Rhetoric from the standpoint of the Lifeworld

The next time the idea of a visual rhetoric is encountered within semiotics, the import of the term is entirely different, but much more specific : within the framework of the « general rhetoric » first suggested by Groupe µ (1970), rhetoric is conceived as a generalisation from the repertory of rhetorical figures developed from Antiquity onward, and later on (Groupe µ 1976, etc.) this idea is applied to pictures. Indeed, the most general procedure underlying all the rhetorical figures could best be described as the production of meaning resulting from a divergence in relation to that which is expected. In more familiar term, it is the deviation from the norm, which is, in the case of verbal language, that of grammar. Indeed, a classical view, attested from Aristotle and Quintilian to French 17^th to 19^th century taxonomists such as Dumarsais and Fontanier, conceives the nature of the rhetorical figures as being divergence : one sign where another is expected to appear. This means, however, that the rebirth of rhetoric within contemporary semiotics has largely been the return of one of its parts only, elocutio, itself reduced to the analysis of figures. Indeed, in the work of influential semioticians such as Roman Jakobson and Barthes, the figures themselves have been shortened to a binary couple: metaphor and metonymy.

The three sciences of communication : rhetoric, semiotics, hermeneutics

The « new rhetoric » represented by Groupe µ (as Ricœur 1975 : 177ff calls it), is thus not so new, in the sense of continuing comfortably within a tradition dominating Western thought since the times of Peter Ramus in the 16^th century for which rhetoric is merely the way of giving artful expression to an already complete thought. It is new, however, in that it furnished us with the tools for analysing how such expression is bought about, going beyond the level of figures, and permitting a meaningful application to domains other than verbal language.

As a contrast, the other school to which the term « new rhetoric » is often attached, inaugurated by Chaïm Perelman and Lucie Olbrechts-Tyteca, which originated outside semiotics, has nothing new to say about the figures, but it does return the focus to the original heritage of rhetoric as a theory of argumentation and persuasion. Indeed, if, as Perelman (1977) claims, the purpose of rhetoric is the produce adhesion to the arguments proposed, then the scope of rhetoric will go far beyond that which is usually suggested by the term persuasion. Clearly, even pictures, and not only openly propagandistic pictures, aim at producing adhesion to the values of their producers. In this sense, as is often proposed nowadays (cf. Foss, Foss, & Trapp 2002), rhetoric is the general theory of communication. If so, however, rhetoric is not a part of semiotics, as Groupe µ would have it, but identical to it in its entirety. However, since semiotics and rhetoric (as well as hermeneutics, which constitutes a third perspective on communication) represent different historical traditions, they may still enrich each other. This is what we will suggest in the following.

This practice has produced at least two symmetrical, equally negative, consequences : by reducing all kinds of semiosis to the mass media kind, in particular to that employed by radio and telegraphy, we become unable to understand the peculiarity of more direct forms of communication ; and by treating all semiosis as being on a par, we deprive ourselves of the means to understand the intricacies added to direct communication by means of different varieties of technological mediation. Taken together, this means that we dispose of no way of explaining the effects of the multiple mediations having accrued to the immediately given world of our experience in the last century. Beyond this, we may even discover a third, even more serious consequence : by projecting the communication model onto each and every form for conveying meaning, we lose sight of that which is really common to all kinds of semiosis.

The most well-known criticism, of course, is that the model relies on a spatial metaphor, i.e., it emphasises the analogy with communication in the sense of trains, cars, etc., construing all meaning as some kind of object travelling from one point in space to another. The identification of communication with transport is probably suggested by the spatial layout of the diagram itself, rather than by the media modelled (cf. Fig. 1.).

Fig.1 The communication model of information theory

Or it may have some even deeper source, as suggested by Reddy’s (1979) well-known analysis of the « conduit metaphor ». Interestingly, the transport model of communication was criticised already in 1929 by Voločinov (1986), well before it was embodied in the mathematical theory of communication. Yet it continues to be the basis of almost everything that is written about communication.

Curiously, the temporal metaphor also embodied in the communication model has not come in for scrutiny : what is accomplished by the sender as well as by the receiver are acts in time, which are close but do not coincide. This is true of the telegraph, but not of everyday face-to-face interaction, nor of the messenger travelling during many years. It applies even more awkwardly to the case of media having to be recreated before being received, such as a piece of music or, in a different way, a movie. The temporal presupposition seems even more beside the point in the case of prehistoric frescoes having been painted once and for all at some forlorn location. Once again, the communication model obliterates precisely those changes that characterise the age of information : even pictures have now become temporal acts, as testified by the television image and the picture imported from some Web server.

Instead of a continuous process initiated by a subject and affecting another, communication really should be seen as a double set of acts, which may coincide spatially and/or temporally, but often do not, and which are initiated by at least two different subjects, the sender and the receiver, or, to choose more appropriate terms, the creator and the concretiser. Curiously, the case of the radio, and to some extent even the telegraph, should really have suggested this model : no matter how much a program is broadcast, no communication will take place until somebody puts his radio receiver on. Nowadays, when we have to start up our computer, connect to the Internet service provider, start the e-mail program and then fetch our mail on the server, we get an even more acute idea of the double initiative required for communication to take place.

According to the conception of the Prague school of semiotics (Cf. Fig. 2.), as it was notably developed by Mukařovský and Vodička, norms, which in part are purely aesthetic, and in part have an extra-aesthetic origin, determine the production of the artefact by its creator, both directly, as a canon, or set of rules, and in the form of a repertory of exemplary works of art which are offered for imitation.

Fig. 2. The Prague school model

In order to become an aesthetical object, the artefact must be perceived by the art public, and this process of perception, termed concretisation, itself depends on the existence of norms, which are ideally more or less identical to those employed by the creator. More commonly, and more interestingly, the norms may have been modified and even exchanged for others since the artefact was created, in which case a new interpretation of the artefact will result. Concretisation involves the determination of the dominants appearing in the structure of the work of art, that is, the elements that are to receive emphasis and which will then organise the remaining elements of the structure according to their purpose ; it also allows the perceiver to fill in lacking details from his own experience.

According to an idea, suggested by Jurij Lotman, the sender and receiver of any situation of communication start out with « codes » — or, as I would prefer to say, systems of interpretation —, which overlap only in part, struggling to homogenise the system of interpretation as the communication proceeds (cf. Fig. 3. and Sonesson 1995 ; 1997).

Fig. 3. General model of communication

We can extend this idea by referring to the conception elsewhere propounded by Lotman and his collaborators in the so-called Tartu school, according to which cultures may be sender-oriented and receiver-oriented, and by transferring these properties to situations of communication. The communicative act may then be said to be sender-oriented, to the extent that it is considered to be the task of the receiver to recover that part of the system of interpretation, which is not shared between the participants. It will be receiver-oriented, to the extent that the task of recovering knowledge not held in common is assigned to the sender. Art, as conceived in the 20^th century, has been characteristically sender-oriented ; mass media, in the received sense of the term (which is not really applicable to all modern media), have been noticeably receiver-oriented.

With reference to this model, it is easy to see that rhetoric, semiotics and hermeneutics situate themselves at different points of the process of communication. They all take in the whole process, but from different perspectives. Rhetoric takes the point of view of the creator of the message: it asks how one it to express oneself in order to obtain adherence on the part of the receiver. Hermeneutics takes the point of view of the receiver: its question concerns how to understand the message of the other (and/or his work). Semiotics takes up a position in between, that is, within the phase going from the artefact to its concretisation : it asks what resources are available for making the process occur.

It is misleading, of course, to say that rhetoric is only involved with the creator : on the contrary, it concerns the relationship of the creator to the receiver, mediated, or not, by the resources at their disposal. The real question, then, is how the creator has to act in order to produce adhesion in the other or, alternatively, how he is to use the resources at his disposal in order to obtain that same adhesion. The hermeneutical question really involves what the receiver is to do in order to be able to understand the other (or the work of the other) or, alternatively, how he is to make use of the resources available to be able to understand the other (or his work). The semiotical question, finally, relates to the nature and kind of resources being available at the moment of communication or, alternatively, to the way in which the thoughts of the creator have been mediated by available resources to the receiver. The alternative versions of the rhetorical and hermeneutical questions are semiotically informed variants. The alternative version of the semiotical question is a rhetorically and hermeneutically informed variety. At the centre of semiotics, then, is the question how : in what way is meaning produced, conveyed, and collected.

From history to theory : Principles of Lifeworld rhetoric

As a historically attested enterprise, rhetoric is a complex build, which has been exploited for different purposes ever since Antiquity. First of all, there is the over-all purpose : the ability to persuade by means of discourse. If we generalise rhetoric at this level to semiotical domains other than language, understanding persuasive intent in the most narrow sense of the phrase, we end up with a rather uninteresting result : visual rhetoric simply involves the study of pictures used in publicity and propaganda. Since is has gained a reputation for being able to persuade people of the truth of any falsehood, rhetoric has often become synonymous with the abuse of language, and the manipulation of the audience. This, of course, is not a sense of the term that could yield any interesting results by being generalised.

Inventio, of course, is the art of finding out what to talk about, but it also already involves a certain organisation of the material, taking into account both social and psychological aspects, such as what is taken for granted within society (the topoi) and the way to influence people (ethos and pathos). This is an aspect which so far have been more or less neglected in the attempt to create a visual rhetoric ; in the following, however, I will take at least the former aspect into account, to the extent that I will have recourse to the notion of the Lifeworld, also known as the ecological sphere.

