Approches to the Lifeworld core of pictorial rhetoric (1)

Göran Sonesson

Lund University,

Sweden

In Visio. La revue de l’association internationale de semiótique visuelle, 1:3, 1996/97, 49-76.

Among the human sciences, it is a peculiar feature of semiotics (as well as of that particularly well-known part of it, linguistics) that it builds models of the objects it endeavours to analyse, which are then modified in confrontation with its objects of study. These models embody some particular theory about how meaning comes into being. In the case of pictorial semiotics, there has been, for some time now, three leading models of semiosis: those of the Greimas school, the Groupe µ, and the Quebec school; according the Fernande Saint-Martin (1994:2), however, one may also distinguish a fourth school, represented by the present author, which she calls the ‘Swedish school’. In this essay, I will in fact to some degree substantiate her claim, by referring to the work of some other ‘members’ of that ‘school’ (Castro 1996; M. Johansson 1996; T.D. Johansson 1996; Marner 1994; 1996a, b; person 1996). In that sense, what follows is meant to delineate some central issues of such a school.

In several earlier publications, I have tried to characterise the models of the three above-mentioned schools (Sonesson 1992a, c; 1993a,b; 1996a); briefly put, according to the Greimas model, all pictorial meaning derives from binary oppositions, the terms of which are distributed into two series which serve to separate at least two fields dividing the picture; the Quebéc school claims visual meaning is embodied in topological and Gestalt terms; and the Groupe µ holds that pictorial meaning emerges from the ruptures suffered by the norms which are posited to hold for all pictures (‘general norms’), or for the very pictures which they transgress (‘local norms’).

In some respects, these models seem irreducible; it is even possible that each of them simply is more apt to analyse some kinds of pictures (cf. Sonesson 1992a, c; 1993a, b). To some extent, however, it may be possible to integrate traits of the different models, building one synthetic model having a wider application and a more secure theoretical foundation. If so, this will involve starting out from the elementary predicates of the models, such as opposition, configuration, transgression, and the like. In the process, the meaning of these predicates will obviously change.

Configurations, in an even wider sense than that postulated by the Quebec school, are certainly basic to all perception, and thus to the perception of pictures, as are the topological space preceding Euclidean space in the experience of every child (cf. Sonesson 1992b; 1993a). Gestalts may sometimes emerge from, or be overridden by, structural relationships, and as ‘good forms’, or prototypes, they regularly serve as the normative instances against which deviations become visible (cf. Sonesson 1989a; 1992a). The relationship between oppositions and the transgression of norms is less clear-cut. On one hand, oppositions, realised as contrasts, may be a particular case of the rupture of norms: it constitutes, as expressed by Jakobson already in his notion of the projection of equivalencies to the axe of selection, more similarity than would be expected. But it also seems feasible, for instance, to look upon the rhetorical production of meaning as a particular case of an opposition, in which one of term is given, as part of that which is expected, and even prescribed, whereas the other term emerges as something new and unexpected, something which comes into existence here and now. It is this latter possibility which we are going to explore in the following.

Rhetoric characterises the way meaning is brought about by means of breaking the norms for how meaning would normally be produced. This means the rhetorical model can never be a total model of meaning: some meaning is already embodied in the norms which are overridden by rhetoric. In the visual domain this is not really such an extraordinary condition: it only means that, contrary to phonology, but just like visual structure (as Sonesson 1989a has shown against Lévi-Strauss and the Greimas school), rhetoric is regulatory rather than constitutive, i.e. it modulates a meaning which is already there, rather than creating it anew. In this sense, configurations are more elementary constituents of visual meaning; however, considered as operations applied by pictorial semiosis to the visual world, so as to change its import, oppositions and transgressions may be more important.

In constructing their visual rhetoric, Groupe µ actually spends most of their time studying how meaning is created in ordinary, non-rhetorical circumstances, establishing the principles which rhetoric is concerned to overrule. In their view, forms, textures and colours are the three principle elements of visual meaning. These elements may, or may not, be relevant at some level of analysis, but the real question is elsewhere: it concerns what makes up a norm, and what constitutes a transgression of it.

