Chapter 3 of Visual Grammar

[Contents] [Previous: Chapter 2 Architecture] [Next: Chapter 4 Rules]

In the last two chapters I have looked at visual language from a semiotic and cognitive viewpoint. I have central to these chapters were the questions does visual language exists, and how does visual grammar look like? Closely linked to these questions is another question which I will try to answer this chapter: what does visual grammar contribute to the analysis of visual representations? It is a question that must not be overlooked. This thesis is not merely a mental exercise, I do not construct a visual grammar for the sake of visual grammar itself. Visual grammar is a tool for analysis and as such it is important to keep in mind to what goal this tool is created.

In the this chapter, I will examine visual language from an approach that can arguably be called post-modern, although that label itself clarifies only a little. I will examine language within a larger socio-cultural context. The 'post-modern' approach, as we will see, does not regard language as a unified system as might have been suggested in the previous chapters. Rather language, and visual language, can be seen as a set of subsystems that all interact with each other, the language itself and other systems outside language. In this chapter I will concentrate establishing this notion subsystems in visual language and the role visual grammar can play in the analysis of such subsystems.

The idea of subsystems is derived from a discussion of the conception of language by the Russian scholar Mikhail Bakhtin in the first section of this chapter. In the following two sections, I will investigate two different, although related, aspects of visual subsystems: modality and intertextuality. In section 3.4 I will incorporate subsystems into the Modular Parallel Architecture. In the last section of this chapter I will discuss the analytical potential of visual subsystems for the study of visual grammar. This is an important section, because I aim to exploit this analytic potential in chapter 5.

3.1 Language Revisited

Language is a highly social phenomenon. This is one of the most important claims Mikhail Bakhtin makes in his essay Discourse of the Novel. The social nature of language makes language as diverse and dynamic as culture itself. For example, to think about the English language as one system is according to Bakhtin a misconception. He distinguishes several forces that are active within in language. Some of these forces are so strong and structured that they can be called textual systems in themselves; poetics has its own patterns distinctive from, say, scientific language. All of the textual systems that are part of a language share a common ground, which is a vocabulary and a set of grammatical rules (in the traditional use of the term; the grammar one learns at school). However, each of these textual system has its own conventions for the construction of texts and unique set of connotations.

The notion of a textual system is very broad. It can be said that different social, cultural and historical groups create, develop or have their own textual systems. Such groups or subcultures usually develop a highly codified system of signs that besides speech include gestures, dress, behaviour and etiquette; it goes beyond linguistics. In a society there exist different groups and it is possible, even likely, for one person to belong to several of these groups at once. Bakhtin gives the example of an illiterate Russian peasant of the first half of the 20th century that 'speaks' the language of the Slavonic Church to pray, sings songs in the language of folklore, has language to communicate with his family and uses the official language to address the local authorities (Bakhtin 1981: 295). Today, most people show totally different behaviour at work, at home and at the sports club. They wear different clothes for such occasions and probably employ a different 'language' to address their co-workers, friends and family. All these different behaviours, dress-codes and 'languages' are regarded as different semiotic modes, based on different semiotic codes, and therefore form separate systems in themselves.

When considering languages, Bakhtin distinguishes between centripetal and centrifugal forces. Centripetal forces unify languages and strengthen one dominant world view. According to Bakhtin, poetics and traditional linguistic scholarship are part of the centripetal forces of language, because they contribute to the existence of one (ideological and grammatical) 'correct' language (ibid. 270)[1]. On the other hand, the centrifugal forces include various dialects, professional languages, the languages of different generations, parody and irony. These textual systems need a central unitary language; it provides a common ground that is required for communication. However, the collision between these systems and the language of which they are part, keeps language diverse, alive and ensures its development (ibid. 271-272).

