures”, Boehm argues, was masked by the intuitive idea that pictures are transparent. If a picture is like a window, there is little to be analyzed. According to Boehm, we must therefore break free of this association in order to understand the deep nature of pictures. To do so, we must strip pictures of everything they have traditionally been taken to communicate, such as stories (cf. Boehm 2010, 29f.). The fact that stories can be communicated in different media and thus cannot be essential for pictures might well be the reason why Bildwissenschaften generally ignore the narrative dimension of pictures. Pericolo also identifies a more general prejudice concerning this dimension of pictures:

The structures and mechanisms of pictorial narrative do not appear to offer enough ground for analysis and interpretation. That is, examining the means by which an artist narrates through an image does not seem to affect the exegesis of the artwork’s ‘content’, the only element that lends itself to being easily decoded in terms of ideology, culture, society, and history. (Pericolo 2011, 3)

A strong preference for “decoding ideology” can also be detected in the Visual Studies and is particularly obvious in W. J. T. Mitchell’s treatment of Lessing’s Laocoon, (Mitchell 1984), where he provides a list of no less than ten distinctions between poetry and painting to be found in Laocoon, but omits the fundamental distinction between actions told and bodies described. Mitchell is more interested in ideology than narrative analysis or semiotics.3

A recent work on narrative (Altman 2008) exemplifies a third approach to pictures as narratives which appears to be promising: Altman proposes a definition of narrative that has been specifically devised for the single still picture. Introducing his analysis of pictures by Brueghel, Altman writes: “only by stretching our definition can we assure coverage of all types of narrative” (ibid., 198). Rather than trying to demonstrate that pictures can comply with the definition he uses for literary text, Altman presupposes that pictures are narratives. He then adapts his own initial definition by abandoning the criterion of recurrence, a criterion deeply linked to the unfolding of a story in time that seemed essential in the first part of his work.4

But this is problematic. If recurrence really was not a necessary criterion of narrative – which it might or might not be – why did Altman introduce it in the first place? Ultimately, Altman’s approach offers an unsatisfactory solution to the problem we are discussing: while holding on to the idea that pictures can be narratives, the definition of narrative he applies to them is not quite as strict as the one he applies to texts. This introduction of a double standard of narrative makes it seem like a picture could not comply with the real standard. It thus plays into the hands of the skeptic.

I will here try to work with a different approach. I will accept one of the most narrow definitions of narrative and try to show that some pictures can comply with it. As far as I can see Wendy Steiner (1988), Aron Kibédi Varga (1989, 1990), Werner Wolf (2002, 2003, 2011) and Jean-Marie Schaeffer (2001) came closest to formulating and answering the major problems anyone who wants to establish the possibility of pictorial narrative has to solve. They did not presuppose that pictures can be narratives, but sought to establish whether pictures could comply with the criteria of a plausibly narrow definition of narrative. However, Wolf (2003, 96), Kibédi Varga (1989, 96) and Schaeffer (2001) stop short of recognizing that single still pictures can tell stories. For Wolf (2003, 96), for example, only a series of still pictures can tell a story [it can be genuin narrativ], whereas a single still picture can at best “induce narration” [narrationsindizierend]. Steiner (2004, 168) comes to the conclusion that a painting showing different moments in time can be “a strong narrative”, but does not provide an answer to the objection that pictures do not sufficiently determine the order in which represented events are perceived (challenge 2b, chapter 3, below).5 As this short overview indicates, it still remains to be firmly established that pictures can tell stories. To do so is the aim of this essay.6

2. Narrative and the single still picture: arguments in favour of a narrow definition

While it is true that terms such as discourse, telling or Erzählung naturally evoke verbal forms,7 they are also sometimes used in a media neutral way, and it seems quite natural to say: “This painting tells the story of Adam and Eve”. However, there is disagreement as to what exactly this means. While most theorists accept that it is possible to transmit a story through different media channels, there is widespread disagreement about exactly which media have the potential to do so. On the question of whether every transmission of a story counts as a narrative, Gérard Genette, for instance, favors “a narrow definition of ‘narrative’: haplè diégésis, an exposition of facts by a narrator who verbally signifies the facts (in written or oral form)” (1983, 24).8 In this sense, there is no such thing as theatrical or filmic narration. In direct contrast to Genette, André Gaudreault holds that “any message through which a story, whatever it may be, is communicated, should rightly be considered a narrative” (1988, 84).

These disagreements between different narratologists might appear like no more than quarrels about words, but they are much more than this. They constitute a discussion about what exactly the correct definition of narrative is. This problem has kept narratologists busy for many years. Most definitions of narrative thus far have turned out to be either too broad or too narrow, the most minimal ones including recipes in the realm of narratives, others excluding movies. Many narratologists fear that the concept of narrative could lose its meaning entirely if it becomes too broad. “By narrowing the scope of ‘narrative’,” Rimmon-Kenan writes, “I am trying to defend the term against being emptied of all semantic content: if everything is narrative, nothing is” (2006, 17).

2.1 The fundamental arguments against the possibility to tell a story in one picture

It is clear that if single still pictures cannot tell stories, there is no reason why narratology should be interested in them. The fact that there is no narratology of still pictures would only be a problem if it turned out to be possible to tell ‘a story proper’ with a single still picture. After all, nobody regrets that there is no narratology of soft drinks, shirt patterns, or Swiss cheese. However, if single still pictures can tell stories in principle, then narratological analysis of pictures should be much more frequent than it currently is. In generalized form, the arguments against pictorial narrativity are:

  1. A picture shows. It does not tell. In other words: the non-verbal nature of a picture makes it impossible to tell anything with a picture. A more sophisticated version of this view claims that pictures cannot present definite propositions. As several ‘narrative propositions’ are necessary for a story, a picture cannot tell a story.

  2. A picture is specifically a-temporal in nature. A picture doesn’t determine the order of telling (discourse) of what it shows, either by convention (like books) or presentation (like oral tellings or movies). Each spectator sees the events represented in a picture in a different order. Moreover, a picture does not clearly prescribe the order of that which is being told (story). It does not show what happened first, second, third, etc. in the story world. Thus, as a dual structure of time is necessary for a narrative, a picture cannot tell a story.