The notion of norms, in the work of Groupe µ (for instance when discussing local and general zero levels, 1992: 262ff), is intimately wedded to Greimas’ notion of isotopy. But the isotopy, as Greimas (1966: 96; 1970: 10, 188; 1972: 8) affirms and Groupe µ (1977:30ff) repeats, is involved with redundancy and coherence; and as a formal operation, it merely yields an unordered set of lexemes or, more exactly, semantic features. Indeed, the defining criterion, redundancy, and also homogeneity, rather than the isotopy itself, is employed to define the norm in Groupe µ:s (1992:262) later work: isotopy is said to involve ‘l’homogénéité d’un niveau donné des signifiés’, which amounts to a ‘redondance de sémes /…/ dans le domaine linguistique’ and a ‘redondance de déterminations ou homogénéité des transformations dans le domaine visuel’.

In the mathematical theory of communication, redundancy stands for total predictability, which obviously refers us to the relative certainty of (at least the idealised case of) the expectancies entertained by the receiver of the message. The same general idea of a subject expecting something to occur resurfaces when Greimas (1966:96) talks about the isotopy being ‘confirmed’, or when Groupe µ (1992:225) says, about rhythm, that it ‘crée une attente, qui peut être comblée ou déçue’. The central concept here, then, should not be redundancy, but rather expectancy, considered in relation to its possible fulfilment in time. This idea was expressed in the Bruner & Postman model of hypothesis-testing, later developed by the so-called constructionists, according to which perception consists in filling in the incomplete evidence of the senses with the help of systems of interpretation (cf. Sonesson 1989a, III.3.3.; 1996a); and it is also found, with an emphasis on the continuity of experience, in the Husserlean conception of time consciousness, projecting its protentions and retentions from a single point in time, which was later translated into the social domain by the Prague school of semiotics.

Once we conceive of the norm in terms of that which is expected, indeed, that which is considered to be ‘normal’, we realise that there are cases in which we will expect things to change rather than continue to be the same, that is, we will not expect the same units to re-occur. We certainly expect the whole human body to be human, but we also expect it to change from head into trunk, arms, and feet, at particular places, connections, and numbers. While a picture joining together the bosom of Mae West, the face of a general, and a football player’s hairy legs is undoubtedly deviant (Groupe µ 1978:18), the same holds true of another picture displaying multiple copies of Mae West’s bosom where other body parts should occur. Instead of the isotopy, a schema of interpretation is necessary, in order correctly to distribute expected recurrences and non-recurrences.

Although the funny story was used by Greimas to introduce the notion of isotopy, and is embodied in many of the pictures discussed by Groupe µ, its whole structure in fact consists in making us expect something to change – in fact, to change in an unexpected way, which means we expect the unexpected (Sonesson 1988; 1990; 1996a). In the domain of pictorial art, this is true of the mechanism of Modernism, which goes on for ever producing new differences, over and over again making us expect the unexpected to happen (cf. Sonesson 1994b). It is of course possible to retort that, as Modernism proceeds on its way of creating ever new movements, Modernism itself recurs redundantly; and that, as the joke reaches its point of rupture, it continues conveying the meaning ‘joke’. However, even if recurrence is expected on a certain level of abstraction (‘human being’), the interesting thing may be precisely which part of it is expected (‘head’ rather than ‘feet’), etc.

To apply adjunctions, suppressions, substitutions and, in particular, permutations to an isotopy is simply absurd, if we take the term according to definition, as an unordered set of redundant features. It does not matter whether we analyse Magritte’s ‘Le viol’ as a simple permutation, putting the breasts in the place where the eyes are expected, as Groupe µ (1992:303) suggests, or as a more complex projection of the entire female trunk onto the face, as I have argued elsewhere (Sonesson 1989a; 1992a); in both cases, there will be no rupture of isotopy, for both the eyes and the breasts, as well as the face and the trunk, form part of the same isotopy, which we might call ‘human body’, and even ‘female body’, admitting that, in our culture, the shape of the hair-do is a female as the breasts.

Instead of the notion of isotopy, we thus need a concept such as the scheme of interpretation, which accounts for the way the world is expected to be organised, permitting a comparison with the way it actually turns out to be. Such a concept is adumbrated in Groupe µ:s (1992:99ff, 149ff, 291ff) work, in the discussion of Palmer’s model of perceptual information processing, according to which objects are organised into parts, related to the whole by ‘subordination’, and integrated into more extended objects, related to them by ‘superordination’; in addition to which they also have so-called global properties distributed over their whole extension. Groupe µ even suggests that relations between units of the same level, ‘co-ordination’, and those between units in time, ‘pre-ordination’, should be added to the model.

