Online Magazine of the Visual Narrative - ISSN 1780-678X
|
Online Magazine of the Visual Narrative - ISSN 1780-678X |
|||||
|
||||||
|
||||||
|
Issue 7. History and Theory of the Graphic Novel |
||||
The Possibility of Minimal Units in the Filmic Image Part 2 |
||||
|
Author: Sophie De Grauwe Abstract (E): The question of minimal units in the filmic image has been extensively discussed in traditional semiotics (e.g. Eco, Metz). Recently, the subject of minimal units has come up again in the work of Groupe µ, who propose an analytic framework of the still image in which minimal units play an important part. This paper evaluates the arguments and models of Metz, Eco and Groupe µ from the perspective of social semiotics. What will be discussed is the possibility of minimal units in relation to the characteristics of the cinematographic system. To establish these characteristics, the role of (1) abstraction and discreteness, and (2) iconicity, economy and arbitrariness in the cinematographic system are discussed. The paper concludes with a note on the integration of cinematographic grammar. Abstract (F): La questions des unités minimales de l'image filmique a longtemps fait l'objet de discussions très pointues dans la sémiotique traditionnelle (représentée par Eco et Metz). Récemment, la question a refait surface dans les travaux du Groupe µ, qui a proposé un nouveau cadre analytique de l'image fixe dans lequel la notion d'unités minimales joue un rôle clé. Cet article examine les arguments et les modèles de Metz, Eco et du Groupe µ de la perspective de la sémiotique sociale. Plus concrètement, il s'attachera à discuter la possibilité de décrire des unités minimales en relations avec les caractéristiques du système cinématographique. Ces caractéristiques seront établies à partir d'une réflexion plus large sur deux éléments: d'abord le rôle de l'abstraction et du caractère discret des signes, ensuite les concepts d'iconicité, d'économie et d'arbitraire. L'article se termine par une note sur l'intégration de la grammaire cinématographique.
|
|||
|
|
3. Discreteness/AbstractionVerbal language and images are often presented as opposed on the level of discreteness. For example, Lev Manovich states that verbal language is discrete on most scales, and he gives the example of ranking (sentences, consisting of words, in their turn consisting of morphemes, each of them being discrete units). Photographs, on the other hand, do not have discrete units. And the units of modern media that are discrete (film frames, halftone dots) do not act as units of meaning. Metz also makes a difference between the discreteness of verbal language and the non-discreteness of the filmic image. When you analyse the filmic shot, he says, you pass from "un ensemble non-discret à des ensembles non-discrets plus petits" (Metz 1971b: 119). The elements that make up the filmic shot are not only unlimited in number, but also "in their nature". Discrete elements can only be found on a level of higher dimension, e.g. shots which combine to sequences. In order to be able to judge these claims, we should first of all look at discreteness in verbal language. It appears that we should make a difference between physical and systemic discreteness. Before a semiotic analysis, spoken utterances form a continuous whole, with few pauses to establish physically discrete units. Words are linked in a continuous phonic chain. The discreteness shown by written language already constitutes a level of abstraction not present in spoken utterances, the primary object of linguistics. Verbal discreteness must therefore appear on another level. This is the systemic level. Only after linguistic analysis are words and other units established. However, even on the systemic level not all units are clearly "discretable". In fact, as Halliday has pointed out, the claim that "each structural unit has clearly defined boundaries" (Halliday 1994: 35) is completely true only for the ideational level. On the interpersonal and textual levels, many units are only tentatively determined. That is because on these levels prosodic features play an important role. For example, the textual meaning of "New" (i.e. the marking of an element in the clause as new information) is realised by tonic prominence. The textual meaning of "Given" usually precedes the New. The boundary between Given and New is often hard to determine. The prosodic features of stress, length and tone establish contrast, not by absolute positions (which I shall call absolute contrast), but by different relative positions (relative contrast). For example, the actual pitch or tone is dependent on the speaker's individual voice. As individual voices vary considerably in pitch, a system of relative positions is necessary to make pitch a useful feature. The prosodic system is a gradational system where meaning is realised by contrast between different degrees of a feature. As a result, the establishment of discrete units is more problematic because there are no clear boundaries. In the phonemic system, gradation is reduced. With certain phonemic features, contrast is established by realising a distinction between the feature and its absence, e.g. the feature of voice. Other features establish contrast by realising distinctions between degrees of a feature, e.g. openness, height, etc. The feature of articulatory position is an intermediate case. Robins states that, for this feature, contrast is established by realising a distinction between different positions, making each position into a different feature (e.g. alveolar, uvular, etc.). However, since these features are mutually exclusive (dental and alveolar cannot be realised together), one could also speak of one gradational feature. Still, the possibility to locate quite specific positions (like alveolar, dental), in contrast with more vague determinations as high-middle-low, seems to make this feature slightly more determined by absolute contrast than the feature of height. Overall, the combination of these different contrastive devices establishes the phonemic system as more reliant on absolute contrast than the prosodic system. As a result, the phonemic system permits clearer boundaries between units [1] . We can say that, in the filmic image, the concrete image is continuous. The physically discrete units (the frames) are semiotically irrelevant (see section 2.2). If there is discreteness, it would have to be situated on the abstract, systemic level. Groupe µ's analysis makes it possible to postulate such a systemic discreteness. In the discussion of systemic discreteness, I will first comment on the difference between the ideational level and other levels, as in the discussion of verbal language. Then, I will comment on the system on the level of the minimal units, the visual parameters. But before that, we need to introduce another concept used by Groupe µ, viz. contour. The perceptual and cognitive analysis of images occurs through an exaggeration of both similarities and differences in the image. Eventually, the integration of all variables of the image through this process leads to the constitution of what social semiotics calls represented participants, separating them from the background (the figure/ground dichotomy of psychology). This goes hand in hand with the creation of a virtual contour. A contour is precisely the result of this process of perceptual and cognitive exaggeration. It exaggerates the difference between figure and ground and it emphasises the similarities constituting the figure. I call the contour "virtual" because it is an abstract entity, i.e., it is not a concrete given present in the image, but rather a perceptual and cognitive creation. However, its creation is often stimulated in the filmic image by the use of lighting, the use of colour, etc. The discussion of contour indicates that, as in verbal language, units are more clearly discreted on the level of representation than on the other levels of the cinematographic system. Groupe µ's entities are established by contours and by frameline. For example, /full human body/ is established by its contour and by the full inclusion of the body in the frame. /Human head/ is established by its contour and by the framelines, cutting off the rest of the human body. As discussed above, contour is an abstraction. Frameline, on the other hand, is a concrete discretionary tool both in physical and textual terms. So, in contrast with verbal language, signifiers on the representational level are discreted by abstracting systemic means as well as by concrete textual means. As a result of this combination, there is a lot of mobility between entities and sub-entities. What is an entity in one image can be a sub-entity in another. But this mobility does not seem to make the discreteness of entities problematic. Of course it is possible to create images in which it is difficult to discrete entities. But this