Given the fact that films and comics have enjoyed much narratological interest recently, despite the fact that both media do not necessarily present propositions, I will not focus on the first problem here. Instead, I would like to analyze the different roles which have been attributed to time in definitions of narrative. For each of them, I hope to be able to show that the criteria based on time should either not be part of a definition of narrative, or that they do not actually exclude pictures from the realm of narratives. As we will see, many theorists work with a simplistic vision of picture perception and underestimate the potential of pictures to express temporal relationships. I will not question the idea that story-time is needed. Moreover, for the sake of the argument, I will mostly accept the need for a ‘temporal program’ to be inscribed in presentations.9 While I will sometimes evoke cognitivist arguments, my demonstration would have been possible before the ‘cognitive turn’ in narratology took place. My argument is therefore different from those championed by Werner Wolf (2002, 2003, 2011), or Michael Ranta (2011).

While I am generally skeptical of the need for a strict definition of narrativehood, I think that taking the time problem seriously can teach us much about pictorial storytelling. The time problem is one of the oldest and most serious reasons for refusing pictures admission into the realm of narratives. By explicitly resolving this problem, I hope to convince narratologists that the study of pictures can be just as much a part of narratology as the study of short stories.

2.2 Narrative: definitions and intuitions

Criticizing the way ‘narrative’ has been used since the narrative turn took place, Shlomith Rimmon-Kenan writes: “Today, narratives are detected in film, drama, opera, music, and the visual arts. The last two are neither verbal nor governed by double temporality nor, strictly speaking, even by events” (2006, 10). In her final definition, she replaces verbal narration with transmitting agency “to account for phenomena like film” but maintains that double temporality and events are essential criteria for narrative: “narratives are governed by a dual time-scheme owing to the ontological gap between the succession of signs and the temporality of the events (in whatever expanded definition)” (ibid., 16).

Like Rimmon-Kenan, many narratologists have replaced the idea of a narrator with a more general notion of narration (cf. Bordwell 1985, 33) or narrative agency, which need not necessarily be a person or a representation of a person. This shift accounts for our intuition that movies tell stories, although we also feel that they lack an explicit or implicit narrator. Given this broader vision of narration, double-temporality and propositional content linked to verbal form sometimes appear to be the last ramparts defending narrative from being ‘flooded’with productions in non-verbal media. As we already saw, many narratologists worry that the concept of narrative might be watered down to a point where it would become nearly meaningless.

There are several reasons why I think this fear is unjustified. But the main one is this: it underestimates peoples’ capacity to intuitively distinguish stories from non-stories. While for the past many years certain definitions of narrative offered by narratologists would have classified cooking recipes or weather forecasts as narratives, the layman never mistakenly believed that either of these were narratives. Our intuitions are sufficient to sort most artefacts into stories and non-stories. It is easy for any competent speaker of a language to determine that Cinderella tells a story and that the weather forecast does not, that Pretty Woman tells a story while most recipes do not (cf. Prince 1999, 43). In other words: definitions need to stand the test of intuitions, not the opposite. The problem is usually with definitions, not intuitions. In fact, it is because some definitions yield counterintuitive classifications that they have been criticized.

In order to find out which texts people naturally think ‘tell a story’ Françoise Revaz (1997) thus uses focus groups. Not surprisingly the subjects of her groups agree that neither the weather forecast nor a recipe tells a story. As a result, she introduces a difference between the larger category of action texts (textes d’action) and the much narrower category of narratives (récits). While recipes describe actions and are therefore action texts, this does not make them narratives, mainly because they lack basic plot structure with complication and resolution (cf. ibid., 8f.).

I believe that theorists who defend a position which seems to fly in the face of common sense or intuition should be able to explain how people could have been so wrong for so long. In other words: they should be able to offer what philosophers call an error theory. Such a theory tries to show “that, though our position is mistaken, our error was nevertheless an understandable one to have made given the true facts of the matter” (Baggini / Fosl 2010, 96). In the case at hand no such theory has been provided yet. It would have to explain why so many people have wrongly believed for centuries that at least some single still pictures tell stories as we have already realized, and why it is still so natural for us to do so.

3. Addressing the problems of time

Time plays an essential role in most definitions of narrative, both on the representation and presentation side. Story events have to be located in time and the presentation of these events must itself have a temporal structure. However, unlike a moving picture or a verbal narrative, a single still picture does not seem to have a definite temporal structure. In a picture everything is presented simultaneously. While most authors admit that all elements of a picture are not perceived at the same moment in time, some argue that the aesthetic effect is as if we were taking in the whole picture “at a glance” (Chatman 1981, 118). Others deny this as well, but try to show that a picture’s perception is not consistently structured: different viewers see the picture elements at different moments of their viewing process and even the same viewer may follow a variety of different paths through the picture on different occasions. Each of these formulations seems to lead to the same basic problem: telling a story means to communicate temporal developments, and a single still picture is somehow not able to represent such developments or to represent them in a specific order. There are two basic ways to assert a picture’s inability to present temporality:

(1) Focus on pictures as representations. Pictures cannot represent temporal relationships.

(2) Focus on picture perception. Pictures either (a) are not perceived in time at all, or (b) cannot sufficiently determine how they are perceived in time.

Authors have at various times put forward a version of one of these options or versions of all three at the same time, either asserting them independently or presupposing that one implies the other.

3.1 The problem of representing time

There is no change of state without time, and even according to most liberal narratologists there is no narrative without at least one change of state. If (1) were correct, then, one would have to conclude that single still pictures are incapable of telling stories. The simplest version of (1) quite simply refuses the notion that pictures can show events that take time. In an article comparing different media in terms of their semiotics, Michael Titzmann explicitly states that pictorial propositions can only express synchronous states [synchrone Zustandshaftigkeit] but not temporal sequences of states [diachrone Zustandsfolgen]. Therefore only texts can tell stories (cf. Titzmann 1990, 379). As a result, a picture which shows several moments in time by showing the same person engaged in different activities at different locations in the picture space does not really show a temporal development.