Dispositio more specifically consists in putting discourse in order, and thus it has something to do with what we today would call the structure of the argument. One of the different « new rhetoric » developed in recent times, that of Charles Perelman, conceives of rhetoric as a theory of argumentation, but so far is has not been directly applied to pictures (except, in passing by Meyer 2005 ; et alia 1999). And yet it is of course important to establish whether pictures are capable of giving expression to an argumentative structure, which is something that has often been denied in theory, but everyday seems to be borne out in the present-day practice of the mass media. It is at least conceivable that, fundamentally, an argument can be carried out in an identical manner in pictures as in verbal language.

As for elocutio, which is the stylistic elaboration of the argument, it involves, most notably, rhetorical figures, such as metonymy mentioned by Barthes. There have been numerous attempts simply to apply the existing repertories of rhetorical figures to other domains, such as pictures, but the result has never been satisfactory. If not at the level of argumentation, then pictures are certainly entirely different from language at the level of « style », that is, more properly speaking, as far as the semiotic resources at their disposal are concerned. There is, to begin with, nothing similar to words and sentences in pictures. In fact, the differences are even more wide-ranging, as we shall see presently. The second tradition of « new rhetoric » thus has been the one initiated by Groupe µ, which attempts to go beyond the traditional figures, to discover a set of general operations responsible for the functioning of these figures in verbal language, which can then be seen to work in another way, given the kind of resources offered by other semiotic domains, such as pictures.

It is thus to the initiative of the Groupe µ that we own the first real transposition of the rhetorical model to the visual domain, in a version at first narrowly copied on verbal rhetoric, which then was completely reorganised by taking into account the specificity of visuality. This generalisation takes place on the level of elocutio, and, more precisely, in the domain of figures.That aspect of rhetoric, which is exploited in the theory, developed by Groupe µ is thus the capacity for producing meaning by creating expectations (itself based on pre-existent expectations) only to leave them unfulfilled.

To conceive rhetoric of visuality, is necessary, I contend, to start out from the Lifeworld, the Lebenswelt of Husserlean phenomenology, the first layer of reality that, for the subject of perception, is « taken for granted » (Sonesson 1994a, b, c, 1995a). In the Lifeworld, which is simply the real world, as experienced by human beings, there are certain regularities, or as Husserl says, certain « ways in which things tend to behave ». Taking his cue from Husserl, the psychologist James Gibson has maintained that it is the laws of this world, not that described by the natural sciences, which are upset by magic. This is true, in my view, also of the kind of semiotic magic known as rhetoric. In the Lifeworld, which A. J. Greimas described as « the natural world », and which Gibson calls « ecological physics », there are objects (or, as Gibson says, « substances ») which are relatively independent, while at the same time maintaining more or less strong relations of integration with regard to each other, going from simple coexistence to the relationship between the whole and its parts. In relation to the structures of the Lifeworld, there will therefore be a first rhetoric which consists in an overturn of the relations of integration, which is similar to magic such as Gibson conceives it. This is what I will later on call the dimension of indexicality.

But common sense also expects things which appear together to be sufficiently different to be told apart, without going to the opposite extreme of excluding each other. In this sense, there is rhetoric of too much resemblance as well as of too much difference. It will be termed the dimension of iconicity. A third rhetoric takes its starting point in the sign character of the picture. Our expectations are disappointed both by finding a too large part of reality in the picture, and by finding ever further levels of fiction within the pictorial content. Since it refers to the very sign character of the picture, this can be termed the dimension of symbolicity. Finally, pictures a such are have distinct uses in a given society and are, because of this very fact, ascribed to various categories, thus giving rise to a fourth rhetoric in the case of which our expectations, as regards the social categories pertaining to the pictures, are not fulfilled. Whereas the first three dimensions all concern the structure of perception, including, in the case of the third dimension, the picture itself as a perceptual object, the fourth dimension involves the categories of which picture signs are members: in this sense, it is a dimension of socio-cultural categorisation.

Consequently a problem arises: does this standard involve normativity or normality ? Initially, at a stage close to perception, normality no doubt prevails. We are in the Lifeworld, in the sense of an Umwelt common to all human beings. At a later stage, which is found especially in dimensions three to four, the rupture is defined, not so much with respect to expectations as such, as in relation to that which is wished for or prescribed. We then find ourselves within a particular sociocultural Lebenswelt. Actually, one passes from normality to normativity through several intermediate stages, as was already recognised by the Prague school (cf. Sonesson 1992a, c ; 1994b).

In spite of these differences, the revisions proposed all stem from a contention which I appear to share with the Groupe µ : the fundamental importance of basing rhetoric on what we know about the perceptual and the cognitive functioning of human beings. Following the example of the Groupe µ, I will abandon the traditional figures of verbal rhetoric, but I will retain the principle according to which the production of meaning originates in the transgression of norms. But this principle will, in my view, be better served by going back to the general operations which appear in the early work of the Groupe µ (1970) bearing on verbal rhetoric : beyond the two binary couples opposing presence to absence and disjunction to the conjunction, which the Groupe µ (1992) has employed in their recent publications, I will return to the general operations of addition, suppression, substitution, permutation, augmentation and diminution.

Pictorial disposition : The argument from perception

The problem with this view of rhetoric is of course that it is a generalization of elocutio. The kind of dialectic suggested by Groupe µ is real (and introduces a social perspective elsewhere absent in their work), and it is accounted for by the Prague school model. In order to avoid terminological confusion, however, we here prefer to relate semiotics and rhetoric as suggested above: together with hermeneutics, they are two differently directed perspectives on the same process of communication, centred on the creator, the means of communication, and the receiver, respectively.

However, the rhetorical tradition clearly imposes another task on us: if we are going to talk about a pictorial rhetoric, then there must be some means for pictures, if not to conduct an argument, then at least to make simple statements. And this seems to ruin our attempt from the beginning : many philosophers and semioticians, such as Wittgenstein, Peirce, Sol Worth, and Kjœrup, have explicitly denied the possibility of pictures affirming anything: more precisely, many of them have added, pictures may affirm something only if a verbal label is attached to them. In a sense, this is trivially true: if we define a statement as being a verbal construct, then, of course, pictures cannot state anything. Pictures, as pictures, are simply not made up of verbal constituents. However, if we define a statement simply as an operation, by means of which a particular property is said to pertain to (avoiding the more value-laden term «‑be true of) a certain entity, then, of course, it may be possible for picture to make statements in the way of pictures. But even this may seem impossible, if pictures, as if often thought, simply reproduce the world of our experience.

If we are to believe many semioticians, from Lessing to Goodman, pictures are not only able to describe the whole of space, but they cannot avoid doing so: they can only show « fully determinate entities », that is, really as it is, in a single chunk. This is certainly not true : as I have shown, against Goodman, the « density » of pictures is only relative, and all kinds of abstraction are found in them (cf. Sonesson 1989a : 226ff and 324ff ; 1995a). This applies to the expression plane, in the case of more or less schematic pictures, but is also applies to the content plane of some pictures the expression plane of which is fully dense. Thus, for all practical purposes, many pictures are not signs of individuals, but of abstract roles in more or less generic situations.

The difficulty posed by narrativity in pictures, as Bayer (1975) reads Lessing, is that the picture is unable to abstract : Homer may show the gods drinking and discussing at the same time, but pictures are unable to convey all this information. To my mind, it is not the amount of information, which is crucial (the picture may easily carry more), but the possibility to organise it: verbal language is able to convey relative importance, newness, theme, etc., and the picture may possess corresponding mechanisms, which we do not know of. But the space of representation in the picture is, at the same time, the representation of the space of ordinary human perception, which hampers any kind of organisation by other systems. Some modifications to this principle was introduced by Cubism, Matisse, some forms of collages and synthetic pictures, and also by some visual systems of information, logotypes, Blissymbolics, traffic signs, etc.

In the following, I mean to suggest that some kinds of thematic devices exist also in pictures, permitting pictorial texts to acquire the force of statements, and thus to function rhetorically, as well as narratively. In an earlier discussion (Sonesson 1989a : 333ff), I felt compelled to conclude that, contrary to pictorial signs generally, pictorial metaphors, and thus perhaps all figures of visual rhetoric, are not necessarily asymmetrical. Put in another way, it is often impossible to decide which term is the « vehicle » and which is the « tenor ». Thus it is not clear whether Magritte, in « Le viol », is trying to tell us that female trunks resemble faces, or rather that faces are similar to trunks.

The direction of the comparison is often made clear by the accompanying verbal text, by the genre of the picture, or by the larger context. For instance, once we have identified the pictorial category « advertisement », we understand that the super-market chain B&W wants to tell us that their fruits are as valuable as a crown, not to inform us about crowns, which, in some curious way, resemble heaps of fruit. Indeed, one may wonder if we really need to identify the genre here: given our social value system, according to which a crown is assigned much more import than a bunch of fruits, it makes much more sense stating that fruits are like crowns, than the reverse. Applying Breton’s or Bataille’s doctrines to the same surrealist photographs, we will inevitable come up with different interpretations, but in both cases, that which is depicted is clearly meant to be subordinated to something else — as long as we remain inside the framework of surrealist discourse, allowing the linguistic sub-text to determine the pictorial one (cf. Marner 1996 : 17f ; 1999). In Magritte’s « Le viol », an interpretation along Bataille’s lines will certainly reduce the head to a trunk, rather than the reverse, but the result would probably be the same, only with an opposite evaluation, if we follow Breton’s style.

In these cases, genre conventions, common sociocultural understandings, or explicit theoretical discourses permit us to discover which of the terms of the comparison is the most « prominent » one. But the question is whether there is also some mechanism internal to the pictorial sign, or at least independent of verbal texts and genre categories, which is able to decide the direction of the metaphorical relation. According to Groupe µ (1976 : 47), Julien Key's « chafetière » is normally thought to represent a cat disguised as a coffee pot, rather than the inverse, the reason being that the feline parts and features have been more completely reproduced. We know from the advertisement function of the picture that it is to the coffee pot, and thus to the coffee, that the feline properties are to be transferred. Is then the term most completely rendered in the picture the one, which is the vehicle of the metaphor, or is it instead the tenor, the internal mechanism having been overridden, in this instance, by the publicly acknowledged function of advertisement pictures ?