For Bakhtin novels are the most excellent examples of centrifugal texts. Novels are multiform in style and confront the reader with a diversity of textual systems. Individual speech uttered by various characters and the narrator are examples of such different textual systems that can be incorporated within a novel. Also, many novels include semi-literary or extra-artistic utterances, such as diary-entries and moral, scientific or philosophical statements (ibid. 261-262). Again, these are manifestations of different textual systems within one text.

Bakhtin's integrates language with a larger socio-cultural context. As a result, he distinguishes several subsystems or sublanguages within language that roughly follow social and cultural strata. Each linguistic subsystem differs from the language it is part of and from other subsystems within the same language. I assume that this also holds true for visual language, the existence of visual genres that are used at different occasions and by different social groups supports this assumption. It also opens up the possibility of analysis that focuses on the way subsystems differ from each other and the (super)system of which they are part. Such analysis will overlap with the analysis of different visual genres such as cultural posters, political posters, comics and art.

I will proceed by investigating two important features of visual language: (visual) modality and (visual) intertextuality. The notion of intertextuality has been developed with a direct link to Bakhtin's ideas, thus visual intertextuality will tie in nicely with visual subsystems. Visual modality, as discussed by Kress and Van Leeuwen, is a concept that has many similarities with visual subsystems, and therefore deserves some attention in this chapter. I do not wish to claim that with the discussion of modality and intertextuality the potential influence of subsystems on the visual system is exhausted. However, I do feel that within the scope and goal of this thesis, these are sufficient.

3.2 Modality

Although the notion of subsystems as presented here, or anything similar, is not prevalent within Reading Images, an interesting example of such an analysis is found when Kress and Van Leeuwen discuss visual modality. The concept of visual modality comes pretty close to the analysis of visual subsystems, even though the aim of Kress & Van Leeuwen is different. Modality is a term Kress and Van Leeuwen borrow from linguistics. But here it is used to refer to the credibility or truth-value of a message (Kress & Van Leeuwen 1996: 159-160). However, their use of modality in visual analysis implies a more potent use of the term, in which modality refers not to credibility but to the different modes an image can have.

Verbal representations can include modality cues or modality markers. In verbal representations, modality markers serve to align readers with some statements which creates some distance between the reader and other statements. This is usually done by clearly attributing statements to certain people (ibid. 160). For Kress and Van Leeuwen, S1 and S2 below have different modalities.

S1 Aboriginal people had no religion.
S2 This made them think that Aboriginal people had no religion.

S1 has a higher modality it is presented as being more likely to be true than S2. In the last sentence the statement 'Aboriginal people had no religion' is attributed to 'them'. It is quite conceivable that 'they' were wrong in their conception because something 'made them think so'. Whereas the same statement in S1 is presented without being attributed to a person or a group which makes it seem to be more objective[2].

Visual representations can also have a high or low modality. However, the way this is done for pictures is more complex than for verbal texts. It is a popular conception that a picture (or the camera) does not lie. This gives a picture a 'natural' high modality. Even though the illusions created by cinematographers are nowadays pretty convincing, but clearly not always real.[3] In any case, Umberto Eco has shown that the ability to lie is one of the key characteristics of any semiotic system, including visual language (Eco, 1979: 6-7). According to Kress en Van Leeuwen things can be depicted in such a way that they seem real or fictional. Use of colour is one example by which visual representation can lend a high or low modality to the things depicted. In natural images, a natural use of colours would mark high modality, whereas the absence of colour or unnatural vividness of colours marks low modality[4]. However, within the context of a technical or scientific diagram absence of colour is the preferred mode and natural use of colour marks low modality (Kress & Van Leeuwen 161-165).