Why exactly is a picture supposed to lack the capacity to represent diachrony? A reason independent of the time structure of visual presentations, which we will treat later, might be the absence of prepositions. While it is easy to say “before”, “five years later”, or “the next day” in texts, this may seem impossible with pictures. One might argue, however, that pictures are only less precise than verbal language in this respect. Storytelling does not presuppose precise specifications of temporal distance between the events told. We do not need expressions like “10 months, 3 weeks and 2 days later” to specify temporal order. “Many years later” or even simply “later” are good enough and pictures can provide for something equivalent to this: when we see a depiction of a man alive and then we see a depiction of his death, we know ceteris paribus that his actions must have preceded his death. Obviously, we may make incorrect deductions, but this is no different than in real life or when reading novels.10 The fact that we are sometimes misled certainly does not imply that no picture ever represents temporal order clearly. Our everyday knowledge is quite reliable in helping us grasp temporal order in pictorial representations of stories. The clues we use may be different, but they certainly are not less reliable than the clues provided by texts. What is also true in life and when watching movies is true for pictures: “In general, the spectator actively seeks to connect events by means of cause and effect. Given an incident, we tend to hypothesize what might have caused it, or what it might in turn cause. That is, we look for causal motivation” (Bordwell / Thompson 1990, 58).

The second reason why many scholars believe that pictures can only show synchronous states is because they think that the structure of pictorial presentation prevents pictures from showing temporal order; in terms of the arguments above this would mean that (2) implies (1). In a text one sentence follows the other. If the sentences describe events, then the description of one event follows the description of another event. This is what is meant by the formulation “the succession of signs” when Rimmon-Kenan says that there needs to be an “ontological gap between the succession of signs and the temporality of the events (in whatever expanded definition)” in a narrative (Rimmon-Kenan 2006, 16). There seems to be nothing equivalent to the succession of words and sentences in pictures.

3.1.1 Representing temporal developments: Lessing’s semiotics

Connecting this difference to the impossibility of narrative has a long history: according to G. E. Lessing, the difference between ‘signs which coexist in space’ and ‘signs which follow each other in time’ is one of the fundamental differences between pictures and texts. In order to understand why Christian Metz, Shlomith Rimmon-Kenan and others have deemed the succession of signs essential for storytelling, one must look at how this idea is connected to narrative in Lessing’s text:

I reason thus: if it is true that painting and poetry, in their imitations, make use of entirely different means; the former employing figures and colors in space, and the latter articulate sounds in time – if these signs indisputably require a suitable relation to the thing betokened, then it is clear, that signs arranged near to one another, can only express objects, of which the wholes or parts exist near one another; while consecutive signs can only express objects, of which the wholes or parts are themselves consecutive. (Lessing 1766, 101)11

Based on this explanation, Lessing then goes on to argue that pictures can only represent bodies, and that texts should focus on actions.

Like Lessing, Rimmon-Kenan sees the successive nature of signs in a medium as the key factor regarding the temporality of telling (and thus that medium’s narrative potential). But she does not offer any justification for this principle. How, exactly, is the succession of signs connected to narrative by Lessing? As we have seen, after introducing a distinction between signs that follow each other in time and signs that coexist in space, Lessing goes on to say that “these signs indisputably require a suitable relation [ein bequemes Verhältnis] to the thing betokened [bezeichnet]” (Lessing 1766, 101). In other words: if you try to tell successive events through signs which are not successive themselves, you are doomed to fail. As stories clearly imply succession in time, pictures are not suitable for telling stories.

This premise may certainly seem correct at first glance, but in reality anything can represent anything else. Every diagram showing changes in a company’s performance on the stock market uses signs in space to signify changes over time. Thus Lessing’s idea that signs in space cannot show temporal developments, if it is only based on his declaration that there must be a suitable relation between signs and signified, is surprisingly weak, and it is hard to understand why it still plays such an important role today. Rimmon-Kenan, who does not argue for her position concerning the succession of signs, has to face the same problem as Lessing. There is no good reason to suppose that signs in space cannot signify developments in time.

3.1.2 Representing temporal developments: Lessing’s norms

Lessing’s second distinction between pictures and texts, i.e. that pictures operate with natural signs while texts use conventional signs, lends weight to his argument. Because pictures are based on natural signs, and natural signs can only represent what they resemble, pictorial signs which are spatially near one another can only represent bodies where elements are also near each other in space. The fact is that Lessing opposes anything conventional in pictures, but this does not mean that using convention in pictures is impossible. Just as the convention we use when we understand that a woman with a sword and scales is a personification of justice, there are in fact clues and conventions which tell us that the events represented in a picture have succeeded each other in time. In many medieval and early Renaissance paintings, the depiction of the same person in different places alerts us to the fact that difference in spatial position is to be interpreted as difference in the moments when the events depicted take place (cf. Goodman 1981, 333). Quite aware of the existence of this convention, Lessing explicitly rejects it on aesthetic grounds:

Succession of time is the department of the poet, as space is that of the painter. To introduce two necessarily distant points of time into one and the same painting, as Fr. Mazzuoli has the rape of the Sabine women and their subsequent reconciliation of their husbands and relations, or, as Titian has the whole history of the prodigal son, his disorderly life, his misery, and his repentance, is an encroachment upon the sphere of the poet, which good taste could never justify. (Lessing 1766, 120)

As it appears here, the problem is really that ‘good taste could never justify this kind of encroachment’, not that the semiotic nature of pictures makes it impossible to tell stories: it is a problem of norms, not of facts.

Still, the idea that the essence of a medium implies rules for its use, and that these rules are different for different media, is quite intuitive. In fact, describing what makes something special quite naturally leads to explaining how it should be treated and what it should (and should not) be used for. It is against the backdrop of this kind of reasoning that Lessing first endorses the idea of media essentialism: if pictures are good at capturing spatial relationships and bad at capturing temporal relationships, he says, then we ought to use them for something closer to description than to narration. In the end, Lessing thus implicitly admits that it is possible to tell a story with a single still picture. After all Fra Mazzuoli and Titian tell stories. W. J. T. Mitchell is thus right to underline that “the ‘laws of genre’ which appeared to be dictated by nature turn out to be artificial, man-made statutes” (1984, 105).