In any event, the case of the metaphor in absentia seems much clearer : it must be the absent term which was expected, e.g. the ice-pail, that is likened to the term present in its stead, i.e. the Coliseum, in the publicity for the aperitif Punto e Mes placed within the Roman building, complete with ice cubes and all. If, on analogy with this, that element that has to be abductively supplemented, even if only partially, is the tenor, then, in Key's poster, the coffee pot is the tenor, and the cat is the vehicle. However, the Coliseum picture could well turn out to be too simple for us to draw any more definitive conclusions : if the Coliseum had required any complementary object, which were to substitute for the bottle and the ice, as the latter requires an ice-pail, which substitutes for the Coliseum, the issue would have been much more complex, and this is obviously what happens, for instance, when a bird’s head, in Max Ernst’s print, « Les rencontres de deux sourires », appears in the place of human one, leaving out both the human head and the bird’s body.

In a satirical picture, like Pasch's « Hen painting », showing hens with female faces, at the time recognisable as those of court ladies, we expect the « worst possible » reading to be the correct one, and we know the themes broached to be such things as human beings and human affairs, so there can be no doubt that Pasch's painting is about court ladies being similar to hens, rather than the reverse. But this is probably not so only because we recognise the court ladies ; rather, following Lifeworld principles, we reduce that which is less close at hand to that which is more familiar, and in the human world, as opposed to that of hens, human beings are ipso facto more directly relevant. In a similar vein, though of course much further away on the scale of familiarity, Raoul Ubac’s fossil of the Eiffel tower certainly transfers fossility to that which is more familiar, the Eiffel tower, not Eiffel-tower-ness, what-ever that is, to fossils (cf. Marner 1996).

What we here discuss is often addressed in terms of subject and predicates, topic and comment, and so on. As Halliday (1967 : 201ff ; & Hasan 1976 : 325f) has observed, these terms tend to confuse at least two different distinctions, the theme and the rheme, and the given and the new : that which is the subject of discussion, as against that which we have to say about it; and that which is taken to be known beforehand, i.e. « recoverable » from some other source, as against that which is introduced as new information. Every unit of information must contain something new, whereas the given element is optional. These terms are not specifically linguistic : in language they are mapped into intonation groups, and the like, but in pictures they may well appear as something else.

It seems to me that facts, which are recoverable from the structure of the common Lifeworld, or from the norms present in some particular socio-cultural world, must be treated as givens. Consider the case of Max Ernst’s bird-headed man: the theme here must be taken to be « human being » (although the scheme for human being is only partially fulfilled), whereas the rheme becomes something like « bird-likeness ». What is given (known and recoverable from the picture) is the human trunk, in relation to which the new information must be « head of bird ». Similarly, Ubac’s photograph is clearly about the Eiffel tower, to which is attributes fossility (itself induced by some kind of iconic relationship to solarisation) ; however, the Eiffel tower’s shape and general appearance seem to form part of the given, while only some difficult-to-describe plastic property (resulting from solarisation) could be considered to be new.

These considerations do leave us with a vocabulary for talking about information structure in pictures; they do no, however, answer definitely the question whether thematic relationships may sometimes be reversible in pictorial signs. More fundamentally, however, the suggestion here is that what appears as mere elocutio, in reality serves the function of dispositio, that is, of framing an argument, at least in the case of visual rhetoric. As Aristotle observed, and as Ricœur (1975) has insisted, the metaphor (understood as a stand-in for all figures), is a cognitive operation, involving both the discovery and the communication of relationships (notably similarity), which were not known to obtain beforehand. The figure does not simply decorate : it makes us see. In this sense, whether it is reversible or not, it is a thematic device.

This is why the rhetoric based on homogeneous transformations, which we are going to study in the following in only the most obvious case of visual rhetoric. As mentioned above, Groupe µ (1992 :295) only recognise what they call « heterogeneous transformations », the result of which is a picture showing at the same time various styles of reproduction. The typical case might be the Cubist collage. This is so because they compare the picture to itself, attending to the homogeneity of its expression plane. But we may also compare the picture with that of which it is a picture, that is, in Perelmanean terms, with the structure of reality. Since it is an iconic sign, the picture posits at the same time its own resemblance and dissimilitude with respect to its object : because of the first feature, the sign creates an expectation of identity that, by means of its second aspect, it necessarily disappoints. Unlike language, which is not primarily iconic, the picture is from the beginning a statement about a reality, which, by the same token, it transforms. Homogeneous transformations are already rhetorical, in the double sense of elocutio and dispositio: the resulting picture deviates from reality, and, by this means affirms the otherness of reality.

The µ model within the space of the hermeneutic circle : lucunae and repleteness

The rhetorical model of Groupe µ is no doubt, together with the work of Jean-Marie Floch, the single most important intellectual feat realised within visual semiotics. If this is granted, one may wonder why we cannot simply accept the model as it stands today. The present author has for a long time been one of the most faithful « fellow travellers » of the Groupe µ : at first I was very impressed by the early article on « visual metaphor » (Groupe µ 1976, etc.), yet deciding to reformulate their analysis of « the chafetière » (the cat which is also a coffee pot) in indexical terms derived from a phenomenology of perception (Sonesson 1989a ; 1990), but then I greeted the Traité du signe visuel (Groupe µ 1996a) as a contribution of the very first order to the semiotics of pictures (Sonesson 1996a), to the point of forgetting the foundations that I had formerly searched in indexicality. But I soon found reasons for doubts : to the dimension of the conjunctions and the disjunctions which are present or absent, I thought it was necessary to add a « generalised oxymorology », later to become my second dimension (cf. Sonesson 1996b), and even to substitute continuous axes for the discontinuous terms posited by structural analysis (Sonesson 1996c ; 1997a ; 2001a ; 2004b, 2008).

But these footnotes to the project envisaged by the Groupe µ really result from much deeper misgivings. As a start, the problem is that there are three (or at least two) different theories of rhetoric in the Traité, which are never properly connected. First of all, there are the comprehensive operations stemming from the early « general rhetoric » (Groupe µ 1970), such as addition, suppression, substitution, permutation, augmentation and diminution, which are never really applied to visual semiosis. Then there are a variety of transformations, which account for the difference between what we see in perceptual reality and within the picture sign. Finally, there is the structural cross-classification in terms of absence and presence, combined with conjunction and disjunction. Only the last one seems to be operational, but as soon as you try to use it, it breaks down. Instead, the operations of general rhetoric would need to be mapped directly onto the perceptual worlds by means of different kinds of transformations.

In order to discuss the failings of this model, we must start by clarifying what will here be understood by the term model : it is a grid defined by a method. From this point of view, a method can be defined as a set of procedures or operations applicable to a given phenomenon, the application of which transforms the latter into objects to be studied, with the objective to formulate generalisations about the object of study. In semiotics, the object of study is signification, of which the picture is one kind. The objects studied can be concrete significations, for examples pictures, but they may also be other phenomena, such as, for instance, intuitions. The model serves as mediation between the objects studied and the object of study.

In present-day semiotics, there are at least four different methods. First, there is text analysis, the object of which it to describe, in a exhaustive way and from a particular point of view which we give ourselves in the concrete case, a given picture, or a set of pictures, allowing us to extract from the analysis a model which can be applied to other pictures. This is the most common method, which appears, in the semiotics of pictures, in the work of Floch, Thürlemann, Saint-Martin, etc. In the case of system analysis, the researcher, guided by his intuition as a member of the human community, determines which terms go together, as well as the limits of the variation allowed within each term. We find this method in the work of the followers of Peirce but also, partially, in that of Eco. The experimental method has also been used in semiotics, especially in the study of pictures, notably in the work of Tardy, Lindekens, Krampen and Espe. In that case, we build an artificial «‑text‑» which must then be evaluated with regard to the system or else completed by a text contributed by the experimental subject.

In discussing elsewhere the methods of pictorial semiotics (cf. Sonesson 1992c), I have reserved a special part for classificatory analysis, represented especially by the Groupe µ : this method combines the character of a conceptual combinatory found in system analysis with the choice of a concrete example for every combination of properties, derived from an elementary text analysis. By constructing a sufficient number of combinatorial tables, one could in principle produce an exhaustive analysis of a single picture, that is a text analysis, but this really seems completely utopian, given the quantity of tables, which would have to be to constructed.

These four methods may be seen as different interventions within the hermeneutic circle which goes from the system to the text and back again, from the general principles to the single occurrences and vice versa. Groupe µ entirely rejects what I have here called text analysis, e.g. the exhaustive study of a single work : indeed, they maintain that this analysis proves nothing and that it gets lost in particularities. However, I think text analysis possesses a distinct heuristic value, simply because it demands the exhaustiveness of texts ; that is, it requires the analytical procedures to tell us everything there is to say about the studied objects. For this reason, it serves as a test for the results produced by other methods. Classificatory analysis, as well as system analysis, only demand the exhaustiveness of the system, that is, they require all the possibilities contained in the combinatory to be exhausted.

While one must admire the elegance of this analysis, I believe that the symmetry on which it is based is illusory and not very enlightening in the end (Sonesson 1996a, b). As I noted above, the µ model, just as any other, can be questioned in several ways : because the descriptive terms are not adequate to put the analysed objects into opposition ; because certain objects, that is, in this case, the pictures, which, at a pre-theoretical level, seem to be different in interesting ways, are not distinguished by the model ; or because there are other objects, in this case other pictures, which do not enter naturally into any of the categories supplied by the model.