will probably be due to a low level of lighting or a poor image quality, not the use of framing [2] . Representational processes, on the other hand, are less clearly discretable. They can be discreted on the syntagmatic and paradigmatic level through features like movement, duration of movement, direction, etc. In contrast with verbal language, however, they are always connected to represented participants on the textual level. They can be discreted and abstracted from these participants on the basis of features like movement, but this involves a more profound abstraction than necessary in verbal language. This is related to the difference in concreteness between verbal language and the filmic image on the level of the instances (see further below). On the interpersonal and compositional levels, the contrasting systems in use are gradational systems: contrast is established by making a distinction between degrees of a feature. For example, social distance, realised by size of frame, and salience, realised by, among others, dimension, are gradational features. As a result, it is not possible to establish clear boundaries. On the level of visual parameters, the contrasting system is again a gradational system. For example, marques of "form" (e.g. /circularity/) are determined by the parameters of /dimension/, /position/ and /orientation/. These features establish contrast by realising distinctions between degrees of the feature, e.g., between /big/ and /small/ there is a scale of gradation. As mentioned in section 2.3, we should point out that there is no black-and-white binary distinction between absolute and relative contrasts. Relative contrast may be less relative in some cases than in others, and absolute contrast is not as absolute as is sometimes maintained. As regards the former, the feature of /brightness/, for example, is a gradational feature, i.e., whether something is light or dark is a matter of degree and is determined by its relative position in comparison with other elements. However, it also displays some absolute positioning, as the viewer also has some pre-established idea about what is /light/ and what is /dark/. As regards absolute contrast, the "absoluteness" is attenuated by prototypicality and context-dependency. The categories established by absolute contrast are prototypical categories which means that the so-called clear boundaries between the categories are an abstraction based on the clear boundaries between the central members of the categories. There are, however, also peripheral members of the categories which render the boundaries between the categories vague. Also, context-dependency in the use of language allows adaptations of both signifier and signified according to situational and textual contexts. These features which close the gap between language as a system and language in use are responsible for the flexibility of language and the possibility of change. They are at the basis of the systemic-functional approach which investigates meaning as a process, rather than studying a code of fixed signs. The discussion so far makes it clear that filmic image and verbal language are similar on the level of the concrete image and utterance as regards discreteness. In both media, the text is continuous at this level. In both media, the viewer/listener has to go through certain discretion and abstraction processes to establish discrete units. But there is a difference on the level of systemic discreteness. Verbal language makes more use of absolute contrast (which establishes clearer boundaries between units) than does the filmic image. However, as the role of absolute contrast is not as prominent in verbal language as is sometimes assumed (there is also extensive use made of relative contrast), the difference between verbal language and the filmic image on the level of discreteness is less pronounced than is often claimed. But there is another difference between the two media that constitutes a more significant distinction. Both in verbal language and in the filmic image, the instance is more concrete than the type, on the level of content as well as expression. However, there is a difference regarding referential abstraction. In verbal language, the concretisation of the referential meaning of the instance is realised by the context of situation and the textual context. In the filmic image, the concretisation is established not only by these contexts, but also by the instance itself, its concrete expression. Unlike verbal language, the concreteness of the instance on the level of expression is directly linked with the concreteness of its referential content. The iconic (and indexical; see section 4.4) referential link between referent, signifier and type entails that the cinematographic instance is bound to the same level of concreteness as the referent (although there are some ways to diminish the level of concreteness of the signifier; see further below). This is in contrast with verbal language where the arbitrary relation between signifier and signified ensures the necessary distance between referent and signified, so that the linguistic instance is not bound to the level of concreteness of the referent. The iconic relation between referent, signifier and type not only leads to an extreme concreteness of the instance in relation to its type, but also to the extreme concreteness of the instance as a result of the fusion of different types in one instance. Abstraction of a concrete given not only leads to the selection of its pertinent features into a type, but also to the analysis of the concrete given into several types rather than one type, i.e. to the establishment of several selections of pertinent features such as a represented participant and a process. For example, in an image of a man eating, the participant "man" and the process "eating" are integrated, while a verbal depiction of this scene will provide an analytical differentiation of the scene into "man" and "eating". As regards the image, the analytical distinction of "man" and "eating" is facilitated by the image because the movement of the jaws permits an interpretation as "process + participant" rather than just "participant" (as movement is linked to process). But this is just an analysis on the basis of the concrete image: on the level of the concrete image itself, process and participant are integrated. Therefore we can say that a two-fold abstraction is realised in verbal instances which is not realised in cinematographic instances: abstraction of a concrete given (1) through selection of its pertinent features so that a type is established; (2) through its analysis into several types rather than one [3] . As a result, verbal language can bypass a significant part of the process of abstraction that is necessary for the filmic image. As mentioned, this is a result of the arbitrariness of verbal language. Arbitrariness induces a distance from reality which makes an arbitrary system excellently suited for the manipulation of reality. However, this does not mean that the lack of arbitrariness leads to the impossibility of manipulation of reality. Further in this section and in section 4, we will study how the filmic image deals with the absence of arbitrariness. Also, this difference regarding abstraction does not mean that the filmic image as a semiotic medium only functions on the concrete level. What it does mean is that verbal language bypasses the process of abstraction from the concreteness of the referent: concrete verbal instances already establish a certain degree of abstraction. In the filmic image, on the other hand, instances have roughly the same level of concreteness as the referent, so that each time a much more elaborate process of abstraction has to be gone through than in the case of verbal language. In other words, the distance between the level of the (concrete) instances and the (abstract) types is much greater in the filmic image than in verbal language. But just because the instances are more concrete this does not mean that there is no abstract level at all. A certain level of abstraction is necessary to understand the filmic image: without it, there could be no recognition of concrete instances. Recognition entails a comparison with previous instances that are similar or dissimilar. Recognition of similarities involves a process of abstraction. In the filmic image, the process of abstraction relies upon visual perception and cognition. However, the filmic image has several tools to enhance certain abstractions. Grouping, distancing and stereotypical representation can all enhance abstraction of represented participants, as mentioned by Van Leeuwen (Van Leeuwen 2001). Lighting and focus can also enhance abstraction of represented participants as well as of the setting in general. The analysis of the continuous image into discrete units is enhanced by the same elements, in addition