Kress en Van Leeuwen identify 8 visual modality markers. Each is a gradual scale, and for natural images maximum modality corresponds to a position somewhere between the two extremes (ibid. 165-167). It is doubtful that the eight modality markers Kress and Van Leeuwen identify are easily distinguished or exhaustive. The four modality markers that I think are the most clear-cut and applicable are:

  1. Colour saturation, runs from absence of colour to full colour saturation.

  2. Colour modulation, runs from the use of plain colours only to the use of many individual shades.

  3. Contextualization, runs from the absence of background to the most detailed background.

  4. Depth, runs from absence of depth to deep perspective.

The values of the modality markers of most natural images will resemble the positions represented in figure 3.1. There are some examples of pieces of (modern) art that have a high modality according to some of the scales, but low modality on some other scales, often with intriguing results. Figure 3.2 is an example taken from Kress and Van Leeuwen. This image has high modality according to the contextualization and depth scale, but it has low modality according the colour saturation and modulation scales.

Figure 3.1 - Natural modality slides

Figure 3.2 - Patrick White (Louis Kahan 1963)

Apart from the modality markers, Kress and Van Leeuwen also distinguish between four 'coding orientations' that are employed by different social groups and that attribute high modality to a different values among the modality scales. How the modality is distributed for each of these coding orientations is represented in figure 3.3. Incidentally, each coding orientation uses the same distribution for each Modality marker. Thus the Technical coding orientation attributes high modality to low values on colour saturation, modulation, depth and contextualization (ibid. 170).

Figure 3.3 - Coding Orientations

These coding orientations reflect the different values and conventions in use by the social groups that employ them. The groups that can be associated with the various orientations are distinguished from culture in general in different ways. The technological orientation can be associated with certain professional groups. Although Kress and Van Leeuwen do not differentiate the technological orientation beyond this initial level, it is conceivable that this can be done for various particular technological professions. Architects and mechanical engineers will probably prefer the existence of some kind of depth (or isometric perspective), whereas more abstract professions (such as linguistics) usually prefer total absence of depth.

Abstract coding orientations are used by the socio-cultural elites (ibid. 170). But there seems to be in-betweens between the images favoured by the socio-cultural elite and natural imagery that are omitted by Kress and Van Leeuwen. An interesting example is film-posters. Cult movies often have film-posters that tend to be more 'abstract' their imagery shows a shift towards the abstract coding orientation than the posters of a popular movie. Figures 3.4 and 3.5 feature two film posters. The first is for the cult movie Ghost World. The poster's absence of depth, and context, as well as the strongly modulated and vivid colours, are examples of this shift. It is a striking contrast to the poster of Lord of the Rings, The Fellowship of the Ring, which imagery is by most standards more natural, even though the first poster makes use of a photograph and the latter is a drawing (of an imagined scene).

The sensory coding orientation is encountered in context where pictures need to be visually pleasing. Examples are illustrations in cooking books, advertising and certain types of art.

Figure 3.4 - Ghost World

Figure 3.5 - The Lord of the Rings

One other very useful application of modality seems to go unnoticed in Reading Images. Kress and Van Leeuwen do not discuss the meaning or occurrences of different local modalities within one picture, even though some of the pictures in their books contain excellent examples of this. Figure 3.6 is one of such examples. Kress and Van Leeuwen discuss this image in the context of visual symbolism. The portrait of late artist Brett Whiteley in the image is considered to be symbolic because "of the non-naturalistic and highly conspicuous way in which it is lit" (Kress & Van Leeuwen, 1996: 109, my italics). Where the largest part of the picture is mostly black and white, the portrait features more greyscales. This can be considered to be a shift in modality. I agree with Kress and Van Leeuwen about the symbolic character of the portrait, but attributing it to a local modality shift, allows for a better and more accurate method to identify visual symbolism[5].

Figure 3.6 - Arkie Whiteley

I am inclined to think of these coding orientations as different modes of representation that are aligned to different conventions of graphic representation, and by extent to different subsystems of visual language. Also, all modality markers can be associated with the means of articulation discussed in chapter 1, although some such as depth can only be associated to a complex of means of articulation and its conceptual effects. All in all, the idea that there exist certain patterns of articulation that appeal to different visual subsystems is an idea that I will pursue in chapter 4. It also ties in nicely with the MPA, as I will show in section 3.4. Modality is a valuable tool for the analysis of images and visual grammar.