Thus, the model becomes similar to the one I introduced (in Sonesson 1989a,I.3.2), on the basis of Koestler’s notion of holarchy, Benveniste’s distributional and integrative relations, and Rosch’s notion of basic levels. The difference is that I did not take account of ‘pre-ordination’, and that Groupe µ does not attend to Rosch’s (et al. 1976:383) evidence for the existence of a level of organisation, a basic level, which is privileged, in so far as it describes ‘intrinsically separate things’. Moreover, Groupe µ:s notion of global properties turns out to be very restricted, corresponding more to what I (in Sonesson 1989a,I.3.4.), following the Leipzig school of Ganzheitspsychologie, have called configurations, as opposed to other holistic properties. They thus ignore, or at least fail to single out, such holistic, non-configurational properties as ‘angularity’ and ‘roundness’, which have been shown to be perceived before any other features by children, and to be differently located in the brain (see references and discussion in Sonesson 1989a,I.3.4 and I.4.4.). They are therefore unable to account for some subtle effects of visual semiosis.

If the norm is equivalent to an isotopy, then it must be characterised by homogeneity, and we will expect heterogeneity, and nothing else, to be rhetorical. If, however, norms cannot be resolved into isotopies, there is no reason to expect all heterogeneity to be rhetorical, or all homogeneity not to be so.

One may wonder between which elements homogeneity and heterogeneity are supposed to obtain. When discussing the so-called ‘rhetoric of transformations’, Groupe µ (1992:295ff) distinguishes homogeneous and heterogeneous transformations, not by comparing the state of the world with that which is rendered in the picture, but in relation to the result internal to the picture. A heterogeneous transformation, in this sense, in one in which different transformations have applied to a homogeneous state of the world rendering the picture itself heterogeneous.

It is natural, then, that Groupe µ should claim transformations to be rhetorical only to the extent that they apply heterogeneously; for this means transformative rhetoric will indeed depend on a rupture of isotopy. However, when discussing concrete transformations, Groupe µ (1992:307f) also attends to some homogeneous transformations, and this is also the reasonable choice, if historical circumstances are taken into account, and also, as I will argue, when considering the structure of the iconic sign. But there is no rupture of isotopy in homogeneous transformations: they rather constitute an infraction of the historical norm stipulating the way in which objects are to be rendered in pictures.

Pictorial rhetoric, in Groupe µ:s restricted sense, thus looms large in Cubist collages, in Dadaist works of art and, more recently, in much Post-modernist art and a lot of advertisements. On the other hand, the fact of transforming a real personage into a stick man is not in itself rhetorical, whether the drawing is in black and white or in several colours; but to introduce a figure in black-and-white inside a drawing which is otherwise coloured is to produce a rhetorical figure. In the same way, the red flag appearing in Eisenstein’s otherwise black-and-white Potemkin-film is rhetorical (p.295). If this is the case, it seems to me that the ‘rhetoric of transformation’ is a misnomer: what is rhetorical is then the combinations, not the transformations. As more extreme case, I offer the example of Picasso’s ‘paraphrase’ of ‘Las Meninas’, or, even better, Hamilton’s ‘paraphrase’ of ‘Picasso’s Las Meninas’, which contains a still more complete sample-card of different iconical transformations.

Now consider the kind of transformation by means of which an ordinary landscape is made into an impressionistic painting. This transformation, I submit, is doubly rhetorical: first, it breaks the general norm, in vigour at the time, for how the perceptual appearances of the world are to be rendered in fine art; and, in the second place, it breaks the norm, still in place, for how a normal picture would render those appearances. This could equally be formulated in terms familiar from the Russian Formalists (from Sklovskij and Jakubinskij): the impressionistic painting disrupts the habits of perception, which are acquired (thus becoming ‘automatized’) in our ongoing everyday experience, not of standard language, but of another standardised medium, ‘non-artistic’ pictures, thereby making them ‘strange’, or ‘actualised’, for us; and at the same time it breaks down the expectancies created by earlier artistic movements, which where once revolutionary, but has since then hardened into standardised artistic forms. In the later re-formulation of the Prague school, it transgresses the norms for the standard medium in which it is couched, but also the norms set up by immediately preceding artistic movements.