to the frameline. The concreteness of cinematographic instances has often hidden the possibility of abstraction within the cinematographic system. Metz, for example, has often emphasised the discursive nature of the filmic image. Although gradational semiotic systems (which are, as mentioned above, also present in verbal language) let discourse play a bigger part, this does not mean that there is no abstract systemic level. The system plays a role as a resource for constituting meaning, just as can be observed in systems of absolute contrast, but instead of offering a limited number of meanings realised by absolute contrast it offers a large number of possible meanings realised by relative contrast. Not only does Metz focus on the concreteness of the filmic image, but he also emphasises the systemic basis of verbal language. This is because his focus is on words. Words, as abstract systemic units, are taken as the basis of linguistic grammar, as opposed to systemic-functional linguistics which takes the sentence (or the clause) as its basic unit. The focus on words [4] obscures the concreteness of verbal language. This not only because it diverts the attention from syntagmatic structure, which explicitly brings in the discursive level of language, but also because of the ease with which words-in-discourse can be disregarded. As mentioned before, the meaning of a word in an utterance is more concrete than its meaning in the system. However, since this concreteness of meaning is only established through the context of situation and text, and not through its expression, this concreteness is less obvious. The differences in expression have normally no effect on their referential meaning. As a result, the establishment of words as abstract entities is facilitated, and it is easy to use concrete words in utterance as equivalents of their abstract types, i.e., to use the instance for the type. This is still enhanced in written language, where differences in expression can be abstracted from more easily. In the filmic image, on the contrary, differences in referential meaning between different instances of one type are noticeable in their expression. In addition, instances refer to several types at the same time (as mentioned above). Therefore it is not possible to use the instance for the type, and it is less obvious to talk about the filmic image in an abstract way. Thus we can conclude that the polarisation of filmic image and verbal language, which frequently occurs in traditional semiotics, is enhanced by two factors. Firstly, the extreme concreteness of the filmic image (on its discursive level) obscures its abstract level, secondly, the less concrete nature of linguistic utterances (in comparison with cinematographic discourse) obscures their concrete nature (in comparison with the abstract system). To conclude, we can say that the differences between the visual parameter system (i.e. the system of minimal units of the image, viz. chromemes, formemes and texturemes) and the (linguistic) phonemic system are found on two levels. Firstly, phonemes are arbitrary, while visual parameters are not. This difference results in a difference in referential abstraction on the level of discourse between the two systems. Secondly, the visual parameter system uses a gradational system, while the phonemic system is founded on a mixed system with both absolute and gradational features (although the latter are limited by economy). This results in a difference in systemic discreteness, although the difference in discreteness is less pronounced than is often thought (especially when we look at the discreteness of represented participants). The differences mentioned are related to referential meaning, as it is here that arbitrariness matters most. However, the study of how meaning is constituted would not be complete without giving interpersonal and textual meaning due attention. On these levels, verbal language uses a type of contrastive system different from the phonemic system. The prosodic system, the importance of which is highlighted by Halliday, functions as a gradational system. In such gradational systems, iconicity comes in and relativises the arbitrary nature of verbal language, thus bringing it closer to the filmic image on these non-referential levels. In the next chapter, the role of iconicity in verbal language and the filmic image respectively is studied, and the arbitrariness of verbal language is relativised. 4. Iconicity, arbitrariness and economy
4.1. Traditional semioticsIn traditional semiotics, the iconicity of verbal language is often treated only marginally, as an exception in an arbitrary system. The second articulation of verbal language is emphasised, and Metz even defines "the preferred form of signs" as "arbitrary, conventional, and codified" (Metz 1974: 76). Thus, as we will see, a large part of verbal language appears not to have the "preferred form of signs". This focus is due to the structuralist background of traditional semiotics, where the "code" is defined as a synchronic, static, arbitrary correlation of the systems of signified and signifier. As a result of this focus, a lot of effort is dedicated to finding partial arbitrariness in the filmic image, in order to be able to justify in some way the postulation of a cinematographic semiotic system. A good solution to this problem is to separate the concept of arbitrariness from the concept of conventionality, as is the case in social semiotics (see section 1.1). To function in a community a semiotic system needs to be conventional, i.e., it must be socially accepted as a system. Iconic motivation can also be conventionalised. Arbitrariness is a good tool for manipulating reality, but as verbal language also shows, iconic motivation does not necessarily prevent manipulation of reality. In addition, as will be developed later, the cinematographic system has found other ways to circumvent its disadvantage. In A Theory of Semiotics (1976), Eco provides an extensive discussion of iconicity. He claims that iconicity is actually an umbrella-term for (too) widely differing phenomena, and as a solution he proposes to get rid of the notion of iconic signs. It would be better instead to concentrate on the procedures governing signs, or modes of sign production. The study of these procedures shows that iconic signs cannot constitute a unique category since the procedures that govern them are not exclusive to "iconic signs". The elimination of the category of iconic signs is also necessary to get rid of the fixed and stable correlation between discrete signifiers and signified, which is not suited for the contextual and transitory constitution of iconic signs. I agree that traditional semiotics' focus on signs instead of processes leads to a fixation of semiotic practice. However, I do not think that this fixation is only a problem in relation to iconic signs, but also in relation to the linguistic system. To solve this problem of fixation, Halliday's systemic-functional linguistics has postulated the linguistic system as a flexible system and focuses on processes of meaning constitution instead of fixed relations and signs. The rejection of the iconic sign also entails a rejection of iconicity. One of the reasons Eco gives for this is that the notion is an umbrella-term for many differing phenomena. Firstly, he objects to the use of the term "iconicity" for similarities which are not geometrically or spatially "iconic" (i.e. for similarities other than those determined by the geometrical term of "similitude"), this being "a mere metaphorical license" (Eco 1976: 199). This applies, for example, to Peirce's identification of logical diagrams as iconic signs (see Figure 7 for an example). Pierce claims that the correlation between conceptual and spatial relations as displayed in logical diagrams is iconic. Eco claims that the correlation between the two kinds of relation is a convention which posits a proportional analogy, and that to call this complex isomorphism iconic is merely metaphorical. However, after making clear that we should focus on relations between expression and content, instead of between expression and object, as Peirce does, and while ascertaining that any geometrical iconic relationship requires transformations, and thus conventions, Eco is prepared to consider models of abstract conceptual relations as virtual space (but without disclaiming his earlier judgement with regard to these models). Then, Eco proceeds by considering other phenomena of assumed iconicity, and rejects some of them, for example specular reflections and replicas, as cases of iconicity. Eco also stipulates a gradated continuum of iconicity from minimal convention, such as congruences, to maximal convention, as in stylisations. Finally, Eco decides that the notion of "iconicity" is too wide-ranging to be tenable, and he rejects it together with the concept of sign.