3.3 Intertextuality

Bakhtin's ideas are the precursors of the present notion of intertextuality. As such, intertextuality is an interesting concept in regard to differentiation and visual grammar. However, there is some debate whether on should use intertextuality to refer to an inherent quality of specific texts the fact that especially literary texts often quote other literary sources directly or indirectly or to refer to a general characteristics of all texts (Mai 1991: 31). The inventor of the term, Julia Kristeva, used the term in a way that is closer to the latter view.

The term intertextuality was coined by Kristeva in her essay Word, Dialogue, and Novel first published in 1968. In this essay, intertextuality was only mentioned in passing but not explored further. The importance the term has today, was not foreseen by Kristeva at the time. Word, Dialogue, and Novel discusses Bakhtin's ideas and it was his insight that "any text is constructed as a mosaic of quotations; any text is the absorption and transformation of another" that Kristeva labels intertextuality (Kristeva, 1980: 66). The fact that intertextuality ties in nicely with subsystems becomes clear from the following passage:

First, we must think of genres as imperfect semiological systems "signifying beneath the surface of language but never without it"; and secondly discover relations among larger narrative units such as sentences, questions-and-answers, dialogues, et cetera, not necessarily on the basis of linguistic models justified by the principle of semantic expansion. We could thus posit and demonstrate the hypothesis that any evolution of literary genres is an unconscious exteriorization of linguistic structures at their different levels. (ibid. 66)

Thus, intertextuality for Kristeva is broader than the direct quotations of a text in another text, as it is sometimes understood. In fact it is "the transposition of a system of signs into another system of signs" (ibid. 15). Thus copying the form of, for example, a newspaper article, and all the stylistics associated with it, into a novel is a good example of intertextuality. It is not necessary that the intertextual part of a text is an exact quotation, as long as its form or style is recognisable as belonging to another type of text. This makes different genres of texts excellent vehicles of intertextuality, as genres usually have well-known and easily distinguishable stylistics.

Figure 3.7 - Guitar and Wineglass (Pablo Picasso 1912)

Figure 3.8 - Cadres Make All The Difference (A. Kondurow 1988)

The collage can be considered to be one of the archetypical forms an intertextual representation can take. A collage is assembled from several fragments of text or representations that are recycled into a new representation. Usually, a collage consists only of such fragments, but this is not necessary. James Joyce's Finnegans Wake is a good example of what might be called a literary collage, but the typical collage is a visual representation. Picasso's Guitar and Wineglass (figure 3.7) is an early example of a collage. In this painting, fragments are literary glued together to form a new representation. The painting includes a piece of musical notation, a drawing, wallpaper and a newspaper. All of these have their own distinct stylistic and are easily recognizable as such. At the same time they are brought together to form a guitar and a wineglass as the title implies. By contrast, consider figure 3.8. This painting shares many traits with Guitar and Wineglass it consists of fragments taken from various sources. However this time all fragments are painted and not glued on the canvas. Even the title of the work is a quotation, it is said to be a statement frequently made by Stalin. The artist made some effort to make several elements of the work look different: the postal stamp looks quite different from the portraits of Stalin, the heap of hands is executed with a different technique than the lettering in the central pane. Visual intertextuality of this kind is often marked by distinctive and different patterns of articulation, and is actually akin to local modality differences discussed in section 3.2.

Figure 3.9 - Double Nude (Dianna and William Beckman 1978)

Figure 3.9 is an example of another form of visual intertextuality which is commonly encountered in the visual arts. It is an oil-painting executed with photo-realistic quality. In every sense it follows the techniques of naturalistic paintings of the 19th century and even the subject matter nudes is not uncommon for paintings of that period. However, the poses of the models are not at all like the poses normally encountered in naturalistic paintings. This painting takes the form of naturalistic paintings and uses this to represent something naturalistic painters would never dream to represent, even though most naked people in fact look like this. In that respect, the painting can be seen as a comment on naturalistic paintings of nudes that try to make the nudes look better or try to justify their nakedness.