I would argue, however, that there is still a third norm with which the impressionist picture breaks, in the improbable company of the contemporary Artistic Salon picture: the norm of the ordinary world, of ‘real’ reality. Whether the ordinary, three-dimensional, living man is transformed into a stick-man, a number of contours, some patches of a photographic plate, a series of pointillist colour spots, or some Cubist shapes, the result is certainly heterogeneous to the point of departure. There is a sense in which every picture is rhetorical in relation to what is represents, as is indeed every iconic sign: they create an impression of similarity on the background of a fundamental difference. But this impression of similarity then gives way to the perception of a new difference (Sonesson 1989a; 1994e; 1996c). In verbal language, rhetoric comes much later: there is never any suggestion (except perhaps in onomatopoetic words) of a primary similarity from which the difference has to be regained (Sonesson 1996b).

According to Groupe µ (1992:262ff), we should distinguish between general and local norms. A general norm is established in a code, existing prior to the occurrences which conform to it, as well as to those which transgress it. Thus, according to a norm of the perceptual world, a head is normally attached to a body below it, and if there is no body present, or the body is above the head, norms have been transgressed. Local norms are established by the very complex of signs which then goes on to break them. Thus, many of Vasarely’s works feature endless horizontal and vertical repetitions of circles of a particular size and colour, but the expectations that this will go on for ever is suddenly deceived with the appearance of a square, or a smaller, or differently-coloured, circle. This distinction is useful, but it is not sufficient. Some place should, for instance, be made for genre norms, which are mentioned several times, without being given any systematic status.

In the Potemkin example quoted above, two norms, one local, and the other general, are really broken at one stroke: the general norm resulting from the technical possibilities of the time, which prohibits colour films (the red colour was painted onto the film strip), and the local norm making us expect a film which is otherwise black-and-white to be entirely black and white. As seen today, however, the film as a totality breaks another, more general norm, which requires films, at least of this kind, to be made in colour. A similar case could be made for the rhetoricalness, to us and to the public of the time, of the first sound movies.

This means that norms will be relative, not only to historical moments and particular societies, but also to certain pictorial genres. If so, it will no longer be possible to conceive of semiotics as a domain outside history and social parameters, as Groupe µ, like the Greimas school, wants to do. As I have often pointed out (e.g. in Sonesson 1989a,I.4.), following Prieto, who himself quoted Saussure, semiotic objects only exist for their users, that is, they have only the kind of existence that they are accorded by their use in a given social group; and thus, once we pretend to go beyond sociality, there is nothing left to study.

The relevance of sociological concepts to semiotics must, in a way, be limited: otherwise, semiotics will be resorbed into sociology, or sociology into semiotics. Thus, only those social features should be considered relevant to semiotics which, in some way or other, impose constraints on the very formal nature of the sign systems studied... It seems to us that the concept norm, which plays an important part in the theory developed by Groupe µ, must be a social concept (although it is not treated as such by Groupe µ, with the exception of those passages in which they are concerned to explain why they do not want to explore it further). Perhaps, then, we may take the norm to be precisely that part of sociology which has to be retained by semiotics (cf. Sonesson 1996a).

If, like Groupe µ, we take a rhetorical view of semiotics, then norms are relevant to the extent that they specify the rules which might be broken by means of rhetorical transgressions. But such rules are clearly different for different picture types, let alone different kinds of ‘visual signs’: thus, although there may not be any rhetorical effect in leaving out the frame of a contemporary work of art (as suggested by Groupe µ 1992:389f), there certainly would be one if the work is question is a kitsch painting; and if the picture is not a piece of fine art, the presence of the picture frame would instead be that which breaks the norm. Again, it is a pity that Groupe µ (1992:377) dissolves the frame into the more abstract concept of edging, refusing to distinguish the frame and the plinth: for indeed, a plinth under a picture, or a frame around a statue, would certainly carry rhetorical import.

This model is an adaptation of the conception of perception propounded by Husserlean phenomenology, with an added social dimension (cf. Sonesson 1992a; 1993c; 1994c). We may therefore restore to the model its general import, applying it to all objects of perception, while retaining its social character. In this sense, the production of any artifact given to perception involves norms of different levels of generality, and so does its concretisation into the perceptual experience of a given subject of perception. On the other hand, the idea of there being a perpetual movement in which de-automatized norms are exchanged for automatized ones, an ever-repeated dialectics of ‘struggle and reformation’ (in the terms of the Prague theses) applied to established artistic forms, really reproduces the conception of art presupposed, and even explicitly formulated, by the exponents of Modernism, as I have argued elsewhere (Sonesson 1994b).