The problem that the term "iconicity" is used for phenomena that are not iconic can be resolved quite simply: by excluding them from the category of iconicity. Of course, you need an accepted definition of iconicity to do this. But this is exactly what Eco does in the first stage of his argumentation: on the basis of a definition he excludes certain phenomena from the category of iconicity (see above). What seems to bother Eco most are the apparent inconsistencies displayed by "iconic signs". They are motivated and conventional at the same time; they vary in their degree of iconicity; and the establishment of relative contrast in texts seems to undermine the systemic basis of iconic signs. These objections originate in the structuralist background of traditional semiotics. Firstly, if "convention" is separated from "arbitrariness", as in social semiotics, there is no contradiction between "motivation" and "convention". As a matter of fact, there can be motivated conventions not only in visual media, but also in verbal language (see sections 4.2 and 4.3). Secondly, social semiotics accepts gradational scales and relative contrast as relevant and useful elements of semiotic systems. They can be used to deal with semiotic reality in a differentiated and flexible way, and again, are also used in verbal language. In traditional semiotics, on the other hand, a binary and black-and-white view of semiotic reality prevents the acceptance of such flexible tools. Text and system are completely separated and opposed, preventing the acceptance of relative contrast as a semiotic mechanism. They base their argument on the linguistic system, which would be the supreme exponent of such a binary view. However, as we have seen and as we will see further, the linguistic system also makes use of gradational systems, undermining the binary view of traditional semiotics. So, in a social semiotic perspective, we can hold on to the notion of iconicity, as the apparent inconsistencies pointed out by Eco are not considered contradictory. What would be included in the definition of iconicity? Eco is quite clear about some phenomena, like specular reflections, doubles and replicas, which he excludes from the category. With regard to logical diagrams, he is less clear. On the one hand, he calls it "metaphorical licence" to consider them iconic, on the other hand, he includes them in his discussion of iconic transformations. It is not clear if the latter means that his first statement is not valid anymore, as it is never clearly contradicted. If it is supposed to be still valid, it is not clear how he reconciles the two statements. Let's consider his first statement, that only geometrical similitude can establish iconicity. The advantage of this restriction is that geometric similitude is a precise and scientific concept, founded on observable spatial parameters. Also, "icon" originally refers to visual similarity. So in that sense the use of the notion of iconicity for similarities other than visual similarity is indeed a metaphor. However, although I do not object to a specification or even a restriction of the notion of iconicity per se, I do not find etymology a sufficient basis for this. Meanings of words change all the time (through metaphors for example), and there is no real reason why the first meaning would be the "best" meaning (in this case: iconicity as based on visual similarity rather than similarity in general). There remains the need for a scientific criterion. Geometrical similitude can be defined as "the property shared by two figures that have equal angles and sides that are proportionally equivalent" (Eco 1976: 196). Eco states with regard to this definition that "the criterion for similitude is based on precise rules that select some parameters as pertinent and disregard some others as irrelevant". So on the one hand, there are precise rules, on the other hand, these rules are conventional. If we accept logical diagrams as iconic, the rules will still be conventional, but they may be less clear. For Eco, this does not seem to be very desirable, and he sees it as contrary to the "practicability" of iconicity as a concept. In order to comment on that, let's examine another definition of iconicity. In (functional) linguistics, the importance of iconicity for the constitution and functioning of the grammatical system has been accepted. In Natural Syntax (1985), John Haiman bases his concept of linguistic iconicity on Peirce's diagrammatic iconicity. Like the latter, linguistic iconicity concerns similarity between the structure of a certain meaning and the structure of its corresponding form. John Haiman defines iconic motivation as follows: "Given two minimally contrasting forms and the information that they differ with respect to a single semantic feature, a language learner who is unfamiliar with the forms (or perhaps even with the language) will be able to assign to each form its appropriate meaning. He will be able to do this on the assumption that the difference in form will in some respects be an icon of the difference in meaning." (Haiman 1985: 71) As we can see from this definition, we are not presented with exact laws but with human judgement. In order to make the latter semiotically relevant, i.e. socially confirmed by convention, I think it is best specified as social judgement. Elsewhere, Haiman declares that true identification between different categories is made "when they share distinctive, or essential, properties" (59). This is similar to Eco's assertion that "one decides to recognise as similar two things because one chooses certain elements as pertinent and disregards certain others" (Eco 1976: 196). What is made explicit in Eco's definition is that similarity is "in the eye of the beholder", i.e., these judgements are made on the basis of (likewise socially conditioned) human perception and cognition. This is also true for geometrical similitude but in that case the category is explicitly restricted to a scientific context by stating prescriptive rules. Maybe it will be possible, with further investigation, to state some rules as to what is perceived as similar in a social and semiotic context, but if so, they will probably be less clear and exact than the geometrical rules. This is due to their context of use. Geometrical similitude is a scientific category functioning in a specialised scientific context which demands clear and prescriptive rules. General semiotic similarity functions in the wider, non-specialised social area of communication and representation which allows for, and even demands, more flexible, not necessarily logical, rules. Eco's pursuit of scientific exactness fits perfectly into the programme of traditional semiotics', and is certainly admirable. However, Haiman has pointed out two major problems with this approach. Firstly, the search for consistent and predictive rules can result in a neglect of the actual empirical data. Secondly, there is a fundamental difference between the laws of physics and the laws of social processes like language. The former are not modified by use, whereas the latter are. In this case, the use of the notion of iconicity as a socially and cognitively determined and flexible category, functioning in the social reality of language, is to be preferred to the application of a stringent scientific category such as geometrical similitude to the domain of a flexible social reality.
4.2. Linguistic iconicityAccording to John Haiman, the arbitrariness of language is situated at the level of the lexicon, while iconicity is situated at the level of grammatical relations. Words are arbitrary symbols; the grammatical rules governing them are diagrammatically iconic. This distinction is already suggested by Saussure. However, at the level of grammatical relations, iconicity is opposed by the principle of economy. Economy is a limitation of the number of signs and results in, and is motivated by, a greater simplicity and manipulability of the linguistic system. Haiman claims that there is an inverse relation between economy and iconicity: the more iconicity, the less economy there is, and vice versa. This statement is refined by the distinction between two kinds of economy: paradigmatic economy and syntagmatic economy. Paradigmatic economy is compatible with iconicity. It "economises on the inventory of signs within a system" (Haiman 1985: 158). The fewer signs there are in a system, the more periphrases, circumlocutions, etc. will be necessary to express what a more elaborate system would express in words. Combinations of words will be more transparent than single words, as already stated by Saussure [5] . Syntagmatic economy, on the other hand, is opposed to transparency. It is the tendency "to economise on the length or complexity of any utterance or message" (ibid.). The less complex an utterance is, the less transparent it is. These competing motivations are responsible for inconsistencies in the linguistic system and the dynamic nature of linguistic rules. Linguistic iconicity as conceived in functional linguistics differs in various ways from the many conceptions of linguistic iconicity throughout history as presented in Genette's Mimologiques (1976). Firstly, most defenders of linguistic iconicity in the preceding centuries focused on sound symbolism or phonological iconicity. Functional linguistics, however, accepts the arbitrariness of (single-morpheme) words. Linguistic iconicity is to be situated at the level of grammar. Secondly, it differs from the historical conceptions of syntactic iconicity presented by Genette. The iconic motivation of word order was the subject of heated debates over several centuries. Mostly, one word order was favoured over the other as the right reflection of the structure of thought. Cases were made for the acceptance of either French or Latin word order as the right or natural word order. That is to say, iconic motivation was ascribed necessarily universal justification power, and variation was mostly labelled as wrong or unnatural. Only rarely was it linked to cultural diversification, so that different word orders would express different culturally determined ways of thinking, without one being better than the other. Present-day functional linguistics tries to account for variation in several ways. Haiman does not focus on the link between language and culture but on competing iconic motivations. For some phenomena, it is possible to have competing iconic motivations, for