3.1.3 Representing temporal developments: Lessing’s theory of illusion

Lessing’s discussion of poetry may help us better understand the exact nature of his argument concerning pictures. As the signs of language are arbitrary, he states, the signs of language can represent anything, including relations that are essentially spatial. In general, a poet should then have more freedom than a painter. However, Lessing quickly makes his argument more specific and restricts the freedom of the poet:

The poet does not merely wish to be intelligible; the prose writer indeed is contented with simply rendering his descriptions clear and distinct, but the poet has a higher aim. He must awaken in us conceptions so lively, that, from the rapidity with which they arise, the same impression should be made upon our senses, which the sight of the material objects, that these conceptions represent, would produce. In this moment of illusion, we should cease to be conscious of the instruments, by which this effect is obtained, – I mean words. (Lessing 1766, 111f.)

If she just wants to convey information, a writer may represent bodies and actions, whatever she deems relevant, but if she wants to make a strong (aesthetic) impression she should limit herself to actions only. This important passage has been rarely quoted, but it is essential to completely understand Lessing’s argument. It shows that Lessing would not have accepted the generalization of his argument from the domain of artistic texts and pictures to texts and pictures in general. The telos of all arts is, according to Lessing, deception, illusion. In poetry and painting “appearance takes the form of reality” and “the deception is, in either case, pleasing” (ibid., xiii). Lessing thus presupposes that the aim of verbal and visual art is to conceal art, but he does not presuppose that this is the aim of writing or creating pictures in general. As visual artists and writers use different kinds of signs, they must follow different rules to achieve striking renderings. The visual artist can hope to deceive a beholder when representing bodies but not when trying to depict actions.

Lessing’s point, it turns out, is not that pictures cannot tell stories, or that they should not tell them for ideological reasons (as W. J. T. Mitchell suggests) but simply that they cannot tell them as effectively as texts. But this statement is not quite as strong as the other. His imperative is not categorical. It is hypothetical. Lessing does not simply tell artists: “You should not imitate actions in pictures” or even “You should not imitate actions in pictures because the imitation of action should only be done in texts”, which is how W. J. T. Mitchell reads him. Lessing says: “If (and only if) you want to create pleasing illusions, then you should use pictures for description rather than narration”. Lessing’s point does not seem to be completely wrong: when the aim is to create representations which get close to deceiving spectators, the limitations Lessing suggests are definitely worth considering. While failing in its generalized version, Lessing’s argument might still be successful in its hypothetical form. After all, Trompe l’oeil painting most frequently depicts inanimate objects and not actions.

The claim that it is easier or somehow more effective to tell stories through texts, is much weaker than those made by Rimmon-Kenan or Chatman, who argue that it is impossible to tell stories with pictures. Lessing’s claim turns out to be psychological rather than conceptual. Saying that pictures should not show several moments if artists want to make a strong impression is less problematic than making the conceptual claim that it is impossible for pictures to show several moments. It is true that the kind of picture Lessing evokes, in which several moments are explicitly shown, has become unfamiliar to us:

We generally assume that a convincing representation of space implies the simultaneity of that which it includes, as would be essentially the case in a photograph, for example, in which everything is shown in the same place at more or less the same moment. Consequently it strikes us as odd and illogical when we see several successive episodes, with the same character or characters in each of them […]. (Andrews 1995, 3)

This impression seems to be particularly strong if a picture, such as a Renaissance painting, convincingly evokes space. As Andrews underlines, ‘it strikes us as odd and illogical’ when several different moments in time are shown in the same image. This, however, does not mean that such images are intrinsically odd or illogical, as we shall see.

3.1.4 The photographic convention and its unjustified privileges

Denying that pictures of the kind evoked by Andrews actually do show several moments means bestowing an arbitrary sense of privilege upon our viewing habits and pictorial standards. These standards are based on what could be called the photographic convention. According to the photographic convention, it is decided that every picture should always only show one moment in time, just like a standard photograph. But the photographic convention is only one possible convention for depiction. The photographic convention did not emerge as a result of progress in art either. It might be argued that some of the earliest depictions of mankind (e.g. Lascaux) already used what I call the photographic convention. As Richard Brilliant (1984, 52 and passim) has shown, it reappeared in Greek Antiquity, but then disappeared again.

In the Middle Ages, as mentioned earlier, many pictures were polychronic: whenever a spectator saw the same person several times in the same space, she would know that rather than showing a cohort of identical twins all dressed identically the pictures represented different moments in time. As Andrews underlines, the 14th century was more photographic in the temporal sense than the 15th century (cf. Andrews 1995, 8). Only after the 16th century did the photographic convention again start to become the standard way of creating pictures. Cubism and Futurism then systematically subverted it again. Today’s production of (photographic and non-photographic) pictures is clearly dominated by the photographic convention. This very short historical sketch shows that there is nothing necessary about the photographic convention and makes it highly unlikely that it is superior.

However, if this is simply a convention, how could we have come to believe that there was a necessary limitation to the temporal structure of a picture? I think that there are at least two reasons why this could have happened. The first is the erroneous belief that a picture does not have a “temporal ‘program’ inscribed in the work” itself (Chatman 1990, 7) because it does not dictate when, exactly, a spectator sees a specific picture element. The second is the possibility of entrenchment. Conventions, which are arbitrary by definition, can become so ‘entrenched’ that they come to appear necessary. When the norm is followed consistently, two elements which were initially linked through a norm, can appear to be logically, naturally, or conceptually connected. Instead of believing that pictures should only ever show one moment in time as Lessing still did, we have come to believe that they can only show one moment. If we accept that pictures always necessarily represent one moment only, this implies that a picture that depicts the same person at several moments is not merely incorrect or strange, but rather does not exist. While few modern scholars would have explicitly endorsed this kind of idea, many of them have done so implicitly. When an author explicitly accepts that one can interpret (cf. Goodman 1981, 333) a picture that really depicts identical twins in different spatial locations as if it showed the same person at different moments, that it can represent (cf. Schaeffer 2001, 19) or evoke temporal developments but not show them, it means that he applies the photographic convention to pictures which operate with another convention. All these terms downgrade the representational range of pictures and presuppose that the semantics of pictures we are most familiar with must necessarily govern every picture. It then becomes seemingly impossible for a beholder to immediately understand the work of creators who adhere to different conventions. To understand such a picture, we need to go through a process of interpretations, deductions, and transpositions until we have attained a transcript in a sign system which conceptually allows for the representation of several moments: a verbal description, for instance.