When a concretisation takes the form of the reading of a literary work or the perceiving of an art object, it is the sake of the individual subject, and cannot be seized by analysis. However, in some cases, concretisation, in its primary phase, involves a collective subject, which creates a new artifact, offered to the perception of further perceivers: this occurs with the execution of music, the dramatisation of a theatrical work, and the representation of a ballet. In these instances, the concretisation may undoubtedly, instead of following the norm of an ideal rendering, opt for its transgression. Let us consider the case of ballet, which is made up of purely visual signs, which are only slightly pictorial. In the version of the ‘Swan Lake’, staged by Mats Ek for the Cullberg ballet, transgressions of norms occur on several different levels: transgressions of the code of ‘classical ballet’ which ‘Swan Lake’ shares with many other spectacles of dancing (movements which are not allowed by the code, contiguous with other movements which serve to establish ‘classical ballet’ as the local norm); transgressions of the exemplary work of art which ‘Swan Lake’ constitutes in our culture (most obviously, modifications of the ‘romantic’ narrative structure in the direction of more trivial reality, in which, after the peripatetic moment, in which the White Swan is saved from the sorcerer, the prince retains the Black Swan as his mistress); transgressions of the norms of concretisation (which are harder to exemplify, because they really have to be shown: it is the particular way in which the grande chassée is executed).

In terms of the Prague school model, isotopy, as referred to above, would mean that concretisation not only totally coincides with production, but this coincidence is effortlessly obtained. According to an idea, suggested both by Moles and Lotman, the sender and receiver of any situation of communication start out with codes which only partially overlap, struggling to homogenise the codes as the communication proceeds: isotopy would be the kind of condition in which this homogenisations is obtained as a matter of course. In both cases time consciousness, including a difference between real and projected time, would have to be integrated into the model (cf. Figure 2 and Sonesson 1994c; 1995b).

Beyond the particular socio-cultural world, there is, I submit, a number of universals traits taken for granted in every conceivable world inhabited by human beings. The description of these universals have been made in terms of what Husserl and his followers, Schütz and Gurwitsch, have called the science of the Lifeworld, as well as in the form of James Gibson’s ‘ecological physics’ and Greimas’ semiotics of the ‘natural world’ (cf. Sonesson 1989a, 1994a,c,e; 1996a). Husserl, Gibson, and Greimas all invented this science because they realised that the ‘natural world’, as we experience it, is not identical to the one known to physics, but is culturally constructed. Like Husserl’s Lifeworld and Gibson’s ecological physics, but unlike Greimas’ natural world, I will suppose this particular level to be a privileged version of the world, ‘the world taken for granted’, in Schütz’s phrase, from the standpoint of which other worlds, such as those of the natural sciences, may be invented and observed (cf. Sonesson 1989a,I.1.4, I.2.1, and passim).

It is a basic property of the Lifeworld that everything in it is given in a subjective-relative manner. This means, for example, that a thing of any kind will always be perceived from a certain point of view, in a perspective that lets a part of the object form the centre of attention. What is perceived is the object, though it is always given through one or more of its perspectives or noemata, which themselves are unattended. Gibson observes that, when we are confronted with the-cat-from-one-side, the-cat-from-above, the-cat-from-the-front, etc., what we see is all the time the same invariant cat. To Husserl, this seeing of the whole in one of its parts is related to the etc principle, our knowledge of being able, at any one point, to turn the dice over, or go round it, to look at the others sides.

Everything in the Lifeworld is given in ‘open horizons’, that is, reality is not framed off like a picture, but goes on indefinitely, however vaguely indicated. Beginning with our centre of attention, the experienced world gradually fades away, without there being any definite limits, and we only have to change the centre of attention in order to extend the field of distinct experience. Every object has an outer horizon, i.e. the background field of other, nearby objects, and an inner horizon, the parts and attributes that are presently out of view or just unattended. To both the horizons, the etc principle applies.