example the position of focus. One tendency in languages is to have "given" information first in the sentence and "new" information later. However, another tendency is to put the focus first in the sentence (as the speaker wants to express first what is most prominent in his mind). These tendencies compete, with variation between languages as a result. To confront the criticism that the result cannot be distinguished from arbitrariness, Haiman claims that, if there are competing iconic motivations, then there is variation between languages, and if there are no competing motivations, then there is near-unanimity between languages. For example, the competing motivations for focus location result in variation between different languages. But if a language has coded grammatical focus in a specific position, then the interrogative pronoun will be attracted to this position. This is a universal tendency, arising from the absence of other, competing motivations for the motivated relation between focus and interrogative pronoun. Also, Haiman proposes a means to demonstrate the existence of competing motivations or tendencies. As has been said, if two tendencies compete in one structure, there will be variation. Now if a supposedly related structure has the same variation, then what is demonstrated is both that the two structures are related and that the two tendencies exist. For example, the prototypical coordinate sentence "S1 (and) S2" displays the same kind of variation as the prototypical conditional sentence "If S1, S2". On the one hand, S1 often has the same tense and mood as S2, in both structures. On the other hand, there is often only one marking for tense and mood in S1 and S2 if one is sufficient. Haiman claims that "the existence of the same kinds of variation in related structures can be plausibly attributed to the existence of the same pressures" (Haiman 1985: 252f.), the pressures being the tendency to create symmetry and the tendency to destroy it for economic reasons. Which tendency wins out in a given language, or within a subsystem of a given language, depends on a variety of reasons. Firstly, there are some hierarchical relations between the tendencies, so that, in given circumstances, one tendency has more weight than the other. Secondly, the outcome can be decided on the basis of structural factors. Thirdly, it can be determined by social factors. Social factors are for example popularity, prestige, and sense of humour of the resulting expression. An example of a structural factor is the tendency in linguistics to avoid centre embedding, and thus discontinuous constituents. For example, the sub-standard expression "Now we'll be able to see what each other is up to" may be prevented from spreading or surviving because of this tendency, which avoids the use of "each...other" in this expression. Haiman also mentions that sometimes there may be a link between language and culture, which may account for what tendency survives in a given language. This is elaborated on by Halliday. He claims that there is a mutually constitutive role between grammar and culture. There are only some universal features, one of which is "the ‚metafunctional' hypothesis": "in all languages the content systems are organised into ideational, interpersonal and textual components" (Halliday 1994: xxxiv). But the ways in which these metafunctions are realised differ between languages. As for the link between culture and grammar, Halliday emphasises that caution is necessary. We should not link one particular feature of a grammar with a particular aspect of culture; rather, we should link one such feature with a large number of other, very general grammatical features, look at the reasons within the grammar why this feature is favoured, then "relate it to the patterns of language use" in its society, "the historical changes that have taken place" in a certain period, and "the ideological systems that underlie them" (Halliday 1994: xxviii). All this is necessary to avoid ad hoc explanations. Finally, another objection to grammatical iconicity mentioned in Mimologiques is that the only plausible instances of grammatical iconicity pertain to marginal phenomena in the grammar of a language, for example, the reflection of the temporal succession of events in the word order. However, the notion of grammatical iconicity has evolved away from the discussion of "which word order is best", developing functionally motivated explanations for the existence of motivations (as explained above). This has allowed the extension of iconic motivation to the whole of grammar: from asymmetrical relations to symmetrical relations; to other word orders and positions (see further for some examples); to the iconic motivation of clause reduction (viz. conceptual closeness); to word classes (the division of words into verbs and nouns matches "the analysis of experience into processes and participants" (Halliday 1994: xviii)); etc.
4.3. Iconicity, arbitrariness and economyAccording to me, arbitrariness and iconicity are linked to different meaning systems. Iconicity typically coincides with gradational systems, i.e. with relative contrast, while arbitrariness typically works with systems of absolute contrast. If we look at verbal language, we see that iconic relations usually involve gradational features. Stress, for example, which is iconically linked with salience, is a gradational feature. Other examples are conceptual distance, social distance, and complexity. Haiman claims that, the greater the conceptual distance between two concepts, the further apart they will be realised in an utterance. He proposes a scale of decreasing linguistic distance (see Figure 8).
In Figure 8, # stands for word boundary, + stands for morpheme boundary. "[d] represents analysis; [e] agglutination; and [f] synthesis of the morphemes X and Y into a single morph - which may be, but is not necessarily, identical with either X or Y." (Haiman 1985: 105) An example of the iconicity of linguistic distance can be seen in the distance between verb and object. Greenberg has formulated a universal law which states that there is no language where the "morphological bulk" of the direct case affix exceeds that of the oblique case affix. Haiman interprets morphological bulk as the number of syllables. The linguistic distance between verb and direct and oblique object can be related to their conceptual distance: direct objects are directly affect, while oblique objects are not. For example, in Spanish "contestar la pregunta" ("answer the question (acc.)") means "to succeed in answering the question", while "contestar a la pregunta" ("answer the question (dat.)") means "attempt to answer the question, but not succeed" (Haiman 1985: 137, original italics). Haiman also states as a principle that "the more polite the register, the longer the message". The length of a message serves to obfuscate the message, and it can go as far as to make it almost unintelligible. As a result, the "addressee is protected from the contents of the message from a social inferior" (Haiman 1985: 154). The length of the message consequently signals and preserves social distance. In verbal language, however, gradational scales are often limited by economy. For example, difference in number is usually expressed by difference in length. The scale is in principle limitless, but usually there is only a differentiation of two or three numbers, thus extremely reducing the potential gradational scale. In addition, the third number is often not expressed by a longer form than the second number. This is due to the principle of economy. Absolute contrast is necessary in an arbitrary system and is the result of economy. In iconic systems, meaning can easily be derived from relative contrast because of the iconic motivation underlying it. In arbitrary systems, there is no iconic relation between meaning and form, so that, to achieve some sort of stable ground for meaning, absolute contrast is necessary. Absolute contrast is based on a selection of relatively stable points on a gradational scale (keeping in mind the prototypical nature of these points, or rather, categories). As a consequence, absolute contrast entails economy: the number of possibilities (the selected points) is limited. This economy is necessary in verbal language, more so than in the visual medium. Verbal language has one dimension (or one and a half, if we consider the simultaneity of the phonemic and the prosodic system), while the filmic image has three dimensions (two spatial dimensions and one temporal dimension). As a result, the relationship of the two media with regard to reality is different. In verbal language, multidimensional reality has to be represented in a one (and a half)-dimensional medium. In the filmic image, the four prominent dimensions of experienced reality, viz. space (three dimensions) and time, are projected onto three dimensions (two dimensions of space and one of time). The reduction is therefore less severe. Verbal language has to find other ways to represent the richness of reality. If the many dimensions of reality would in some hypothetical way be completely projected onto the one dimension of language, the result would be intolerably complex. Therefore, economy is necessary to limit the complexity. As a result, however, iconicity is also reduced. In the filmic image, the limitations are less severe. The four dimensions of space and time do not constitute the whole of experienced reality, but they do constitute a prominent part of it. There is of course a necessary projection of the other dimensions onto the audiovisual, which entails a limitation. For example, touch and taste have to be evoked through visual features like lighting and colour respectively. This evocation involves synaesthesia: visual features evoke features of other senses. As a result, the representation of these non-audiovisual dimensions is still iconically motivated. In verbal language, on the other hand, arbitrariness of the inventory of signs is introduced because it is much more difficult to iconically represent all dimensions (on the basis of synaesthesia). Verbal language can only use sound in the construction of the lexicon (at least as far as single-morpheme words are concerned), which is not as useful synaesthetically as the visual. The synaesthetic advantage of the visual over the auditive, and the fact that the projection of multidimensional experienced reality onto four dimensions