Here I am going to retain two critical observations which will allow us to develop a more comprehensive model : first, the descriptive predicates appear to explain nothing, because they can only be understood on the basis of the examples, simply because the world of perception abounds in every kind of cases intermediate between the conjunctions and the disjunctions ; and, in the second place, the distinction between present elements and absent elements is analytically inert, because, except in some marginal cases, any rhetorical operation supposes at the same moment an absence and a presence.

Extracts from the µ-tological bestiary : from Captain Haddock to the cat-coffee pot and back again

Let us pick the twin cases of Captain Haddock, which is described as a conjoined figure in absentia, and the « chafetière », which is supposed to be a disjoined figure in praesentia. Indeed, one may wonder if it had not been better to say that both figures are in praesentia, the first one, that is the picture of Captain Haddock, being disjoined, because one of its parts is separated from the whole, while the « chafetière » is conjoined, because, in that case, two objects are merged. In fact, this terminology is also deceptive, because we could just as well turn it around: the picture of Captain Haddock is conjoined, because it involves a part attached to a totality, while the « chafetière » is disjoined, because it associates two separate objects. When presenting these examples to my students, I have invariably found that the classification is experienced as being mysterious and even arbitrary, and even when, in the end, the students manage to understand it, this happens because the examples serve to clarify the terms, whereas it is of course the model which is supposed to make sense of the examples rather than the reverse. This difficulty of comprehension is not fortuitous : in fact, on one hand, each absence supposes a presence, and vice versa, and on the other hand, there are a multiplicity of intermediate cases between conjunction and disjunction.

Instead of seeing the bottles as replacements for the pupils in the eyes of Captain Haddock, we could conceive the whole as an interpenetration of bottles and Captain Haddock, exactly as la chafetière is presented as an interpenetration of the cat and the coffee-pot. No doubt, we first identify the scheme, globally, as a person (and, more particularly, as Captain Haddock), and we then discover that the part of body where we expect to see the central part of eyes, the pupils, is occupied by other objects, the bottles ; whereas, in the case of the « chafetière », already the global information about the identity of the phenomenon is contradictory. However, this difference seems to have nothing to do with a part of the figure being present or absent.

In fact, in both cases, some elements, which we had expected to be present, are absent (the pupils of Captain Haddock as well as some parts of the cat and the coffee-pot), while at the same time certain elements, which we had expected to be absent, are present (the bottles and some different parts of the cat and the coffee-pot). This appears to be the most common case in rhetoric: actually, it is rather difficult to find examples where there is only an absence or a presence, which goes contrary to our expectations. On the other hand, in both cases, there are elements which are separated in the Lifeworld which go together in the picture (the bottle and the eyes excepting the pupils, as well as the cat and the coffee-pot) and some elements which are conjoint in the Lifeworld which are presented as separated in the picture (the pupils and the other parts of the eyes, as well as some elements of the cat and the coffee-pot).

When rephrasing this analysis in the terms of a more or less great difference with regard to expected integration, I have substituted for the binary couples a continuous scale of more or less indexicality : instead of the oppositions, there are dimensions which are continuous, at least to a certain degree (cf. Sonesson 1997a, 2001b, to seem b). In fact, as is already suggested by the terms of contiguity and factorality, the perception of indexicality is not purely quantitative, but also supposes qualitative jumps. In order to account for both the systematic character of the perceptual structures, and the intrinsic motivation of indexicality, it is necessary to have recourse to mereology, the science of the whole and its parts, once conceived by Husserl, but nowadays taken up by cognitive science.

In the following, we are going to study two aspects (by no means exhaustive) of the problem set by the µ model, the first of which concerns the arrangement of the Lifeworld which serves as a basis for the operations producing rhetorical modifications, whereas the second one involves the organisation of the rhetorical sign itself. In the first case, I am concerned to find a substitute for the conjunctions and disjunctions of the µ model. In the second case, on the other hand, I will attempt to discover a means of modulating the range of absences and presences.

The construction of the Lifeworld in perception and society

All signification has its origin in perception. This was long ago recognised by philosophical phenomenology and has since then been amply illustrated by psychology and cognitive science. As soon as we try to go beyond the linguistic model in semiotics, as I claimed in an already ancient book (Sonesson 1989), we are obliged to take perception into account.

Already in the 1940ties, the Prague school observed that the sign, before being the anything else, was an object of perception. According to Mukařovský (1974), any work of art is an artefact, which acquires a real life only when it is perceived by somebody, who thereby transforms it into a « concretisation », filling in its « empty and indefinite places » by means of his own experience. However, for Mukařovský, these experiences are of a social order : they have been formed in the society in which the perceiving subject lives. Thus every act of perception is overdetermined by norms, canons, and repertories of exemplary works.

This model rests on a number of more general principles derived from the phenomenology of Husserl, notably those which concern the regularities characterising the « Lifeworld », which is this first layer of reality which, for the subject of the perception, « is taken for granted » and which is directly given in perception. According to one of these principles, any object appears in perception in a given perspective, through some of its parts, and conceived in a particular way, while being always perceived as such. This explains the presence of empty and indefinite places in the work of art, as well as in any other sign (cf. Sonesson 1989a, I.2. and 1992a, 1994a, b ; 1996a).

In fact, the importance of the Lifeworld for a perceptual description of the picture, as well as of other semiotic objects, goes well beyond the elements absorbed by the Prague school. The idea of a Lifeworld, understood as a layer of meanings taken for granted, the reality of which is never questioned, was developed by other phenomenologists, notably by Alfred Schütz (1932 ; 1967), in the case of the sociality, and by Aron Gurwitsch (1957; 1974), in the case of perception. We also find a similar conception in the theory of Peirce, who defines « abduction » as a conclusion based on a regularity, which has not been shown to be true but which is nevertheless generally accepted (cf. Sonesson 1989a, I.2 and III.3.); as well as in the notion of « perceptive mass » defined by the Russian Formalist Yakubinskij and taken up by Vygotsky and Bakhtin (cf. Wertsch 1985:84ff).

The science of the Lifeworld was more rediscovered recently, by Greimas (1970:49ff) who describes it as a « semiotics of the natural world », in the sense in which we speak about a semiotics of natural languages, that is, linguistics. In both cases, the naturality in question stems from the feeling of the subject using it. Similar ideas are expressed in another terminology within the domain of « naïve physics » described in the cognitive sciences (cf. Smith 1995a ; & Casati 1994). The basic principles of the Lifeworld were summarised and amplified by the psychologist James Gibson (1978 ; 1980), who talks about « ecological physics », which is at the foundation of that ecological psychology which he created, in order to describe the conditions of possibility of the kind of perception available to a real subject in the everyday environment.

Just as Husserl, Gibson emphasis the peculiarities of perception in the real world, in contradistinction to such perceptions as may be produced artificially in a laboratory. According to his instructive formula, it is the principles of « ecological physics », not those physics as a natural science, which are overruled by magic — and, I would add, by visual rhetoric. As everything which is « taken for granted », these laws become obvious only when they are broken, which is also the case in rhetoric, in proxemics, and in the semiotics of culture (cf. Sonesson 1994b ; 1996a ; 1999a, 2000b, c). Some of the « laws » of « ecological physics » are identical to the regularities of the Lifeworld, as described by Husserl. As Manar Hammad (1989 : 31f) rightly points out, many regularities end up becoming rules. Indeed, as was recognised by the Russian Formalists even before the Prague school, once established the norm will not remain in place for ever, but even while it is upheld, it mostly serves as the background against which transgressions take place. It was precisely by applying this observations to the history of art that I was able to describe Modernism in visual art as a rhetorical device which is all the time repeating itself (cf. Sonesson 1993b ; 1998d).

If all concretisations are social, as Mukařovský claims, they have been formed by the specific circumstances in which the individual lives, but also, beyond the peculiarities of his society and the experiences of the individual in question, by the generic structures of sociality. As I have demonstrated elsewhere (cf. Sonesson 1989a, I.2.2.), this social overdetermination is not only valid at the level of the sign, but even at the level of significations which precede the constitution of the sign: the die is not perceived as such less immediately than the cube which make up its basic shape. Contrary to Greimas, who would like to see « the natural world » as a semiotic system just as any other, it is no doubt necessary to admit, with Gibson as well as Husserl, that the Lifeworld constitutes a fundamental layer of meaning on which all other possible systems of meaning are based. But this does not prevent this elementary layer from being already socially and/or culturally overdetermined: to make up, in Gurwitsch’s term, a particular sociocultural variety of the Lifeworld.

From this point of view, the sign appears as a particular modification of perceptive intentionality. Going beyond what is simply taken for granted by Peirce as well as by Saussure and Hjelmslev, Husserl (1939 : 174f) defines the sign as a complex unity which consists of an entity which is directly perceived although it does not constitute the theme, corresponding to the expression, and another entity which is thematic, while being given in a indirect way, corresponding to the content (cf. Schütz 1967 : 294ff and 1980 : 99ff). The idea according to which a sign of a particular kind, the picture, is a case of indirect perception, was taken up more recently in the psychology of perception, by Gibson (1980 ; 1982 ; cf. Sonesson 1989a, III.3 ; 1992b, c). It is at this level that the question regarding absences and presences has to be posed : an object is absent when it is perceived in an indirect way (at second degree or more) while remaining thematic, whereas the object present is less thematic but more direct.

However, within the Lifeworld, the basic element is not the sign, nor its features. It is, in Gibson’s terms, the « independent, detached object ». And we have to start out from such an « independent object », in order to find a replacement for the disjunctions and the conjunctions of the µ model.