What is interesting about this last example is that the painting appeals to the naturalistic tradition of painting by a certain quality of representation and subject matter. Even the word 'nude' in title contributes to this. At the same time, the picture opposes the naturalistic tradition by representing the nudes in a rather confronting way[6]. This sets up a double meaning or a deliberately ambiguous representation (in a sense the depicted persons are 'double nude' indeed). Bakhtin and Kristeva refer to this kind of ambiguity with the term ambivalence (cf. Kristeva, 1980: 68-72). Something similar can be said of the newspaper fragment in Picasso's Guitar and Wineglass. The easily recognisable newspaper and the seriousness commonly associated with it are contrasted by the words 'LE JOU' they form; it is quite literary a play on words. Figure 3.8 is also pregnant of ambivalent irony set up by the contrasts of subject matter, style and the combination of various elements in the painting.

As we have seen from the previous examples, (visual) intertextuality often draws upon pretexts from various media; a visual representation can use verbal pretexts, musical pretexts and socio-historical context, to name but a few examples. Such 'transvisual' intertextuality seems to be the norm rather than the exception. This enriches the potential of the intertextual image but also complicates the analysis of such images, because a visual grammar should somehow be able to incorporate all these extra-visual influences.

From this section and section 3.2 we can conclude that visual language actually consists of a set of semiotic subsystems. These subsystems all share a common ground which is the reason why we can speak of visual langue. This common ground sometimes makes it difficult to recognise one subsystem from another, because the subsystems share elements and syntactic constructions, or even borrow them from systems outside visual language. The viewer of an image is greatly influenced by the subsystem (or subsystems) he elects to interpret the image with. On a low level, it can influence the credibility the viewer attaches to the image; on a high level, it can lead to different interpretations of the same image. Some images appeal to more than one subsystem which creates ambivalence, while others employ different subsystems next to each other. The latter can be used to mark symbolic meaning, as was discussed in section 3.2, or create a collage of different image types. In the next section, I will discuss how these effects can be incorporated within the Modular Parallel Architecture, presented in chapter 2.

3.4 Differentiation and the Modular Parallel Architecture

The Modular Parallel Architecture is formally well suited to deal with some aspects of visual modality and intertextuality. One of the most important characteristics of the MPA is that visual information is processed by specialised modules and that the results converge in the conceptual module. In this module, the information is brought together and 'meaning' is established by careful consideration of this information. It can thus be easily imagined that the lower-level modules feed contradicting or supplementary information to the conceptual module. A very clear representation would ideally feed the conceptual module only supplementary information. Most of the time this is not the case and deviant information will need to be suppressed. That the mind is well equipped to block out contradicting information can be easily illustrated by images often used in cognitive textbooks. Figure 3.10 is such an image. One can either see two faces or a vase but seeing them at the same time is impossible. When one sees a vase the black forms becomes the background, and vice-versa.

Figure 3.10 - Faces / vace

The information passed from each module in interpreting figure 3.10 can be represented as follows:

Articulation structure:
[[F1 = black] + [F2 = white]]
Syntactic structure:
[[F1= periphery and background] + [F2 =centre and foreground]]
[[F1 = periphery and foreground] + [F2 = centre and background]]
Lexical structure:
[[F1 = two faces] + [F2 = vase]]

In this case the syntactic module passes two plausible configurations and leaves it to the conceptual module to decide which is more likely. The only input the conceptual module gets that might help it in any way is the lexical structure. Thus it can decide that if [F1 = two faces] than it must be also the foreground. However this contradicts [F2 = vase] because it is not compatible with being background. This is why only one of the two solutions can be prevalent at the same time.