involves a much smaller reduction than its projection onto one-and-a-half dimension, result in the development of economy and arbitrariness being much less urgent in the filmic image. Indeed, we can see that economy plays a much more limited role in the filmic image than in verbal language. There is no paradigmatic economy as regards the inventory of represented participants. However, paradigmatic economy is also concerned with the limitation of grammatical gradational scales. Although most grammatical relations are realised by a gradational scale in the filmic image, there are often on the scale some favoured positions. For example, for horizontal angle, one of the favoured positions is the angle which is as close as possible to the eyeline (this is different in film and television because of multi-camera use). For frame size, visual features can limit certain positions. For example, big close-ups are limited by the lines of eyebrows and mouth. On the syntagmatic level, much more research needs to be done. However, we can give some indications. Firstly, we should make a distinction between temporal and spatial syntagmatic economy. The former can be related to average shot length and camera movement. "Average Shot Length" (A.S.L.) has been proposed by Barry Salt as a criterion for cutting rate. A.S.L. equals "the length of a film divided by the number of shots in it" (Salt 1983: 176). By taking into account the length of the film considered, A.S.L. establishes "strict comparability between films of different length" (Salt 1983: 176). The shorter the A.S.L., the more shots there are and thus the higher is the cutting rate. Therefore I would say that a short A.S.L. is commensurate with a high number of viewpoints shown. This means that short A.S.L. realises an increase in complexity and thus a reduction of economy. However, complexity is not only realised by editing, but also by camera movement. Shots with camera movement can also show different viewpoints. Still, these viewpoints are more integrated than if they were rendered through editing. In the case of editing, the complexity can be foregrounded or backgrounded through "shock" jump cuts or continuity editing respectively. These are just some of the elements of what can be considered to constitute economy or the lack thereof. Roughly, mainstream (American) film has shown an increase of temporal syntagmatic complexity throughout its history (not taking into account the occasional decreases in complexity). Rather than being a grammatical limiting device, increase or decrease of temporal complexity has become linked to text-types, for example art films of the 1960s and 1970s show an increased A.S.L. (see Salt 1983). As for spatial syntagmatic economy, we can discern several possibilities, for example the reduction of represented participants through abstracting devices, and the reduction of information through the limitation of depth of field. These are related to coding orientation and text-types. For example, the limitation of depth of field is linked with abstract coding orientation, as modality depends on generality in this coding orientation (see sections 1.2 and 2.3). Similarly, limitation of depth of field can be linked with text-types. Film noir, for example, often uses low-key lighting in shadowy rooms which causes a limitation of depth of field. These are some indications as to what could be researched in relation to economy in the filmic image. We can, however, already draw the conclusion that economy as a grammatical limiting device is much less present than in verbal language. As mentioned before, this is probably due to the multidimensionality of the filmic image, which imposes a much smaller reduction of experienced reality than verbal language. Therefore, the cinematographic system can be posited as another kind of semiotic system which employs a certain kind of system, viz. the gradational system, to its full potential. This kind of system is also present in the linguistic system, but there it is often limited by economy. The lack of economy and arbitrariness have as a consequence that the filmic image is less suitable for the development of an abstract discourse or argumentation than verbal language. The reason is that, as mentioned before, the arbitrary relation between linguistic signifier and signified ensures the necessary distance between referent and signified, so that the linguistic instance is not bound to the level of concreteness of the referent. In other words, concrete linguistic instances are more abstract than concrete cinematographic instances. The low suitability of the filmic image for abstract discourse constitutes a limitation on its ability to manipulate reality. This limitation, however, should not be absolutised. The filmic image may not be particularly suited for the manipulation of reality in terms of abstract discourse, but it is certainly excellently suited for the manipulation of reality in a more concrete way (see next section). Also, the limited suitability of the filmic image for abstract discourse does not imply that it does not constitute a full-fledged semiotic system. The lack of economy and arbitrariness, and the use of full gradational scales, do not imply that meaning is solely dependent on ever-gliding scales and can only be constituted idiosyncratically in individual texts. The production of cinematographic meaning is still based on an underlying system. The necessary stability of a system is ensured by a number of things. Firstly, the scales themselves may be unlimited (although, as mentioned above, this can be relativised in some cases), but the number of scales is limited. For example, on the interpersonal level we can identify the scales of social distance, involvement, power, and image acts. Maybe more scales could be identified on the interpersonal level, but the number of scales will not be infinite. Secondly, the meanings they realise as a scale do not change with every individual text. For example, the dimension or size of represented participants always realises salience. The individual text does determine which position is realised on the scale of dimension, for example "relatively small", and thus which position is realised on the scale of salience. But this does not change the fact that dimension realises salience. In this way, the necessary stability of a semiotic system is achieved.
4.4. Iconicity and manipulation of realityArbitrariness brings about a distancing of reality, so that manipulation of reality becomes possible. This is seen as a primary requirement of a semiotic system. Therefore, iconic media can be considered problematic: how can iconicity allow for manipulation of reality? We should distinguish between different degrees of distance to reality. Arbitrariness establishes the greatest distance from reality. Next comes "pure" iconicity, as in paintings and drawings, and finally a combination of iconicity and indexicality, as in film and photography. The latter distinction was made by Lev Manovich, who defined cinema as "the art of the index" (Manovich 2001: 295). Between physical reality and the photographic medium there is an automatic lens-based recording link which leaves little room for manipulation, less than, for example, in painting and drawing. Purely iconic media are more tied to perceptual reality than (partially) arbitrary media. This is exemplified in a higher degree of concreteness of their instances. Still, certain strategies are used to manipulate reality. Since the three spatial dimensions are reduced to two, and there is no automatic lens-based registration, the impression of realism, if that is the goal, will have to be created, rather than recorded. As a result, it is easy to diverge from realism, for example by emphasising the two-dimensionality. Another possibility is abstraction of represented participants by reduction of detail, by emphasis on their contour, etc., so that the concreteness imposed by the medium is reduced. Other possibilities are changes of colour, form and texture which not necessarily entail a reduction of realism. The result of these image manipulations is that it is possible to create a purely fictional referent and scene, i.e., "to tell a lie". In media that are iconic and indexical the link with perceptual reality is much closer. It is difficult to interfere in the automatic registration of reality. Therefore, instead of manipulating the image, it is necessary to manipulate reality by manipulating pro-filmic reality, through placement of camera, decisions of what to film and what not to film, manipulating the placement of objects and persons, acting and staging. As mentioned before, film and photography also have several strategies for abstraction of represented participants, like lighting, grouping, and distance. All these strategies can also be used in the purely iconic media. Film has the additional and very important strategy of editing, which allows a fragmentation and recombination of represented reality. Film traditionally is an iconic and indexical medium. There has always been the possibility to manipulate the image through optical effects and optical printing, but this was only a limited possibility. As Manovich extensively discusses, digitalisation has made it much easier to manipulate the image, in such a way that the manipulation can be rendered invisible. As a result, the difference between indexical and non-indexical media is reduced. The discussion of iconicity, arbitrariness and economy allows us to conclude that minimal units can be established by relative contrast on the basis of iconic gradational scales, without this lowering their semiotic usefulness. The limited part played by economy does not endanger the semiotic usefulness of minimal units, as it would in verbal language, because it is due to the multidimensionality of the filmic image which greatly diminishes the need for economy. The limited role of economy and arbitrariness does make the filmic image less suitable for abstract discourse than verbal language. But it is still excellently suited for a more concrete manipulation of reality. This, and the stability provided, firstly, by the limitation of the number of visual parameters and, secondly, by the stability of their meaning, allow us to postulate that the production of cinematographic meaning is based on an underlying cinematographic system. In addition, the stability of this system provides the necessary stable semiotic basis for minimal units.