Every particular thing encountered in the Lifeworld is referred to a general type. Typification applies to all kinds of objects, even to human beings: according to Schütz, other people, apart from family members and close friends, are almost exclusively defined by the type to which they are ascribed, and we expect them to behave accordingly. ‘In perceptual experience, the spatial shapes of things are determined only as to type – a margin of latitude is left for variations, deviations, and fluctuations’ (Gurwitsch 1974:26). Thus, there are no circles in the Lifeworld, only things with ‘roundish’ shapes, with ‘circular physiognomy’. Indeed, the ‘good forms’ of Gestalt psychology, and the prototypes of Rosch’s theory, are clearly typifications (see Sonesson 1989a,I.2.1.).

Closely related to the typifications are the regularities which obtain in the Lifeworld, or, as Husserl’s says, ‘the typical ways in which things tend to behave’. In fact, once an object has been assigned to a particular type, we know more or less vaguely what may be expected, or rather protained, from it in the future, and we can then learn to manipulate desirable changes ourselves. Many of the ‘laws of ecological physics’, formulated by Gibson (1982:217ff), and which are defied by magic, are also such ‘regularities /that/ are implicitly known’: that substantial objects tend to persist, that major surfaces are nearly permanent with respect to layout, but that animate objects change as they grow or move; that some objects, like the bud and the pupa transform, but that no object is converted into an object that we would call entirely different, as a frog into a prince; that no substantial object can come into existence except from another substance; that a substantial detached object must come to rest on a horizontal surface of support; that a solid object cannot penetrate another solid surface without breaking it, etc. Contrary to what is often believed, children spontaneously believe in non-magic (Gibson 1982:218); indeed, it is on this background that magic becomes pertinent, as a kind of rhetoric of the ‘natural world’.

The Husserlean description of regularities accords well with the notion of abduction, which Peirce puts alongside the more familiar procedures of deduction and induction, and which reasons from one particular instance to another, not, however, exclusively on the level of individual facts, for the facts, Peirce tells us, are mediated by certain ‘regularities’, principles that are tentatively set up or taken for granted. Peirce wondered how it was possible for so many abductions to prove right, postulating a natural instinct as an explanation. Actually, there is an infinite number of ways to relate facts, but most of them would seem to be humanly inconceivable. The limited number of alternative abductions being really proposed may be due, not to a natural instinct, but to the commonalty of the most general organisational framework of the Lifeworld.

Groupe µ (1992:268f) claims that the pictorial sign underlies the principle of concomitance, according to which the limits of the units of the pictorial layer coincide with those of the units of the plastic layer; and that the three elements of the plastic sign, texture, colour, and form, obey a similar principle, also termed coextensivity. Now, clearly concomitance is something which is expected to obtain, because it is a ‘regularity’, as Peirce would have said, observed in the ordinary, perceptual world, and is then transferred as a matter of course, by abduction, to pictorial signs. It is indeed in the Lifeworld that we would normally observe the coincidence of these general properties, as we will that of more particular features, characterising the Roschean prototypes. This coincidence defines the notion of detached object central to Gibsonean ecology.

There is another property, ascribed by Groupe µ to pictures, which to me seems to be a regularity of the Lifeworld: when discussing the collage, Groupe µ (1978:17ff) claims that the picture norm, on the plastic level, is ‘isotopicality’, or more precisely, ‘isomateriality’ of the expression matter, and ‘allotopicality’, which is to say, ‘allograduality’ of the expression graduality – more simply put, that a single material should be employed to compose the picture, in order to express such distinctions between the intervening signs which make it possible to hold them apart (‘une répartition suspendue à la fonction de reconnaissance des signes iconiques’; 1978:18). In contrast to these norms, it is concluded, collages may therefore be isogradual and isomaterial, or allogradual and allomaterial – and even, as it is added in Groupe µ (1979:185), isogradual and allomaterial. But this only means that the picture requires some intermediate degree of order, which permit us to make distinctions and distribute meaning in the world of our experience. Indeed, Groupe µ (1984:20f; 1992.39ff) itself has hinted at the idea of there being two separate tendencies in rhetoric, the ‘adjunction of order’, and the ‘adjunction of heterogeneity’– which will leave the Lifeworld somewhere in between, more or less where the Tartu school model will soon permit us to rediscover it.