The « independent object » within the structure of the Lifeworld

To conceive a specifically visual rhetoric, it is thus necessary take our point of departure in the Lifeworld, in the sense of a primary layer of reality which, for the subject of the perception, « is taken for granted ». In this world, there are objects (or, as Gibson says, « substances »), which are more or less independent, but which at the same time maintain more or less strong relations of integration, going from simple coexistence to the relation between the whole and its parts – in other words, from contiguity to factorality. As I noted above, the first dimension of rhetoric therefore must consist in breaking up these connections, which is similar to magic as conceived by Gibson. From the point of view of Lifeworld experience, we expect the picture to show us the « independent objects » encountered by perception, neither dissolved into more large-scale entities, nor divided into smaller objects. If the degree of integration of Lifeworld things is modified, there is a transgression of norms and thus rhetoric.

Already in an early book, mainly concerned to refute an erroneous idea of iconicity (Sonesson 1989a), I analysed certain phenomena, rather similar, and sometimes identical, to those which Groupe µ calls figures of visual rhetoric. In order to accomplish these analyses, I took my point of departure in the notion of indexicality, understood as something vaster that the sign carrying this name, namely the connections of contiguity and factorality (the relations of the parts to the whole) which characterise perceived significations. Later analyses inspired me to separate the notions iconicity and indexicality from the concept of sign (cf. Sonesson 1994a, 1995b, 1997b, 1998a, b, 1999b, 2000a, 2001a, c, d). Indeed, from a Peircean point of view, indexicality is simply the property, which transforms something, which is already a sign into an index. However, by a slight adjustment of emphasis, which has at least some justification in the work of Peirce, indexicality could be conceived as a property which, when it is added to the sign function, creates an index, but which, besides, can have other parts to play in the constitution of meaning (Cf. Sonesson 1995b ; 1998a).

If we consider indexicality or iconicity independently of the sign relation, they can be considered as the « grounds » which, according to Peirce, make up the particular point of view from which is conceived the relation between the various parts of the sign. In this sense, the « ground » is equivalent to relevance, in the sense of structuralist semiotics, that which makes the difference between « form » and « substance » – or, to express it in Peircean terms, the abstraction, which, in the case of iconicity, amounts to something like « the blackness of two black things » and, by analogy, in the case of indexicality, the « existential » or « spatial relation » between two things that exist together in space (cf. Peirce 1.558 ; 1.196 ; 2.305 ; 3.361 ; 8.335).

Numerous examples offered by Peirce seem to be in agreement with the conception of Jakobson (1979), according to which indexicality is based on « real contiguity » which can be identified with the syntagmatic axis of language, as well as with the rhetorical figure known as metonymy. For Jakobson, however, metonymy involves not only the relation of contiguity of traditional rhetoric, but also that of the part to the whole, known in rhetoric as a synecdoche. What gets lost here is no doubt the distinction between the « independent object » and its parts. This distinction can be restored within the category of indexicality (cf. Nöth 1975 : 20f), and may be described more generally in terms of contiguity, on one hand, and factorality, on the other one (cf. Sonesson 1989a :40ff ; 1998a).

Proximity or neighbourhood is a basic factor in perception according to Gestalt psychology, and it is also one of the relations included in the topological perception of space. The relation of the part to the whole is fundamental for Gestalt relations as such. Indexicalities which are not as yet signs consist in relations between objects which are not situated at different levels of accessibility or thematisation, or which are not clearly differentiated from each other. In such cases, we can speak about contexts or couplings (in the sense of Husserl). Any experience involving two elements connected by proximity, conceived as a basic perceptual fact, can be considered as a real perceptual context implying contiguity. A real perceptual context implying factorality is any experience of something as being a part of a totality, or as being a totality having parts (cf. Sonesson 1989a, I.2.5).

According to Peirce, the rolling gait of a man is an index for the property of being a sailor: but to be a sailor is a social role, not a singular fact. More exactly, this particular way of walking is a part of a social habitus defining a role, which is a part of a totality (a factorality). But if the relation of a property to the totality of which it is a part is indexical, then it is reasonable to think that indexicality will also explain the relation between an object and a class of which it is member. Such examples are apparently not among those mentioned by Peirce, but they have often been quoted by later semioticians: thus, for example, if a pretzel is an index of a bakery, this is so because it is a member of the class of products sold in the bakery. A class is certainly not a singular object, but it can be considered to be a collection of objects. Often, however, such a class is itself determined by abstract properties. The sample which shown to us by a dressmaker, for example, is the sign of a class of tissues having the same quality of texture and the same pattern, but not the same shape or size. Certain samples, for instance colour samples, can even be indices of the abstract properties themselves (Sonesson 1989a :43ff, 137ff ; 1989b : 60f ; 1998b).

Moreover, I shall retain here the first lesson of Husserl’s study, which consists in putting the emphasis, not on the way the whole results from the addition of its various fragments, or, in a parallel fashion, the way in which the part is derived from the division of the totality; but rather, on the relations of mutual or one-sided dependence (among which are to be found the counterpart and the autonomous object) which exist between the parts and the totality which they establish together. In this sense, the mereological model is not equivalent, in the linguistic domain, to a constituent structure grammar, of the kind envisaged by Chomsky, but to a grammar of dependence, such as is conceived by Tesnière, or a categorical grammar, in the sense of Montague (which moreover is inspired, through several intermediaries, by Husserl’s study ; Cf. Sonesson 1989a III.5.1). Without reference to Husserl, however, Hjelmslev bases his glossematics on the same minimal system of dependencies between the whole and its parts.

The task of the mereology will not only be to account for the relations between the whole and its parts, but also to explain the difference between various kinds of totalities. Husserl opposes configurations to aggregates, and we find attempts of the same kind, but sometimes more developed, in the work of various representative of Ganzheitspsychologie (cf. Sonesson 1989a, I.3.4). Peirce even wrote a very long but rather disorganised list of various kinds of totalities (quoted in Stjernfelt 2000). Even if these different kinds of totality remain to define, it is not too difficult to find illustrations, as we are going to see in the following. Nevertheless, it is rather within current cognitive psychology that we have to go searching for the rudiments of a functional mereology.

The hierarchy of wholes within the Lifeworld

It is doubtless in terms of dependence that we have to define the « independent object », to which Gibson attributes a fundamental role within ecological space . The independent object, in this sense, is a whole – even if there are doubtless relative totalities which form part of an independent object (for example, the face as a totality, itself forming part of a superior totality called a « head » in « Le viol » by Magritte), as well as larger totalities which contain several independent objects (for example, the set constituted by the ice cubes, the bottle of aperitif and the ice pail – of which the last one is missing in the Coliseum picture).

To begin with, the independent object is that which makes the difference between contiguity and factorality. But the examples which I have just mentioned show clearly that this difference is not as absolute as one would think. The independent object is first of all a qualitative notion. It should be remembered that for Groupe µ (1970 : 106ff ; 1977, 48ff, 70ff, etc.), ancient style, the metaphor as well as the metonymy are the result of two combined synecdoches, which go from generality to particularity, or vice versa. According to one of their examples, using « Caesar » to mean « De Bello Gallico » consists in realising a generalising synecdoche (from « Caesar » to « Caesar's life ») followed by a particularising synecdoche (from « Caesar's life » to the book « De Bello Gallico »). Now, as I have pointed out before (Sonesson 1989a : 48), it is not possible to go from just any part of the set « Caesar's life » to any other part. The term « toga » in the sense of « Brutus' dagger » fails to form any functional figure. The « set » must possess an organisation. The Groupe µ (1970 : 100) recognises this in passing, by observing, within brackets, that « the nauticity of the boat is situated in the rudder, but not in the cabin » (cf. Sonesson 1989a : 44). And what is true about a constructed object such as « Caesar's life », also applies to the objects of the perception – such as the boat. They do not form sets, but structured wholes.

Let us begin by presupposing an ontology of common sense, which consists in opposing things (or objects) to events. In fact, it can sometimes be more convenient to speak about spatial objects and temporal objects, respectively (while always allowing the term « objects », without qualification, to serve as the equivalent of spatial objects). This is thus what I would take to be the basic opposition within semiotic ecology: objects, which are (in an essential way) in space, and objects, which are (in a essential way) in time. As for the properties of things (and events, which I shall not discuss any further here), I think that it would be better to derive them in a mereological fashion, that is, as parts of the whole, which constitutes the object. As I have pointed out elsewhere (Sonesson 1989a ; 2001c), there are three main ways of dividing up an object: into its proper parts, in the narrow sense of the term (« the head », « the right leg », etc., in the case of a human body), into its properties (« male », « female », « adult », etc.), and into the perspectives from which it may be perceived.

In order to approach the notion of an independent object in the space, some assistance may be gathered from cognitive psychology. In my critique of structuralism (Sonesson 1989a), I made a wide-ranging use of the concept of prototype formulated by Eleanor Rosch, in order to prove that the world, at least as we perceive it, is organised in itself, and is thus only susceptible of being rearranged at a secondary, and thus « rhetoric », level. However, as I then observed (Sonesson 1989a, I.3.2 and III.5.1), Rosch studies only what I have called the intensional hierarchies, leaving aside the extensional hierarchies, which are exactly those which concern us in the current context. In other words, she studies the hierarchies of the kind « man – mammal – vertebra », while neglecting those of the kind « man – arm – hand ». In fact, Rosch and alia (1976) make no distinction between these two types of hierarchies, but in a note (p. 388), they point out that they have eliminated from their studies all the categories containing a relation of part to the whole. But they never justify this decision.