5. Note On the Integration of GrammarAs mentioned before, I will briefly go into the integration of cinematographic grammar. I will not elaborate on it, as this would lead us away from the scope of this article. A social semiotic approach, supplemented with Groupe µ's analysis of minimal units, gives us an integrated cinematographic grammar. Groupe µ establishes a hierarchical system of subentities, entities and superentities, each also analysable in marques. They oppose an integration of the iconic sign with the plastic sign, based on their possibly different subdivision. However, in linguistic grammar, which is supposed to be an integrated grammar, different systems can have differing subdivisions, like the phonological system and the system of grammar. For example, syllables and morphemes often do not correspond. This is of course due to the arbitrariness of verbal language. But within linguistic grammar there can also be differing subdivisions according to metafunctional level, as explained by Halliday. It is slightly different with the cinematographic system, in that the differing subdivisions of "marques" and "formemes", "chromemes" and "texturemes" are situated at a lower rank than the linguistic examples, but the latter show that it is possible in an integrated system to have differing subdivisions. Halliday further develops the relativity of the uniformity and homogeneity of the linguistic system. He claims that "the kinds of structure found in language are rather varied, and different kinds of meaning tend to be realised in systematically differing ways" (Halliday 1994: 35). As a result, what are supposed to be the general principles of the language system - exhaustiveness, hierarchy, discreteness - still apply, but not necessarily on all metafunctional levels, or on all metafunctional levels in the same way. The principle of exhaustiveness means that the elements of the wording have a function at every rank; but not every element has a function at every metafunctional level. The principle of hierarchy means that elements of any rank are constructed out of the elements of the rank next below. However, there may be a different layering on the different metafunctional levels, for example, in the clause as representation there is more layering than in the clause as message. Discreteness means that "each structural unit has clearly defined boundaries". But as mentioned before, this is only completely true of the clause as representation. The variations established between the different metafunctional levels do not prevent the integration of grammar. The function of the textual level is to integrate the interpersonal and ideational elements. Rhythm in particular functions as an integrating force, both in speech and in the filmic image. In the latter, we can even discern two forms of rhythm, spatial rhythm and temporal rhythm. Van Leeuwen elaborates on temporal rhythm in film in "Rhythmic Structure of the Film Text" (Van Leuwen 1985). The supposition that variation not necessarily endangers the integration of the grammar is very helpful for the conception of a cinematographic system. It recognises that there can be differing systems at work, without this necessitating a total separation of the systems. The latter was typical of the modernist viewpoint of traditional semiotics, with its demand for purity and uniqueness. Eco's conception of the filmic systems agrees with this point of view. He makes a clear division and separation of, among others, the perceptual, the photographic and the cinematographic system. Each level is distinct from the other, and their elements are meaningless on the other levels. These levels are hierarchically and sequentially integrated into each other: each level constructs the elementary units of the next level which are meaningful on the former but meaningless on the latter. The lack of semantic relations between the different levels prevents true integration. The emphasis is on fragmentation rather than integration. In a social semiotic approach, on the other hand, the diversity does not entail a separation, rather the different levels are simultaneously (as opposed to sequentially) and semantically integrated, aided in this by rhythmic structure. In the filmic image, the visual parameters are relevant for all metafunctional levels, but they can be organised differently on the different levels. As these levels are simultaneously and semantically integrated (on the basis of, especially, spatial and temporal rhythm), the diversity does not endanger the integration. This discussion only serves as an introduction to the question of integration, which is not the main concern of this article.
6. ConclusionIn this paper, I have discussed the possibility of minimal units in the filmic image. I started by presenting three answers to the question of minimal units. Firstly, Metz claims there are many different cinematographic codes, each with their own minimal unit. As these codes are not integrated enough to speak of one "langue", one cannot speak of one (kind of) minimal unit relevant for all codes. In section 5, I have claimed that the combination of Groupe µ's analytic framework and the metafunctional framework of the social semiotic approach leads to the stipulation of an integrated grammar in which the visual parameters of Groupe µ are relevant for all metafunctional levels. The integration as described by Halliday does not exclude diversity. Secondly, Eco proposes the "photogramme" as minimal unit for the cinematographic code. However, although this is a physically discrete unit, it has no perceptive and thus no semiotic relevance. Thirdly, Groupe µ proposes an analysis of the still image on the basis of the distinction between iconic and plastic signs. The first can be analysed into marques while the second can be analysed into the visual parameters of form, colour and texture (formemes, chromemes and texturemes). Although Groupe µ proposes a separation of the iconic and the plastic sign, I have claimed that marques are based on the visual parameters of the plastic sign. The difference between iconic and plastic meanings and subdivisions does not prevent their integration. The social semiotic approach allows for differentiation within one system, as opposed to traditional semiotics where differentiation easily leads to separation. After the evaluation of these three positions, I discussed certain issues related to the possibility of minimal units: discreteness and abstraction as well as iconicity, economy and arbitrariness. In the discussion of discreteness and abstraction, it was shown that the difference between verbal language and the filmic image with regard to discreteness is subtler than sometimes maintained. On the physical level, verbal language and the filmic image are both continuous. On the systemic level, there is a difference between the two. In the cinematographic system, discreteness is established by relative contrast, while in the linguistic system, it can be established by relative or absolute contrast. The latter shows that the linguistic system is not as strictly opposed to the cinematographic system as sometimes claimed. In addition, relative and absolute contrast are not related in a binary, black-and-white opposition to each other: relative contrast may establish some degree of absolute positioning, and absolute contrast is slightly relativised by prototypicality and context-dependency. As regards the minimal units of the cinematographic system, relative contrast allows for the considerable flexibility necessary in the filmic image, while a certain degree of absolute positioning (in some cases) and the limited number of visual parameters give the stability necessary for a semiotic system. In the discussion of iconicity, economy and arbitrariness, Eco's rejection of iconicity was linked to traditional semiotics' binary viewpoint, where it is difficult to accept relative contrast or gradational scales as belonging to semiotic systems. In a social semiotic perspective, iconicity can be retained as a flexible and socially determined concept, rather than a rigid scientific concept. Then, iconicity and arbitrariness were linked to different meaning systems. Iconicity was linked to systems of relative contrast on the basis of gradational scales while arbitrariness was linked to systems of absolute contrast. The former are present in both the linguistic and the cinematographic system. The difference is that in the linguistic system, the gradational scales are often limited by economy. Economy, which results in arbitrariness, plays a much more prominent part in the linguistic system than in the cinematographic. The one(and a half)-dimensionality of verbal language necessitates this prominent position of economy, in contrast with the filmic image. In the latter, the three-dimensionality entails a much smaller reduction in the representation of reality, so that economy plays a much more limited role. It is true that iconicity entails an increased difficulty to manipulate reality when compared to arbitrariness. The filmic image is less suitable for abstract discourse. However, some strategies were pointed out which allow iconic media to manipulate reality in a more concrete way, such as manipulation of pro-filmic reality, abstraction of represented participants, changes in colour, form or texture, etc. The discussion of iconicity and economy has allowed us to conclude that relative contrast (on the basis of iconic gradational scales) and the limited part played by economy do not reduce the semiotic usefulness and systemic relevance of the minimal units in the cinematographic system. In the social semiotic approach, differentiation and integration can go together, as opposed to traditional semiotics, where differentiation often leads to separation. As a result, semiotic systems are seen as flexible entities in the social semiotic perspective, rather than rigid codes. This approach allows us to compare the linguistic and the cinematographic system without this necessarily leading to linguistic imperialism. The reason is that differences of the cinematographic system, as compared with the linguistic system, not necessarily mean that the filmic image would not have a semiotic system; rather, it has a different kind of semiotic system. As a result, differences are acceptable for proponents of a cinematographic semiotic system, and they will be more inclined not to ignore them. At the same time, the differences of the cinematographic system are not absolutised, so that similarities are not neglected either. The renunciation of a black-and-white, binary viewpoint allows for a balanced and differentiated approach of semiotic phenomena like the filmic image.