This description of the process of understanding what a picture represents may be approximately correct for an audience for which the photographic convention is the standard. But to say that a picture in which the same person appears to be doing different things at different moments in time really shows one person in different spatial locations at one moment would be like saying that sentences spoken by native speakers of a language without a past and future tense (like Mandarin Chinese) always really refer to the present and that all reference to past and future must somehow be deduced. The belief that monochronic depictions are somehow more natural than polychronic ones is just as absurd as believing that English is somehow more sound than Mandarin Chinese. When we translate Chinese to English, we express the temporal relations differently. But this does not imply that the Chinese language did not express them already.

It is important to note that, as opposed to what Jean-Marie Schaeffer suggests (2001, 13f.), accepting that conventions play a role for understanding pictures in no way obliges us to give up the fundamental difference between texts and pictures. A text is mostly based on arbitrary signs (onomatopoeia aside); thus, we need to learn conventions in order to understand a text. As opposed to this a picture is at least in part based on natural signs. We recognize what the sign stands for based on some kind of resemblance. After having seen a dog face-to-face, we may recognize the same (kind of) dog in pictures – but not in texts. In order to understand that the name “Fido” refers to the dog we have seen, looking at the word is not enough. In C. S. Peirce’s terminology, pictorial signs are predominantly icons: we understand what they show because they look similar to their referents, while most words are predominantly symbols, i.e. they are linked to their objects through habit (or convention) (cf. Peirce 1909, 460f.). We can give up the idea that one kind of picture semantics is more natural than others without giving up the idea that we understand most of what we see in pictures based on our previous visual experience of the objects they depict (or other depictions we have seen earlier).

Cubist painting shows that depiction can be both natural and conventional at the same time.12 Just like our knowledge of real guitars as seen from different viewing positions informs our understanding of Cubist pictures of guitars, our knowledge of real people doing different things at different moments in time informs our understanding of polychronic pictures. One important conclusion to draw from this demonstration is that texts and pictures alike are capable of portraying sequential events as being sequential.

3.2 The problem of perceiving pictures in time

According to a second version of the time problem the dilemma lies not only in the supposed lack of ability of pictures to portray changes of state (which have a certain duration), but in the inability of pictures to portray developments in time. The problem here is with the ability of pictures to present, not their ability to represent. A diagram can show developments that take time using only space on the page, but – unless we use animated diagrams like the ones we sometimes see in PowerPoint presentations – it does not really show them ‘in time’. The argument’s first premise is that in order to be suitable for narrative, a medium must not only be able to somehow represent the temporal structure of a story, it must also have its own time structure (cf. Metz 1968, 27). The common-sensical observation that everything must necessarily be shown all at once in a picture is the argument’s second premise. From these two premises, the conclusion that single still pictures are not suitable for narrative quickly follows. According to this view, the texts and productions mentioned by Bremond (1964) are, however, potential vehicles for stories. Movies, theater and ballet clearly have their own time structure which they impose on us during their reception. If we look away or walk out, we miss a part of the story being told.

3.2.1 Arts of time and arts of space: in favour of a more fine-grained distinction

The arts which present in time, e.g. writing, drama, music, and ballet, are regularly referred to as the arts of time, while the arts which are supposed to lack a temporal structure of presentation, e.g. painting or sculpture are referred to as the arts of space. Though useful, this distinction is quite coarse-grained. There is an essential difference between the presentation-time structures of written texts, many musical works, and movie screenings. When we activate a movie, we have to accept its temporal structure precisely. Its duration is a defining attribute of each movie. There is no equivalent for books. Theorists frequently try to explain away this difference by using pages or paragraphs in books as units equivalent of temporal duration, but this can hardly be justified. It does not matter whether it takes me 16 hours or 20 hours to read Wuthering Heights – in the end, I have read the same book.

To capture this difference between media, Philippe Marion differentiates homochrone media from heterochrone media. Works in the former mode, but not the latter mode impose their time frame on us (cf. Marion 1997, 82). In heterochrone media, reception-time determines the duration of (individual) presentation whereas in homochrone media, presentation-time determines the duration of (individual or collective) reception. Just like novels, pictures are heterochrone. The heterochrony of a novel seems to pose no problem for its narrative status, so why should this be so for pictures? From this perspective there is a bigger difference between a novel read to me and a novel read by me than between a novel and a picture. This also shows that there is a fundamental gap between auditory transmission of texts and their reception in written form. Dismissing this difference as minor, like most theorists since Lessing have, seems rather hasty.

3.2.2 The need for a ‘temporal program’

According to what could be called the standard view, the reading-as-presentation is clearly programmed in texts, but this is not the case in pictures. In order to understand a standard text, we have to read it from top left to bottom right. A writer can thus program the order of reading-as-presentation as she pleases. There is no rule for pictures that is as strong as this convention. Pictures do rarely depict represented events in a conventional order (cf. Goodman 1981, 338), and even when different spectators look at the same picture, they do not follow the same paths. As a matter of fact, most studies on eye-movement seem to contradict rather than confirm the wide-spread notion that there is a path in a picture which all spectators follow in the same order. During prolonged viewing, different spectators tend to focus on the same zones (for example, depictions of faces). However, the only element which seems relatively stable across viewings by different spectators are the zones of fixation and the paths different viewers take regularly, but not the direction or order of viewing of the elements which lie on this path (cf. Engelbrecht et al. 2010, 38). Unlike the order of activation through reading, which reflects order of telling, the order of activation through viewing does not allow the artist to predetermine a (potential) order of showing.

Temporal media are different from spatial media insofar as they control presentation-time. A temporal medium “requires us to begin at a beginning it chooses (the first page, the opening shots of a film, the overture, the rising curtain) and to follow its temporal unfolding to the end it prescribes” (Chatman 1990, 7). A picture, on the other hand, does not have “a ‘temporal program’ inscribed in the work” (ibid.). While the first part of this assertion might be quite right (some media require us to start at a precisely fixed point and end at another), the second certainly is not (pictures can in fact have a temporal program).