In agreement with a distinction of traditional logic, I thus separate the extensional hierarchies, in which each further subcategory occupy smaller and smaller space, and the intensional hierarchies, in which the extension remains constant. It is true that all the levels and all the elements in the first type of hierarchy, unlike those in the second type, « have a concrete existence » (p. 345). In fact, when we descend to the lower level of the hierarchy, the extension occupied by the elements becomes ever smaller in the first hierarchy, but there is no change in the second type. For example, a witch, an old woman, a woman, and a human being occupy space to an equal degree, but within the body scheme of the human being, every level of the hierarchy corresponds to a smaller part of space. According to a classic example, the same event can be described as the act of folding the finger, squeezing a piece of metal, releasing a spring, releasing the cock, firing a gun, firing a pistol shot, shooting at a man, killing a man, committing a murder, and saving four lives. This suggests that the same event (or, in other cases, the same object), while continuing to be thematic, can be redescribed at different intensional levels, with regard to an ever wider context into which it is integrated.

Thus, when descending the intensional scale, one always has to take a wider extension into account, exactly as when one goes up the extensional hierarchy, but the subject of the category, that which must be characterised, remains all the time the same. When a girl is painted in the wider context containing a sword, a charger with the head of a beheaded man, and a maid, she can be redescribed at another intensional level as being « Judith » ; but if the same girl is presented in the context of a charger with a head of a beheaded man and, furthermore, an old couple which can be identified as being her parents, she would correctly be redescribed as « Salomé ».

One may thus wonder if there is a basic level also in the extensional hierarchy, as Rosch has demonstrated in the case of the intensional hierarchy. Intuitively, it seems much more evident that there must be a privileged level in an extensional hierarchy than in the intensional equivalent : the body seems to have priority in relation to the arms as well as in relation to the couple and the group of people. However, the characteristics of the privileged level are perhaps different in the case of the extensional hierarchy: the super-ordinate categories may have fewer attributes in common (for example, « the group ») than the basic level categories (for example, « the body »). Whereas the subordinate categories (for example, « the arm ») seem to possess many attributes which we do not find at the basic level, averaged shapes as well as figures hidden in visual noise may be more easily identified at the basic level than at superior levels.

It could be interesting to repeat some of Rosch’s experiments in the case of the extensional hierarchies. It seems likely that also in this case the objects of basic level are more quickly classified into categories than the objects at any other level. But it is possible that completely different criteria must also be employed to determine the basic level of an extensional hierarchy : Gestalt factors such as common fate in the course of movement, perfect closure, etc. It is a variant of this first criterion that we find in Gibson’s conception. Here we shall suppose that an extensional basic level can be found. This basic level then would correspond to « substances » or to « independent objects » according to Gibson.

Already in Husserl’s work, we find a distinction between the totality and the aggregate. But it is doubtless necessary to posit several intermediate categories : let us suppose that two objects perceived together do not even form an aggregate. It is necessary to distinguish the case of the flock (Fig. 4 :Absolut Venice), the case of the Coliseum taking the place of the ice pail within a set of objects usually going together, and the case of the street lamp which, together with some pieces of linen, forms a totality only because they are presented in a photograph from a certain point of view (Fig. 5.Absolut Naples). This no doubt has something to do with different degrees of dependence, but the minimal system set up by both Husserl and Hjelmslev, which only distinguishes between the one-sided and the bilateral dependence, as well as the absence of dependence, will probably not be sufficient for our purpose. We have separate the environment (the context) which can still without ambiguity be identified with contiguity, from sets of objects often going together such as the ice pail with its ice cubes and its bottle, the accidental constellation as the street lamp and the linen, and totalities which at a superior level form new totalities as in the swarm.

The whole and its parts — from different points of view

Let us now return to the examples considered above, such as the cat-coffee-pot, Captain Haddock, « Le viol » and « Le promenade d’Euclide » by Magritte, « The wave » by Hokusai, and « La Bételgeuse » by Vasarely, to which I will add, among others, a few digital works by Inez von Lansweerde, « L.H.O.O.Q » and « Rasée » by Duchamp, and a whole series of advertising pictures for « Absolut Vodka ». Since no exhaustive description of this field of objects can be made at this point, I am going to work in the same way as the Groupe µ, creating compartments inspired by my previous considerations, which I will then try to fill in.

Let us begin with the case of contiguity, where, in spite of what was said above, presence and absence do not seem to go inevitably together (Cf. Fig. 8).

I have not been able to find any examples of an absence of an expected contiguity, that is, a suppression, which does not also suppose the presence of something unexpected, but the opposite case is very common, notably in advertisements. Indeed, the presence of an unexpected contiguity, that is, an addition, is realised by cases such trivial cases as a crown placed next to a bottle of gin of a particular brand, or the classic naked girl in a car. Even in the case of contiguity, nevertheless, we find more often the presence of an unexpected contiguity, combined with the absence of an expected contiguity, in other words, a substitution.

Several subdivisions involving the mereological character of the depiction are possible here: an object belonging to a particular category can be exchanged for another object of the same category, as for example the dog in « Las Meninas » by Picasso in relation to the dog of Velázquez. It is also possible to substitute an object for an environment, or vice versa, but so far I have not been able to find any examples. On the other hand, we can easily find cases of an object appearing in an environment, which is not, its own: for example, a submarine in an ice rink, which was present in an advertisement for the coffee brand « Gevalia ». Actually, this is really a borderline case: it is indeed difficult to say what is lacking at the place of the submarine, if we take the environment to remain constant. A skater perhaps, or some vehicle which moves on ice. The submarine is here thus the substitute for a rather vague class of things lacking at their place.

Another case can be observed, which is at the limit, this time, of factorality : the series or the set of objects in which one may observe the absence of an object of the set, concurrently with the presence of an object not forming part the same set. The best example which I have been able to find is an advertisement showing the Coliseum in the place of an ice pail in a constellation with a bottle of aperitif and a heap of ice cubes. As has been shown within the semiotics of objects (see Krampen 1979), there is a « syntax of objects » which allows certain things to go « quite naturally » together, such as the coffee-pot and the cup of coffee, the table and the chairs, etc. One thus perceives a rhetorical figure when the ice cubes and the bottle of aperitif, instead of being placed next to an ice pail, appear inside the Coliseum. But, once again, it would be nice to have a description that is adequate from a mereological point of view, of the kind of totality formed by such a set of objects.

At the level of contiguity, it is difficult to find examples of a presence of contiguities expected in unexpected positions, that is, a permutation. The bottle of gin can be indifferently to the right or to the left of the crown. Certain series of objects doubtless have their particular order, which can be modified, as the saucer below the cup, or the ice cubes inside the ice pail. However, intertextual realisations of this figure are more common, for example in the advertisement for the brand of socks « Kindy », perceived in relation to a famous still from a film with Marilyn Monroe having the title « The seven year itch » (cf. Sonesson 1992a, c). If we compare the Kindy girl with her model, we shall see that every time a property of Marilyn is lacking in the Kindy girl, it instead appears in the man at her side, and by comparing both men, the same inversion can be noticed. More simply, we observe that the man and the woman occupy inverted spatial positions. In a rather general sense, this is also true for the paraphrases of « Las Meninas » created by Picasso and Hamilton.

Let us now proceed to the cases involving factorality, beginning with the simplest varieties, which are also the most difficult to find, suppression and addition (Fig. 9).

The pure and simple addition is also difficult to find. A body with two heads functions as an addition of parts, for the same reasons that the suppression of a head has the opposite effect. A case in point would also be « L.H.O.O.Q. » by Duchamp (with regard to Mona Lisa). An object with a supplementary property is difficult to perceive as such. On the other hand, an addition of perspectives may easily be found, as in the case of Cubism, but also in Russian icons, in the religious sense of the term.

The richest domain of factorality is doubtless the result of the operation of substitution, in other words the presence of an unexpected factorality combined with the absence of an expected factorality. Before being able to analyse these cases, it is necessary to begin by making a mereological distinction (having its precedents within Gestalt psychology) between the cases in which a part is perceived on the background of a totality, and the cases in which the totality prevails over its parts. In the first set of cases, we can make the same subdivision as previously, namely between the substitution of parts, properties, and perspectives. We perceive at first the same object as we had anticipated, in which a part has been replaced by a part of another object (from the same category or another one). If the exchanged parts belong to the same category, the head of an animal, for example, it can be presented as a part of a human body, as in the work of Max Ernst ; and if the categories of the replaced elements are different, bottles may for instance be made to occupy the place of the pupils in the eyes of Captain Haddock.

As a result of substitution, the exchange of properties is perfectly possible. We perceive the same object as we had anticipated, with the difference that a property of another object (of the same category or another one) has been substituted for one of its properties. Thus, in a work by Inez van Lansweerde, the features of being male and adult have replaced the opposite properties of the mouth of a little girl. Similarly, blue colour can be substituted in the human body for one of the colours we find in the nature, as happens in certain comic strips, as well as in the statues of Hindu gods (according to examples given by the Groupe µ 1992). As an illustration of the substitution of perspectives, we have the inverted perspective of the Russian icon (according to Uspenskij) and the deformed perspectives constructed by Reutersvärd and Escher (cf. Sonesson 1989a, III.3.4).

In all these cases, the relation of factorality is dominated by the part that is lacking in the totality, but we must also envisage the inverse case, where it is the totality that prevails over the parts, which it absorbs. It is rather difficult to account for these examples, in which the replacement does not concern well-bounded parts, but involves more complex interrelations between several « independent objects » considered as totalities. The cat-coffee-pot represents the simplest case, where several totalities are merged into a unity ; but the borderline case is maybe rather the one in which one totality is entirely present, while the other one is only represented by a characteristic detail (the capsule of a bottle of juice added to an orange).