We can make one solution more prevalent by adding more visual markers that support that solution to the figure. For example we can change the articulation of F1 so that it is 'sky-textured' (see figure 3.11). Although it still is possible to see a vase in figure 3.11, following the analyse above I expect it would costs considerably more cognitive effort to suppress the more dominant solution that the 'vase' is actually a 'the sky in the background'.

Figure 3.11 - Faces / sky-textured vace

This example shows how information is processed in a parallel way, and how more visual markers support, oppose or contradict particular interpretations of an image. Experimenting with this can lead to interesting results. An altered version of the visual representation of the MPA can be made to weaken the technological coding orientation of the original figure (figure 3.12), creating a weak form of ambivalence making the image seem less technical and therefore less correct than its original (figures 2.3).

Figure 3.12 - An altered representation of the MPA

All in all, the convergence of information in the MPA is highly compatible with the notion of ambivalence and visual modality. On the other hand, the social and extra-visual aspects of intertextuality are more difficult to incorporate into the MPA. This is largely due to the fact that these aspects are extra-visual and the MPA, as it was presented in chapter 2, only governs visual language. In order to account for these aspects, the MPA needs to be hooked up to other parts of the brain. Basically, there are two ways to do this. The first, and most simple, way is to assume that the conceptual module is also connected to the verbal language architecture and other cognitive functions. The second, more difficult way is to assume that different modules also have interfaces to other cognitive functions of the brain. I will discuss both options below.

If the visual language architecture is connected to other cognitive functions of the brain only through the conceptual module than all extra-visual influences on the use of visual language will be 'top-down'. In effect the extra-visual influences are the result of other, independent (semiotic or social) systems, which have to be learned somehow. In fact, before extra-visual systems can influence the visual system it needs to be learned that such an influence is possible at all. There is much to say for such an approach. In fact, many extra-visual influences on the visual system do seem to be learned. The intertextual qualities of Double Nude (figure 3.9) cannot be 'seen' if one does not know a thing or two about the painting traditions of the 19th century. The same goes for the many intertextual aspects of figures 3.7 and 3.8.

However, the association between different cognitive functions often seems to be quite natural, and not learned at all. Even small children will have only little trouble identifying popular fairy-tale characters, even if they know them only 'verbally'. This implies that there exist links between the various cognitive functions of the brain that do not need to be learned.

If interfaces exist between the modules of different architectures, then this will complicate the study of visual grammar tremendously, for it will become possible that almost anything can influence the individual performance. Sadly, this often seems to be the case. However, this thesis is not really interested in individual performance, and on an average level the influence of such 'trans-cognitive' interfaces seems to be quite limited. At least the number of possibly important 'trans-cognitive' interfaces is limited, otherwise cognitive study of the brain would not have led to a modular architecture. Formally the MPA easily allows for 'trans-cognitive' connections. It is quite conceivable that the syntactic module is connected to its verbal counterpart, and the lexical module might even be shared.

For the sake of simplicity, I will assume that most 'trans-cognitive' connections are of the first type; top-down connections through the conceptual module which do not influence the visual architecture directly. I will assume an influence on the visual system of the verbal system through the lexical and the syntactical modules. With these assumptions, I will deliberately omit some intertextual aspects of visual grammar, but I think that the omitted aspects are not the most important aspects; visual grammar can be build without taking these into account.

3.5 Subsystems and Visual Grammar

In this chapter I have investigated the notion of visual subsystems and the way they function in visual language. The idea of visual subsystems follows from the literary work of Mikhail Bakhtin, who has studied (verbal) language with a special interest in its social and cultural context. I have tried to show that these subsystems play an important role in visual modality and intertextuality. The MPA model has also proven to be well equipped to incorporate at least some characteristics of a visual language. But what is exactly gained from such an approach to visual language?