3.2.3 Why relative determination of order of reception is sufficient

To understand what is at stake here we mustn’t forget that we are concerned with the narrative potential of different media, not their potential for achieving total mind-control. It is true that filmmakers (for instance) can control what a spectator will see at any given point in the runtime of the film (at the two-minute mark, for example), while painters do not have this kind of control. However, it is important to realize that such control is not necessary for a work to be defined as narrative. The reception of works in the temporal medium of writing is, in this respect, closer to the reception of still pictures than to watching movies. As a matter of fact, hardly anybody would argue that people who read a story more slowly have not read the same story as people who read it faster.

What really matters – because it might actually affect story identity and because it certainly affects the identities of narratives – is the order of reading-as-presentation. A writer only needs to control relative chronology of reception: It may make a significant difference if a reader learns about event A after or before learning about events B and C. But it is not usually deemed essential when exactly he learns about either of them. It is quite clear how to apply these principles to pictures: Even though different spectators may start their viewing at different points and move through pictures in different ways, it might still be that all of them see the depiction of (significant) story event A before seeing the depiction of (significant) story event B. If events A and B form a story, the order of viewing-as-presentation establishes the order of telling.

To create a determinate order of showing-as-telling, an artist needs less than total determination of order of perception. While there are probably very few pictures that achieve total determination of the order of viewing, i.e. pictures whose elements are viewed in the same order by most spectators, there may still be many pictures which sufficiently determine order, i.e. pictures whose representations of significant story elements are viewed in the same order by most spectators. A relatively simple visual technique to make sure that something is not perceived first is to make it small, to put it in the background and to avoid bright colors. There are indeed pictures in which it is scarcely possible that a spectator sees the depiction of event E first, but where she cannot help but consider it significant once she discovers it.

3.2.4 An example of sufficient relative determination of order of reception: Brueghel’s Landscape with the Fall of Icarus

Many Flemish paintings created between the 16th and 18th centuries quite consciously control viewing time in such a way that viewers make one or more significant discoveries if they look at them long enough. A painting by Pieter Brueghel, now commonly referred to as Landscape with the Fall of Icarus, for instance, shows a landscape with a shepherd, some sheep, a farmer plowing, a fisherman fishing and a large ship in the ocean. Only if we contemplate the picture for long enough do we discover two legs somewhere in the sea and realize that the picture tells a story.13 Whether the spectator thinks of Icarus or not, whether she has heard of him before or not, the two legs are quite a discovery for someone who looks at the picture. So much of a discovery, indeed, that it forces her to rethink her understanding of the scene depicted. What seemed to be an idyllic landscape becomes a scene of death by drowning. The other figures depicted as simply minding their own business may suddenly seem to be acting inappropriately.14

This famous painting, whose descriptive title is not part of the original artwork but was only added later, is clearly programmed to hide Icarus from being discovered too early. Needless to say, we know that the man must have fallen into the sea before the situation depicted and that he will probably be gone soon. If we look even longer, we may also discover a purse on the ground and corpse in the bushes. This example shows how a picture can determine in what order different elements are perceived. It also provides for double-temporality. The time of viewing-showing-telling is ‘ontologically independent’ of story-time, as required by Rimmon-Kenan. Scripts, schemas and knowledge of causal relations allow us to understand the order of story events. Conversely, the order of presentation is determined by the way in which we view the picture. This viewing is in turn dependent on the way in which the elements depicted on the canvas have been organized. The two processes are thus of a different kind and largely independent of one another.

What is more, Brueghel creates a tension between the relative importance of the events he depicts and their order of appearance. We instinctively know that we somehow determine the order of showing-as-telling through the way we view the picture. Therefore, we feel in a certain sense responsible for our late perception of the dying man and the corpse: we suddenly find ourselves in the same cognitive position as the other persons in the picture. We cannot blame them for not seeing what has happened, lest we also blame ourselves. Just like the shepherd watching the sky, we had contemplated the beauty of the landscape, the animals, and the great ship, and had not noticed that one man was about to die while another was already dead. This painting has often been related to the Dutch saying “No plough stands still because a man dies” (see, for example, Wyss 1990, 12). The saying is both illustrated and performed by each new viewer in her viewing process (unless, of course, she has read the title of the painting first).15

Let me anticipate an objection here. The Landscape with the Fall of Icarus may seem so exceptional that it does not allow to draw conclusions about the narrative potential of pictures in general. But it is not the only picture of its kind. Many of the works of Joachim Patinir (c. 1480 - 1524), Pieter Aertsen (1507/08 - 1575), Joachim Beuckelaer (c. 1534 - c.1574), and some pictures created by Albrecht Altdorfer (c. 1480 - 1538), Cornelis van Poelenburg (c. 1586 - 1667), Diego Vélazquez (1599 - 1660) and David Teniers (1610 - 1690), are structured in a similar way: at least one essential story event is not visible in them at first sight and when we finally discover it, this discovery is eventful. Art historians have coined the term ‘mannerist inversion’ to designate the kind of pictures which hide a significant event in a more ordinary setting, but Renaissance and Mannerist artists are not the only ones who have used it. Théodore Géricault’s The Raft of the Medusa (1818-1819), Jeff Wall’s Eviction Struggle (1988), or Cindy Sherman’s Untitled #175 (1987) arguably allow for similar experiences.

4. Conclusion

This essay has shown that even the most restrictive definitions of narrative do not exclude all pictures from the realm of narrative. Of course, narratological analysis may yield interesting results even for pictures that are not narratives in a strict sense, i.e. pictures that do not exhibit a clearly inscribed temporal program. Although they are not narratives according to the narrow definition used in this article, such artefacts might still possess a high degree of narrativity, as Mieke Bal, Wendy Steiner, Rick Altman and Werner Wolf have amply demonstrated.


Altman, Rick (2008): A Theory of Narrative. New York.

Andrews, Lew (1995): Story and Space in Renaissance Art. Cambridge.

Baggini, Julian / Fosl, Peter S. (2010): The Philosopher’s Toolkit. Oxford et al.

Bal, Mieke (1991): Reading Rembrandt. Beyond the Word-Image Opposition. Cambridge et al.

Bal, Mieke (1994): On Meaning-Making. Essays in Semiotics. California.

Bal, Mieke (1997): Narratology. Introduction to the Theory of Narrative. Second Edition. Toronto et al.