First, there is the case of the totality to which is added a part from another one. It is thus the relation of the first totality to (the part of) the other one that shows a divergence with regard to the norm. This case can be observed in an advertisement showing an orange with the capsule of a bottle of juice, but also in « Absolut Rome » where the handle-bars of a moped takes on the characteristic shape of the cork on a bottle of « Absolut Vodka » (Fig. 6).

Next is the case of two totalities merged into one, as it is the case of  the « chafetière », « Le viol » by Magritte, and an advertisement where we see the Ballentine bottle in the form of a snake or a light bulb. In this case, again, the divergence from the norm exists in the relation of the first totality in the second one.

In other cases, different kinds of relations are present between several totalities. A set of totalities can be organised so as to form another totality, that is, independent objects can occupy the place of the parts of another independent object. It is the case of the paintings by Arcimboldo, but also that of an advertisement for the supermarket B&W where fruits and vegetables form a crown. In the Arcimboldo pictures, the totality is a single « substance », a head, while the parts correspond to a whole collection of objects of the same kind, of which each is a totality in itself. Similarly, the substitution can be combined with a permutation, when the parts of a totality (as opposed to integral totalities) are redistributed so as to form another totality. It is the case of the marmalade jar considered as a totality produced by parts making up another totality, the slices of an orange, but given the form of the jar, as well as the onion made up of legs and hands. Both when the parts of the totality primarily perceived are built up of other totalities, and when they are redistributed parts of another totality, there is a diagrammatic resemblance.

There is nothing strange in the fact that a number of independent objects at the basis level together form a totality at a superior level, that is, a group. This happens all the time in the Lifeworld. This is the case, for example, with all the pigeons forming a flock. However, it is not very probable, in the Lifeworld, that this multitude of pigeons in contiguity with each other produce a figure having the shape of the bottle of « Absolut Vodka », but this is exactly what we see in the advertisement « Absolut Venice » (Fig. 1). As in the previous cases, thus, it is the relation between the scheme of the totality and the parts, which fill in its slots, which creates a divergence to the norm. However, in present case, the hierarchical relation of the totalities in the extensional scale is perfectly normal. What is strange are the properties (of shape, in this case) ascribed to the superior totality.

In the final example of factorial substitution, there is still a divergence in the relation between different totalities. This case involves several independent objects, which have nothing to do with each other, except because they are in contiguity with one another, and because, seen from a certain perspective (chosen by the photographer), they take on the same total shape as another independent object. Here, the distinction proposed by Husserl seems useful : in the Lifeworld, these objects form a simple aggregate, but in the picture, they have all the appearance of being a configuration. The example that serves to illustrate this case is the street lamp and the pieces of linen suspended over a street in Naples, which together, in « Absolut Naples », take on the shape of a bottle of « Absolut Vodka » (Fig. 5).

Finally, simple permutations are also possible: the presence of factoralities, which are expected, appearing in unexpected positions. It is thus the order and/or the positions of the parts, which produce a divergence with regard to the norm. We find this case realised in the various paraphrases of « Las Meninas », and, in combination with other operations, in « Le viol » by Magritte, as well as in «‑Absolut Athens‑», where the parts of a Greek column have been reorganised so as to form a bottle of « Absolut Vodka » (Fig. 7. :Absolut Athenes). Finally, there are cases of augmentation and diminution, that is, the presence of factoralities, which are expected, but presented in unexpected proportions. The same three examples can be mentioned here: parts of the body (the face and the lower part of the body) in « Le viol » have different proportions from what is expected, in the paraphrases of « Las Meninas » some of the persons have changed size, and the column acquires the semblance of the bottle of « Absolut Vodka », because some of its parts have become much bigger than the others.

Contrary to what takes place in the pictures functioning according to the regime of factorality, contiguity thus often produces a rhetorical effect, although a relatively weak one (except when it is combined with an opposition stemming from the second rhetorical dimension), by means of the simple presence of an unexpected elements, without requiring a specific absence. The simple absence is much more difficult to conceive. And when we proceed to the domain of factorality, the concurrent absence of an expected element and presence of an unexpected element is normally required to produce a rhetorical effect. So much for the structure of the rhetorical sign, conceived in terms of absences and presences.

The second rhetorical dimension : more or less similarity than expected

The second dimension of rhetoric is based on the continuity of the world of perception, which should allow us to separate different objects from each other, without producing too radicals breaks in the flow of experience. There should not be too much resemblance or too much difference. The signifier of the rhetoricalness of all the rhetorical figures consists, as I noted above, in the concurrent presence to visual perception of two elements, which are, in some respect, in opposition to each other. But the concept of opposition here lends itself to quite different interpretations : in fact, Haddock’s body and the bottles are different, but not really opposed, and the same thing applies to the cat and the coffee pot ; whereas the wave and Mount Fuji in Hokusai’s print, as well as the street and the roof in Magritte’s « Promenades d’Euclide » are more similar to each other in the pictures than Lifeworld experience would make us expect.

It seems reasonable to make explicit the requirement that two elements entering into a pairing must possess some further property in order to qualify as rhetorical figures, and that this property, if it is not a similarity perceived on the background of a dissimilarity, could just as well consist in a dissimilarity perceived on the background of a similarity. In relation to Lifeworld experience, both more similarity than expected, and more dissimilarity than expected, will produce a rhetorical effect (cf. Sonesson 1996b, c). If, in addition, we allow this latter case to be realised not only by objects merged into one, but also by objects, which are merely contiguous, this domain will suddenly emerge as being particularly rich in figures. Indeed, most of the cases discovered in the analyses of Floch, and the Greimas school generally, which are based, not on mere dissimilarities, but on oppositions, are of the latter kind (cf. Sonesson 1992a, c).

Trubetzkoy conceived oppositions as relations of dissimilarity on the basis of a fundamental identity, which would seem to be the inversion of the case mostly considered by Groupe µ, the similarity emerging on the background of a basic otherness (Hokusai’s « The wave », etc. ; cf. Sonesson 1989a, I.3.3.). Even if one may have some doubts about the authenticity of the binary oppositions which Floch finds in so many pictures, more clear cases can easily be shown to exist: thus, for instance, in the Kindy publicity, organised around inversion of the respective properties of the man and the woman in relation to the Marilyn Monroe poster; or, more clearly still, the minimal representation showing only a bottle and a tomato, the first being a prototype of angularity, and the second of roundness (Cf. Sonesson 1992a, c).

To the extent that two objects appear more or less isolated in the picture, and are not of the kind that would often appear together in reality, they will function as simple pairings, even though there is no similarity and no opposition between them (Fig. 10).

This really is the case with the cat and the coffee pot, as well as the bottles and Haddock’s pupils, which are simply different, rather than opposed to each other. The force of the rhetorical statement will naturally vary, as a function of the increasing or decreasing degree of tension between the terms involved. Different kinds of contraries may well turn out to be more important than logical contradictions, on one hand, or mere alterity, on the other hand.

Logical contradictions, as such, would seem to be impossible in pictures: even spatially separate visual configurations, which the Greimas school tends to describe in terms of contradictions, are probably more adequately described as being contraries. However, there may be a logical contradiction between a picture and its verbal intertext (for instance, the label « Ceci n’est pas une pipe » in one of the possible interpretations of Magritte’s painting which shows a pipe). While there may be an opposition between the plastic and the pictorial layer of the same picture, as in Brancusi’s « Bird » (where the plastic heaviness of the marble is opposed to the pictorial lightness of the bird supposedly depicted, as Groupe µ 1996 : 22 suggests), it is not a logically contradiction, but perhaps an anthropological universal.

Contraries may be of different kinds, and thus carry different import. The more trivial case is the one which we could call mere contrariety, for instance when a colour of a well-known object is changed into something we know does not exist, such as a blue human being; or, between verbal and visual signs, most of Magritte's titles. A more interesting, and probably rhetorically stronger case, is when contrary terms appearing in a picture are subsumed under important values of a culture, such as in a recent work by the group Casmo, in which the EU stars are opposed to a red little cottage epitomising traditional Swedish life (cf. Sonesson 1993b). Still stronger would be the case in which the contrary predicates themselves appear directly in the picture. The closest we can come to a logical contradiction must be the case in which the contrary terms are subsumed under anthropological universals known to every human culture, such as the opposition of the features « female » + « child » and « male » + « adult » in van Lamsweerde's above-mentioned picture. In fact, there may be an even stronger case, in which abstract contrary terms which are anthropological universals appear directly in the picture, e.g., the circle and square prototypes opposed to each other in the Mandala and the super-ellipse (Cf. Sonesson 1989a, I.2.5 ; 1996b).

The rhetorical effect of a picture depends of the force with which it deviates from the expectancies engendered by the perceptual world to which it refers. We have seen that this force itself may result from the type of opposition existing between the terms involved, as well as from the radicality with which it breaks down the integrative wholes found in the world of our ordinary experience.

The third rhetorical dimension : more or less reality than expected

The first and second dimension of rhetoric taken together still do not account for all rhetorical effects resulting from the picture being seen is a divergence in relation to the perceptual world. A third dimension, which concerns the « level of unreality », or fictional level, of the picture (cf. Sandström 1963; Sonesson 1994c) may also be distinguished (Fig. 11).

To begin with, the pictorial sign itself is a particular kind of fictionality, opposed, as we saw earlier, to the « reality » of the Lifeworld. If the picture is confused with reality itself, as when it is seen through a peep-hole, no rhetorical effect is involved, for there is not tension between fictional levels. But when the fictional status seems indeterminate, as is often the case with Baroque illusionary architecture, the sign function itself has been rendered rhetoric.