BibliographyDe Grauwe, Sophie. 1999. A systemic-functional analysis of the multimodal text of film. David Lynch's Lost Highway. (Unpublished M.A. thesis, University of Leuven). De Grauwe, Sophie. 2000. "The cognitivist approach to film in the light of systemic-functional theory: a changing of the guards?". In: Image & Narrative, No. 1. http://www.imageandnarrative.be Eco, Umberto. 1972. La structure absente. Introduction à la recherche sémiotique. Paris: Mercure de France. Eco, Umberto. 1976. A Theory of Semiotics. Bloomington/Indiana: Indiana University Press. Elkins, James. 1998. On Pictures and the Words That Fail Them. Cambridge: Cambridge University Press. Genette, Gérard. 1976. Mimologiques. Voyage en Cratylie. Paris: Editions du Seuil. Groupe µ. 1992. Traité du signe visuel. Paris: Editions du Seuil. Haiman, John. 1985. Natural Syntax. Cambridge: Cambridge University Press. Halliday, M.A.K. 1994. An Introduction to Functional Grammar. London: Edward Arnold. Iedema, R. 1998. "Culture-specific constructions of conflict: Bonnie & Clyde (US) and The Runner (Iran)". (Unpublished paper, University of New South Wales). Kress, G. and Theo Van Leeuwen. 1996. Reading Images. The Grammar of Visual Design. London: Routledge. Manovich, Lev. 2001. The Language of New Media. Cambridge/Mass.: The MIT Press. Metz, Christian. 1971. Langage et cinéma. Paris: Larousse. Metz, Christian. 1971b. Essais sur la signification au cinéma (Tome I). Paris: Editions Klincksieck. Metz, Christian. 1974. Film Language. New York: Oxford University Press. Putseys, Y. 1995. Sound Patterns in English and Dutch. Notes on Present-day English. Leuven: Acco. Robins, R.H. 1971. General Linguistics. An Introductory Survey. London: Longman. Salt, Barry. 1983. Film Style and Technology: History and Analysis. London: Starword. Thibault, Paul J. 1997. Re-reading Saussure. The Dynamics of Signs in Social Life. London: Routledge. Van Leeuwen, Theo. 1985. "Rhythmic Structure of the Film Text". In: T.A. Van Dijk (ed.). Discourse and Communication. New Approaches to the Analyses of Mass Media Discourse and Communication. Berlin: Walter de Gruyter. Van Leeuwen, Theo and Carey Jewitt (ed.). 2001. Handbook of Visual Analysis. London: Sage. [1] It should be added that in verbal language, the principle of economy often reduces the number of positions, even in gradational systems. This makes a relative contrast established by these systems less relative than if they were influenced by this principle. Still, the gradational nature of the systems prevents that they become systems of absolute contrast. For an elaboration on the principle of economy, see section 4. [2] The discretion of units in verbal language can also be made problematic by the conditions of an utterance, like low speech volume and bad articulation. But it may be that the obscuring of discrete units as a stylistic device is used more frequently with images. [3] I call the analysis of a concrete given into several types a sort of abstraction because abstraction involves analysis: to be able to generalise from a concrete given, i.e. to abstract from it, an analysis into several general types is necessary. If you generalise to one type only, the resulting type, e.g. "eating man", will not be as general as when you analyse into several types, e.g. "eat" and "man". [4] The focus on words is also very much present in Eco's writings. In most cases he refers to words when he discusses verbal language. [5] Transparency refers to the iconic motivation of word compounds, or, by extension, of word combinations. The individual words are arbitrary, so their meaning is opaque. But the meaning of their combination is related in some way to the sum of the meanings of the individual words, as Haiman states, i.e., the structure of the form (as a combination of word forms) iconically motivates the structure of the meaning (as the meaning will be related to a combination of the meanings of the words). This means that some aspect of the meaning of the combination is transparent (viz. the combinatorial aspect). In addition, as Thibault has remarked, the transparency of word compounds is also aided by the intersection of various paradigmatic series in this form. Each word in the compound carries with it a paradigmatic series, which intersects with the paradigmatic series of the other words. The meaning of word compounds is therefore not only motivated by the syntagmatic combination of their constituent parts, but also by the composite meaning arising from the intersection of the paradigmatic series in the form. Bibliography De Grauwe, Sophie (1999) A systemic-functional analysis of the multimodal text of film. David Lynch's Lost Highway. Unpublished M.A. thesis, University of Leuven De Grauwe, Sophie (2000) 'The cognitivist approach to film in the light of systemic-functional theory: a changing of the guards?' in Image & Narrative, 1, 2000, http://www.imageandnarrative.be Eco, Umberto (1972) La structure absente. Introduction à la recherche sémiotique. Paris: Mercure de France Eco, Umberto (1976) A Theory of Semiotics. Bloomington, Indiana: Indiana University Press Elkins, James (1998) On Pictures and the Words That Fail Them. Cambridge: Cambridge University Press Genette, Gérard (1976) Mimologiques. Voyage en Cratylie. Paris: Editions du Seuil Groupe µ (1992) Traité du signe visuel. Paris: Editions du Seuil Haiman, John (1985) Natural Syntax. Cambridge: Cambridge University Press Halliday, M.A.K. (1994) An Introduction to Functional Grammar. London: Edward Arnold Iedema, R. (1998) 'Culture-specific constructions of conflict: Bonnie & Clyde (US) and The Runner (Iran)' Unpublished paper, University of New South Wales Kress, G. & Van Leeuwen, T. (1996) Reading Images. The Grammar of Visual Design. London/New York: Routledge Manovich, Lev (2001) The Language of New Media. Cambridge (Massachusetts)/ London (England): The MIT Press Metz, Christian (1971) Langage et cinéma. Paris: Librairie Larousse Metz, Christian (1971b) Essais sur la signification au cinéma. Tome I. Paris: Editions Klincksieck Metz, Christian (1974) Film Language. New York: Oxford University Press Putseys, Y. (1995) Sound Patterns in English and Dutch. Notes on Present-day English. Leuven: Acco Robins, R.H. (1971) General Linguistics. An Introductory Survey. London: Longman Group Ltd Salt, Barry (1983) Film Style and Technology: History and Analysis. London: Starword Thibault, Paul J. (1997) Re-reading Saussure. The Dynamics of Signs in Social Life. London/New York: Routledge Van Leeuwen, T. (1985) 'Rhythmic Structure of the Film Text' in Van Dijk, T.A. (ed.) Discourse and Communication. New Approaches to the Analyses of Mass Media Discourse and Communication. Berlin: Walter de Gruyter Van Leeuwen, Theo & Jewitt, Carey, Ed. (2001) Handbook of Visual Analysis. London: Sage |
|||
|
|
||||
|
||||
|
||||
|
|
||||||
|
This site is optimized for Netscape 6 and higher site design: Sara Roegiers @ Maerlantcentrum |
||||||