Boehm, Gottfried (2010): “Ce qui se montre. De la différence iconique”. In: Emmanuel Alloa (Ed.), Penser l’Image. Dijon. 27-47.

Bordwell, David (1985): Narration in the Fiction Film. Madison.

Bordwell, David / Thompson, Kristin (1990): Film Art. An Introduction. New York et al.

Bremond, Claude (1964): “Le message narratif”. In: Communications 4 (No. 4), pp. 4-32.

Brilliant, Richard (1984): Visual Narratives. Storytelling in Etruscan and Roman Art. Ithaca / London.

Brown, Lee B. (1986): “Review of Iconology: Image, Text, Ideology by W. J. T. Mitchell”. In: The Journal of Aesthetics and Art Criticism 45 (No. 2), pp. 211-214.

Chatman, Seymour (1981): “What Novels Can Do That Films Can’t (and Vice Versa)”. In: W. J. Thomas Mitchell (Ed.), On Narrative. Chicago, pp. 117-136.

Chatman, Seymour (1983): Story and Discourse. Narrative Structure in Fiction and Film. Ithaca / London.

Chatman, Seymour (1990): Coming to Terms. The Rhetoric of Narrative in Fiction and Film. Ithaca.

Dieterle, Bernard (1988): Erzählte Bilder. Zum narrativen Umgang mit Gemälden. Marburg.

Engelbrecht, M. et al. (2010): “Dem Auge auf der Spur: Eine historische und empirische Studie zur Blickbewegung beim Betrachten von Gemälden”. In: IMAGE – Zeitschrift für interdisziplinäre Bildwissenschaft 11 (January). URL: http://www.gib.uni-tuebingen.de/image-50_en?function=fnArticle&showArticle=159 (27th May 2013).

Gaudreault, André (1988): Du littéraire au filmique. Système du récit. Paris.

Genette, Gérard (1983): “Letter to André Gaudreault 22 of February 1983”, quoted by Gaudreault (1988), Du littéraire au filmique. Paris. Note 8, 29.

Goodman, Nelson (1981): “Twisted Tales; Or Story, Study, and Symphony”. In: Synthese 46 (No. 3). [The Richard Rudner Memorial Issue], pp. 331-349.

Imdahl, Max (1975): “Wandel durch Nachahmung. Rembrandts Zeichnung nach Lastmans ‘Susanna im Bade’”. In: Vorträge Kurt Badt zu Ehren. Konstanz, pp. 19-44.

Karpf, Jutta (1994): Strukturanalyse der mittelalterlichen Bilderzählung. Ein Beitrag zur kunsthistorischen Erzählforschung. Marburg.

Kemp, Wolfgang (1996a): Die Räume der Maler. Zur Bilderzählung seit Giotto. München.

Kemp, Wolfgang (1996b): “Narrative”. In: Robert S. Nelson (Ed.), Critical Terms for Art History. Chicago, pp. 58-69.

Kibédi Varga, Aron (1989): Discours, récit, image. Brussels et al.

Kibédi Varga, Aron (1990): “Visuelle Argumentation und visuelle Narrativität”. In: Wolfgang Harms (Ed.), Text und Bild, Bild und Text. DFG-Symposion 1988. Stuttgart, pp. 356-367.

Leitch, Thomas M. (1986): What Stories Are. Narrative Theory and Interpretation. University Park.

Lessing, Gotthold Ephraim (1766): Laocoon or the Limits of Poetry and Painting. London 1853 [English translation by E. C. Beasley].

Lopes, Dominic (1996): Understanding Pictures. Oxford.

Marion, Philippe (1997): “Narratologie médiatique et médiagénie des récits”. In: Recherches en Communication 7, pp. 61-88.

Metz, Christian (1968): “Remarques pour une phénoménologie du narratif”. In: Christian Metz (Ed.), Essais sur la signification au cinéma. Paris, pp. 25-35.

Mitchell, W. J. T. (1984): “The Politics of Genre: Space and Time in Lessing’s Laocoon”. In: Representations 6 (Spring), pp. 98-115.

Peirce, Charles Sanders (1909): “A Sketch of Logical Critics”. In: The Peirce Edition Project (Ed.), The Essential Peirce. Selected Philosophical Writings. Vol. 2 (1893-1913). Bloomington / Indianapolis 1998, pp. 451-463.

Pericolo, Lorenzo (2011): Caravaggio and Pictorial Narrative: Dislocating the Istoria in Early Modern Painting. London.

Prince, Gerald (1999): “Revisiting Narrativity”. In: Walter Grünzweig / Andreas Solbach (Eds.), Grenzüberschreitungen: Narratologie im Kontext / Transcending Boundaries: Narratology in Context. Tübingen, pp. 43-51.

Prince, Gerald (2012): “Récit minimal et narrativité”. In: Sabrinelle Bedrane et al. (Eds.), Le récit minimal. Du minime au minimalisme. Littérature, arts, media. Paris, pp. 23-32.

Ranta, Michael (2011): “Stories in Pictures (and Non-Pictorial Objects) – A Narratological and Cognitive Psychological Approach”. In: Contemporary Aesthetics 9. URL: http://www.contempaesthetics.org/newvolume/pages/article.php?articleID=619 (27th May 2013).

Revaz, Françoise (1997): Les Textes d’action. Metz.

Rimmon-Kenan, Shlomith (2006): “Concepts of Narrative”. In: Matti Hyvärinen et. al. (Eds.), The Travelling Concept of Narrative. Studies across Disciplines in the Humanities and Social Sciences. Helsinki, pp. 10-19.

Schaeffer, Jean-Marie (2001): “Narration visuelle et interpretation”. In: Jan Baetens / Mireille Ribière (Eds.), Temps, Narration & Image Fixe. Time, Narrative and the Fixed Image. Amsterdam, pp. 11-27.

Sonesson, Göran (1997): “Mute Narratives”. In: Ulla-Britta Lagerroth et al. (Eds.), Interart Poetics. Essays on the Interrelations of the Arts and the Media. Amsterdam / Atlanta, pp. 243-250.

Steiner, Wendy (1988): Pictures of Romance. Form against Context in Painting and Literature. Chicago.