I would say quite a lot. Most importantly, it opens up a new way of testing visual grammar. This is not an unimportant asset for a semiotic model that deals with visual representations, for many semiotic models for visual representations have difficulty to deal with the highly subjective reception of the images they study. Often, it is all to easy to disagree with the interpretation offered as a result of a visual analysis, which makes it very difficult to even establish the most basic tools for visual analysis.

Textual analysis has been the traditional method of semiotics since the 1960s. This method focuses on the meaningful devices in one text and tries to build a semiotic system based on these. It moves from one text to another and tries to improve the semiotic system with each additional iteration in this process. This method has proven its value in many analyses of verbal texts, even though it has its methodological limits. J.K. Sheriff pointed out the limits of the method: meaning is the product of both the text and its reader. Since there exists an unlimited potential for different readers and texts, an enterprise that tries to unravel the nature of language by looking at meaning is almost impossible (Sheriff 1989: 21-22). This is indeed the pitfall into which many visual studies have fallen. It is often easy to disagree with the whole presented theory simply because it is easily justifiable that one does not agree with the proposed meaning of such-and-such devise. Meaning, and especially visual meaning is simply too elusive and subjective for this kind of analysis. Sonesson also argues that textual analysis is to much dependent on the initial texts it studies. It does not leave enough room for what he calls 'imaginative variation' (Sonesson 1989: 20). Or to put it within the context of this thesis: the dependence of textual analysis on existing texts makes it difficult for this method to truly account for the discrete infinity of language.

The traditional psychological method, also described by Sonesson, tries to isolate specific elements of a system and to study these elements empirically. This kind of study is better equipped to study low-level cognitive characteristics of the visual system, such as the specific functions of the eye. To study the higher-level, cultural aspects of visual language it is less suited. However, the cognitive method which cumulates in the simulation of the architecture of the human brain, which can be seen as an extension of the psychological method, has been very successful for linguistic studies. Unfortunately, it would be beyond the scope of this thesis to go deeper into this method.

What is needed, is a method that can be employed to set up reasonably objective tests, without loosing itself into detailed elements of visual grammar. Analysis and classification of visual subsystems on the basis of graphic means, seems to be an ideal way of doing just this. Throughout this chapter, I have tried to make clear that visual subsystems do exists and can be differed on particular stylistics on the particular use and patterns of graphical means. Such an analysis can be quite similar to the statements I have made in regard to the film posters of Ghost World and Lord of the Rings (see section 3.2). However, the analysis done in section 3.2 was rather premature. In chapter 5, I will focus on genres of images that have been well documented and try to associate these with particular use and patterns of graphical means.

In the next chapter, I will present a basic visual grammar that has been build with regard to the conditions discussed so far in chapters 1,2 and 3 and with an eye on the tests I will conduct on them.


[1] This claim was stronger when Bakhtin wrote his essay than it is now. In this post-modern age many centrifugal forces have become accepted and cultivated. An example is the strengthening of local identity by publishing books in local dialects.

[2] Kress and Van Leeuwen point out that a semiotic theory of modality, cannot show whether a given proposition is true or false, only that it is represented as either true or false (Kress en Van Leeuwen, 1996: 159).

[3] It can even be argued that the natural high modality of images has been eroded by computer enhanced imagery

[4] Such as may be encountered in pencil drawings or surrealistic paintings, respectively.

[5] Kress and Van Leeuwen argue that visual symbolism can be identified because 'symbolic attributes' are marked either by being pointed at, convention, salience or simply because they "look out of place" (Kress & Van Leeuwen, 1996: 108). Although the first two characteristics are applicable, the latter two are arguably weaker. Especially these characteristics can be made stronger by the notion of local shift in modality.

[6] The absence of context is an aspect of the painting that also contradicts the natural coding orientation. However, in this picture I do not take it as a very strong or marked deviance from the natural norm, for there is a context which might be a wall. In the poster of Ghost World (figure 3.4), by contrast, the context is totally omitted, leaving only a white plane.

[Contents] [Previous: Chapter 2 Architecture] [Next: Chapter 4 Rules]