The ‘terrestrial environment’ of all animals has continued to possess certain simple invariants during the millions of years of evolutionary history, such as the earth being ‘below’, the air ‘above, and the ‘waters under the earth’ (Gibson 1966: 8ff). The ground is level and rigid, a surface of support, whereas the air is unresisting, a space for locomotion, and also a medium for breathing, an occasional bearer of odours and sounds, and transparent to the visual shapes of things by day. As a whole, the solid terrestrial environment is wrinkled, structured, at different levels, by mounts and hills, trees and other vegetation, stones and sticks, and textured by such things as crystals and plant cells. The observer himself underlies the consequences of the rigidity of the environment, and of his own relationship to gravity. Also linguists trying to explain the existence, in all languages, of a set of small words designating the spatial and temporal dimensions of the environment, have found it necessary to postulate a basic framework of the experimental world, determined, in part, by gravity (cf. Miller & Johnson-Laird 1976).

Inside ‘ecological physics’, in Gibson’s sense, there must be some kind of ‘social physics’, not exactly in the Durkheimian sense, but on the micro-level. Schütz and Mead have talked about the array of ‘things’ of the human world which are peculiar in being ‘at hand’, occupying the ‘manipulatory sphere’; and Wallon has discussed the ‘ultra-choses’, which are outside this sphere, but are seen from there. Even these humble things do not only have a use, but are also there, as Lévi-Strauss would have said, to think with. In this sense, I have referred elsewhere to the hierarchy of prominence of Lifeworld things, and I have in fact been using such a scale in two different, but complementary, ways (see Sonesson 1989a; 1992a; 1994c, e; 1996a; c) On the one hand, objects, such as the human body itself, in particular its face, but also common objects like chairs, must be so central to the human sphere, that they will be recognized with only scant evidence, even though the invariants embodied in a particular picture are found in other objects as well. In this case, the objects at the highest levels of the scale stands the best chance of being selected. On the other hand, I have argued that only objects low down on the scale will be recognized as susceptible of embodying a sign function, without being particularly designated as such, which is true, in our culture, of a sheet of paper or a canvas.

In their study of the basic metaphors which underlie both poetry and ordinary language, Lakoff & Turner (1989:160ff) describe a ’cultural model’ which they call ’The great chain of being’. This model, which ’places beings and their properties on a vertical scale with 'higher' beings and properties above 'lower' beings and properties’(p.167), has been studied by historians of ideas since the time of Lovejoy, but Lakoff & Turner show it to be still current and active in a lot of everyday thinking, as for instance in ordinary adages. This ’commonplace theory about the nature of things’(p.170) would only stand in need of being slightly amended in order to account for the naturalness with which surfaces stand for scenes, rather than the reverse. It is simply an imaginative extension of our position in the middle of the terrestrial environment, from which the directions go up and down.

The world of our experience being subjective-relative, the ego stands at its centre, at the origo, as Bühler (1934) once called it, of its spatial and temporal determinants: the I-here-now. The things which are close at hand are familiar; those which, in some way, are placed far from the origo appear strange. In addition, some of the strange things may appear to be at the limit of our perceptual world, as ultra-things, either above us, or below. Given these presuppositions, two operations, which are inversions of each other, become possible, as Anders Marner (1996b) as pointed out: familiar things may be treated as strange, which is the well-known ‘estrangement’ of the Formalists; and strange things may be presented as familiar, which is a process often found in advertisements, when, for instance, a product on sale is integrated in a typical home, or attributed to the proverbial house-wife. Moreover, objects on the human scale may be treated as if they were ‘high above’, in some metaphorical heaven, or ‘deep below’, in rhetorical hell: these are the characteristic modes of access to surrealism favoured by Breton and Bataille, respectively, in the interpretation of Marner (1996a). But such systems of ascending and descending metaphors are easily found also in other domains.

Figure 3. The Lifeworld core of rhetoric (as suggested in Sonesson 1996c)

This could be seen as the Lifeworld core of rhetoric (cf. Figure 3). In given circumstances, something which carries the general meaning of being strange or familiar most probably contains also a more particular content, derived from the individual properties of the things brought into contact (telling us not only that something is strange or familiar, but that, for instance, a Greek column and a bottle of ‘Absolut’ have something in common). I think we may, however, add a few things to Marner’s ‘great cross’: inside the origo, where the I-here-now is located, there is not only the sphere of that which is most well-known, but also the beginning of a series of other scales having their opposite end somewhere among the ultra-things: that which is most real, most understandable, most personal, etc. All these scales determines the possibilities of rhetoric.