Steiner, Wendy (2004): “Pictorial Narrativity”. In: Marie-Laure Ryan et al. (Eds.), Narrative across Media: The Languages of Storytelling. Lincoln, NE, pp. 145-177.

Thürlemann, Felix (1990): Vom Bild zum Raum: Beiträge zu einer semiotischen Kunstwissenschaft. Köln.

Titzmann, Michael (1990): “Theoretisch-methodologische Probleme einer Semiotik der Text-Bild-Relationen”. In: Wolfgang Harms (Ed.), Text und Bild, Bild und Text. DFG-Symposium 1988. Stuttgart, pp. 368-384.

Wolf, Werner (2002): “Intermediale Erzähltheorie”. In: Ansgar & Vera Nünning (Eds.), Erzähltheorie transgenerisch, intermedial, interdisziplinär. Trier, pp. 23-104.

Wolf, Werner (2003): “Narrative and Narrativity: A Narratological Reconceptualization and Its Applicability to the Visual Arts”. In: Word & Image 19 (No. 3), pp. 180-197.

Wolf, Werner (2011): “Narratology and Media(lity): The Transmedial Expansion of a Literary Discipline and Possible Consequences”. In: Greta Olson (Ed.), Current Trends in Narratology. Berlin / New York, pp. 145-180.

Wyss, Beat (1990): Pieter Bruegel: Landschaft mit Ikarussturz: ein Vexierbild des humanistischen Pessimismus, Frankfurt.

Klaus Speidel
ED 5: Concepts et Langages
Université Paris Sorbonne

URL: http://paris-sorbonne.academia.edu/KlausSpeidel
E-mail: klausspeidel@gmail.com

Bitte zitieren Sie nicht die HTML-Version, sondern ausschließlich die PDF-Datei / Please do not cite the HTML version but only the PDF file:

URN: urn:nbn:de:hbz:468-20130604-083458-0

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License.

1The author wishes to thank Roy Sommer, the editorial team of DIEGESIS, and the anonymous reviewers for their valuable comments on earlier drafts of this essay.

2Kemp is of particular interest because he introduced art history to reader-response theory. His approach thus seems to be compatible with post-classical narratology. However, his points of reference are mainly art historical. One exception is the article “Narrative” (Kemp 1996b), which he contributed to the collection Critical Terms for Art History. In the last part of this short article, Kemp ventures some connections to classical narratology, trying to reconceptualize the concepts of lack, desire and transformation in reader-response terms. The article is very insightful and its focus on reception can be considered an anticipation of some aspects of Werner Wolf’s articles on visual narratology. In his Die Räume der Maler (1996a), Kemp analyzes the spatial and narrative structure of paintings, presupposing that a picture can tell a story. Another important study is Brilliant (1984), where the author analyzes narrative in Etruscan and Roman antiquity. Thürlemann (1990) applies concepts developed by Greimas and Hjelmslev to pictures. Karpf (1994) focuses on medieval picture cycles, applying concepts of classical narratological semiotics to pictures. An early example of narrative analysis of pictures in art history is Max Imdahl (1975), who analyzes how Rembrandt slightly transformed a drawing by his master in order to more convincingly tell the same story. Most of these studies presuppose that pictures can be narratives or commit the fallacy I have mentioned above, i.e. they suppose that the fact that tools of narrative analysis can be applied to pictures shows that the pictures analyzed are narratives.

3This is also the main result of Lee B. Brown’s review of Mitchell’s Iconology (cf. Brown 1986).

4According to Altman’s initial definition, only paintings where the same person is shown at several times, performing different actions, could have been narratives. As this is not the case with the paintings he wants to discuss, he changes his definition.

5The same holds true of Michael Ranta (2011), whose broad cognitivist approach to pictorial narrative implies that even natural objects can evoke narratives.

6In terms of logics, it would be enough if I was able to show that there is at least one thing (one picture) which is part of the group of single still pictures and the group of narratives (it tells a story). This is so because, in terms of logics, the thesis I have to refute is that nothing can be part of the group of single still pictures and the group of narratives at the same time.

7David Bordwell rejects the adoption of the histoire / discours dichotomy used by Christian Metz. For Bordwell the use of “discourse” is linked to enunciation theories of film narration, which postulate an implicit narrator for movies and omit the essential difference between verbal language and pictures (cf. Bordwell 1985, 23-26).

8Unless indicated, all translations from the French and German are mine.

9Göran Sonesson (1997, 244) adopts a different strategy. According to him, “narrative supposes at least two events with a temporal link on the content side. Thus far, then, no particular requirement seems to be imposed on the expression side”. He later ads that “a breach of the structure of expectancy” (ibid., 245), eventfulness is also needed. Based on such a broad definition it then becomes relatively easy to show that pictures can be narratives. I will here work with a much more narrow definition in order to show that the definitions can be as narrow as needed without excluding pictures from being narratives.

10It could even be argued that we are more likely to correctly interpret causal relationships in pictures than in life, because pictures are communicative artefacts: A clue in a picture was created by someone as a clue, and is not only interpreted as such by us. The knowledge we have of time and causation in pictures is just as good as the knowledge we have of them when we face a new situation in real life.

11While he generally speaks about “painting” and “poetry” rather than “pictures” and “texts” Lessing underlines that poetry as used in the Laocoon often refers to all arts “in which the imitation is also progressive” while painting refers to “plastic arts generally” (1766, xviii). The generalization of his arguments to the domain of texts on the one hand and pictures on the other is therefore legitimate.

12Dominic Lopes (1996, 32-36) offers a concise discussion of the potential tensions between a resemblance theory of depiction and the fact that different pictures are constructed and perceived within different systems which often have incompatible rules. He concludes that resemblance must be relativized. I believe that depiction of temporal relationships must similarly be relativized.

13To walk away from the painting before every part of it has been viewed would be just like walking out of a movie before it is finished or stopping to read a novel. It would leave its perception incomplete.

14This is one of the themes of W. H. Auden’s poem “Musée des Beaux-Arts” published in 1939, which takes the form of a poetical comment on several paintings by Brueghel. Auden’s interpretation seems much more convincing than Rick Altman’s.

15I thus believe Rick Altman’s (2008, 213f.) interpretation of the same painting and which presupposes knowledge of the painting’s title to be misguided.