However, in a space already defined as being that of pictorality, expectancies may be contradicted in two ways, either by introducing within this frame something that is more real than the picture, or by allowing it to convey us on to something that appears to be less real than the picture itself. The case in which some part of the picture disconfirms the expected fictionality of the picture space is illustrated by the classical Cubist collage, in which real-world objects, such as a ticket, a newspaper page or the seat of a chair appears as part of the composition. A more extreme example would be when a real object is simply positioned, without any further addition, inside a frame or something else suggesting the delimitation of a fictional space.

In the other direction, adding further levels of unreality to that of pictorial space, distinctions may be established according to the increasing degree of tension between fictional kinds. Here we are concerned with pictures which, as far as the pictorial sign function goes, simply show some object or other which may conceivably exist in our socio-cultural Lifeworld, but which, because of certain circumstances, suggests another object, which they are not pictures of. One variety is when a part of the picture is seen to depict one independent object but which, in the context of one or several others, suggests the presence, without depicting it, of another object: thus, for instance, a cover picture of the French weekly « L’Expresse », where a shape depicting Nixon with a white face and a dark costume, placed in the position of the letter « I » in the title « Nixon en Chine », also suggests that letter ; or, to take a purely pictorial example, when a scrotum, in a picture from the exhibition « Bad Girls », where it appears over the head of a doll from the Disney film « Aladdin », can be seen as a turban. These examples are different from the pure factorality of the pupils in Haddock’s eyes, which must be projected onto the picture from the body scheme, or the pure contiguity of the Mannekin Pis in « Absolut Brussels », which is entirely supplied from our world-knowledge. We need some knowledge to see the « I » in Nixon and the turban in the scrotum, too, be they are also, but only indirectly, and at the abstract levels of configurations, present in the signs.

Good examples of other varieties of this kind of rhetorical figure are found in the publicity for « Absolut Vodka », in which the characteristic Absolut bottle appear each time in the guise of some landmark of well-known European cities. In « Absolut Munich », « Absolut Geneva », and « Absolut Amsterdam », some details suggest the shape of the Absolut bottle. Here we have a picture of a real-world object, such as a pair of « Lederhosen », a clock, or a house, the shape of which appears quite possible in itself, though perhaps not particularly probable, and which takes on a secondary meaning as a bottle, only because of our knowledge about what the Absolut bottle looks like and what is indicated by the text.

In other cases, we get pictures of real-world objects deformed so as to resemble other objects. Thus, in « Absolut Rome » (Fig. 6.Absolut Rome), the vespa is seen from a particular perspective and slightly modified so as to resemble the bottle; and in « Absolut Paris », the shapes and colours add up to something which is clearly identifiable as being of the general type of a Metro entrance, but too narrow to be a real instance. At first, one might be tempted to think of mixed objects, such as « la chafetière », as the extreme case of such deformations: but « la chafetière » really depicts parts of a cat and parts of a coffee pot, while « Absolut Paris » depicts an underground entrance and suggests an Absolut bottle.

Another case of secondary fictionality is when the fictive picture object does not correspond to any single detached object in the picture, but results from the contours of other objects, as is the case with the flock of birds in « Absolut Venice », as well as with the constellation of laundry and a street lamp in « Absolut Naples » (Figure 5). Again, these are not pictures of the Absolut bottle, but the bottle is certainly fictionally present in the pictures.

A particular case involving double levels of unreality is the picture depiction, in which both fictional levels are pictorial signs (cf. Sonesson 1994c). In what could be called the generic picture depiction, the picture depicted in not any particular, recognisable picture; it is simply a shape that suggests the category of a picture. In this case, the primary picture function masquerades as reality itself, while the secondary picture function present itself as being primary : this is the case of the pictures, complete with their frames, painted on the wall of Livia's house, and in other Roman villas.

Specific picture depictions, often known a « visual paraphrases », are of several kinds. No matter whether they refer to a single picture, some part or feature of it, or a class of pictures, the object referred to is always specific, in the sense that it may be pointed out outside the pictorial space in which it is simulated. There are two ways in which specific picture depictions may appear to create a tension between fictional levels: either because of some indeterminacy in the existence of the secondary fictional level, or because of some uncertainty relating to the independence of the primary fictional level. Of the first kind are pictures referring to whole classes of other pictures (or motives), such as, for instance, Manet's « Olympia » which suggests works by Goya, Titian, etc. Here are also found pictures in some borrowed style category, such as the « analytical cubism » which Hamilton takes over from Picasso in « Picasso’s Las Meninas », applying it to motives which Picasso has not treated in that way. Finally, there are the pictures referring to borrowed situations, such as the Hollywood scenes corresponding to no real films in the « Film Stills » by Cindy Sherman.

The other general type, in which it is the subject of the referring act which is in doubt, is represented by the depiction of specific recognisable pictures, such as Picasso’s, Hamilton’s and Witkin’s variations on Velázquez' « Las Meninas ». Another variety is when the depiction refers to personal stereotypes, or « ideotypes » of a particular painter, such as the characteristic bull and pink acrobat in Hamilton’s « Picasso’s Las Meninas », which are found in numerous works by Picasso, but not, of course, in his variation on the work of Velázquez. Finally, the limiting case of this aspect of the secondary pictorial functions is what has become known as « appropriated picture » after some works by Sherrie Levine, which are simply reproductions of the work of others presented as her own work.

Contrary to the two first dimensions, the third one depends on the fact of the picture being a sign, which is intermediate, at the level of fictionality, between the Lifeworld as directly experienced, and further levels of fiction within the sign. Once it is considered a sign, the picture may also be treated as an object of the Lifeworld, and thus be socio-historically overdetermined. The forth dimension is about the categorisation of the objects called pictures within different socio-cultural Lifeworlds.

The fourth rhetorical dimension : More or less divergence than expected from the usage categories

It may not be particularly fitting to talk about the socio-historical aspects as a fourth dimension : they can hardly by visualised as a scale, and they are actually present everywhere, modifying the other aspects (Fig. 12).

Devices which are highly rhetorical at one moment many become so widely used, that all rhetorical effect wears off in the end. This is in fact that aspect of rhetoric that was explicitly embodied in the Prague school model.

Something must be said here, however, on the relativity of rhetoric to pictorial categories. Pictorial signs may be divided into subcategories in at least three different ways (cf. Sonesson 1989b ; 1992a). First of all, it is possible to differentiate pictorial kinds from the point of view of their rules of construction, that is, the rules specifying which traits of the expression plane are relevant for conveying the content, and vice-versa. From this point of view, a photograph differs from a painting and a cut-out; and a linear drawing, to use traditional art historical terms, is different from a painterly one. Blends of these pictorial kinds are clearly rhetorical, for example the combination of a drawing and a painting, an oil painting and a water-colour, or a photograph and a painting, seen in many Modernist and Post-modernist works.

Then we could also distinguish categories of pictures according to the effects, which they are intended to produce (not the actual effects, which may vary, and which cannot really be known). Thus, in our society, publicity pictures are expected (among other things) to sell commodities, pornographic pictures are thought to stimulate sexual imagination, and caricature supposedly hold the depicted person up to ridicule (cf. Sonesson 1988 ; 1990). This is of course a purely social kind of categorisation. Subversions of these functional kinds are found when a news picture is offered where an advertisement is expected, as in the Benetton publicity; or when a caricature is presented as art, as with Duchamp’s Mona Lisa.

Thirdly, pictorial categories may be differentiated on the basis of the channels through which pictures circulate. The picture post card, for instance, follows another trajectory in order to reach the receiver than a publicity poster, a wall painting, a television picture, or the illustration of a weekly review. Art may override the rules characterising such circulatory kinds, too, for instance when art works are presented, not in a gallery but a shopping mall, as the group Casmo did some years ago.

These divisions, although based on empirically recognised categories, are not themselves « empirical », but must be considered theoretical constructs. The different categories coexist. Indeed, certain combinations of categories ascribed on the basis of construction, intended effects, channels, and configurations, tend to co-occur frequently in a given society. Thus, for instance, «‑fine art‑», as it was conceived traditionally before the advent of Modernism, involves a particular intended effect (« aesthetic experience »), certain channels of circulation (« art galleries », « museums », « art books », etc.), and even a small number of accepted modes of construction (« painting », « drawing », « wood-cut », etc.). In fact, Modernism (including Postmodernism) may be seen as a set of operations breaking this syncretism of functions apart. All works produced, in particular within the outgoing tendency of Modernism serve to abolish the expected correlation of construction, function, and circulation which is Western art : through its history, the Modernist work or art has been known to consist in anything which is not an oil painting (constructional kind) with aesthetic function (functional kind) which circulates through galleries and museums (circulatory kinds).

Conclusions

In the present text, I have suggested that, contrary to verbal signs, pictures are immediately rhetorical, because they posit at the same time their own similarity and difference in relation to the perceptual world. Therefore, I have argued, pictorial rhetoric must be based on the common sense structures of perception, overdetermined by the socio-cultural Lifeworld. The primary dimension of pictorial rhetoric, indexicality, derives its meaning from a divergence with regard the relative integration of neighbourhoods, independent objects and higher-order totalities. It may involve contiguity or factorality, but it ordinarily supposes both the presence of something unexpected and the absence of something expected at the same time. Iconicity, which makes up the second dimension, is founded on the expectancy of a relative differentiation of the objects of the world, which does not allow for too much similarity, nor too much dissimilarity. The third dimension depends on the fictional character of the picture as a sign, which can be confounded both by an all too direct experience, and by too numerous levels of fictionality. Finally, the fourth dimension concerns the picture as a social object, forming parts of certain categories of construction, circulation, and function. The advantage of this conception, in relation to the µ model to which it is indebted, consists in its closer attention to the common sense structures of perception.