1.1.The Coloreme

Following considerations of both the structure and the mechanisms of visual perception, we have defined the basic unit of visual language as the coloreme. It corresponds to that aggregate of visual variables perceived in the visual representation by way of an ocular fixation, or focus of the gaze. One might compare it to an area corresponding to the “point of ostension” of a finger directed upon the opaque surface of the same representation.

A coloreme is defined, therefore, as the zone of the visual linguistic field correlated to a centration of the eyes. It is constituted by a mass of energetic matter presenting a given set of visual variables. This primary element of visual language is made up, from a semiotic point of view, of a cluster of visual variables, the properties of which will be described and elaborated upon in a later chapter.

Through the perceptive mechanism of ocular fixation (Arnheim; 1971), the coloreme is immediately structured as a topological region. Visual perception is realized through a positioning of the eye in the direction of the visual field, called an ocular centration or fixation. This fixation can obtain different visual information according to the specific mechanisms of vision it borrows: the fovea, the macula, or the peripheral vision.

The fovea centralis is a small pit or depression at the back of the retina furnished by the strongest condensation of cones, or receptors of colors, forming the point of sharpest vision, but within a distance of 15 feet and along a small arc of one to two degrees. The macula is a yellowish region surrounding the fovea, possessing a lower amount of cones, but covering a wider arc and a further distance. The peripheral vision is the name given to the rest of the retina, quite poor in cones but rich in rods, or receptors of tonality values; it lacks precision and susceptibility to colors, but can apprehend a wider portion of the field (over 90 degrees) as well as quite subtle movements in that field (Hall, 1966).

It should be noted that a viewer can at any moment use one or the other of these mechanisms of vision, or change from one to the other, thus modifying the visual information he is getting from the world. And in spite of apparent contradiction in terms, an ocular fixation may resort to peripheral vision and produce a blurred interpretation of the visual field when trying to encompass a larger and more distant area. Peripheral vision may also serve other purposes, as can be inferred from A. Ehrenzweig’s propositions (1967), which will be discussed later.

Given the richer visual potentialities of the two central sources of vision, we have defined as a coloreme the area of the visual field which is the product of two interrelated zones: (1) a central area more precise, dense and compact, corresponding to foveal vision; and (2) some peripheral layers, less dense, but still rich in colors, corresponding to macular vision.

On the objective plane of representation, the coloreme corresponds to any colored quantity located at the termination point of an ocular fixation and contributing to the formation of a visual percept. The very definition of a percept as an entity structured as a field of forces (Gurvitsch, 1957; 114) requires that the minimal unit ascribed by semiotics to visual language be: a material zone sufficiently large for perceptual mechanisms to be realized. In other words, it cannot be microscopic or in any way inaccessible to normal perception. On the other hand, it cannot be too large, so as to permit an efficient realization of foveal and macular vision.

Invariable in its structural properties of interrelating a harder core to surrounding layers, the coloreme offers the topological characteristic of ‘elasticity’ in its dimension (and form) which excludes the very notion of static metric measurement. This unit reflects the functional dynamism of the organ of sight which is capable of amplifying or reducing the dimensions of the foveal percept within its proper limits of an arc of two degrees, as well as of integrating some of the scope of macular vision.

It is to be expected, however, that the characteristics of a coloreme will be modified when larger portions of macular vision are jointly used to consider the former stimuli of the foveal zone. In the same way the use of peripheric vision will greatly modify the experience of the visual field. While the latter will play a most important role in establishing relations among wider combinations of coloremes, it is not involved in the perception of the basic unit of visual language.

This definition of the element of visual language as a continuous and spatialized topological entity, endowed with somewhat fuzzy boundaries, would appear incompatible with the accepted view of the phonemic unit of verbal language, only if one neglected to consider the actual elasticity of this latter notion. In effect, the phoneme is constituted by a cluster of auditory variables within extended limits, and it can also, according to the individual case, play the role of a morpheme or even of an entire phrase. As put forth by L. Hjelmslev (1963; 63), a verbal phrase is seen:

. . . to consist of only one clause and a clause of only one word. This phenomenon is constantly turning up in the most various texts. In the Latin imperative “i” / “go” or in the English interjection “ah” we have an entity that may be said to be at the same time a sentence, a clause, and a word. In each of these cases, also, we find a syllable that includes only one part of a syllable.

In the same way that the acoustical analysis of sounds provides little information as to the nature and function of the basic units of verbal language (the phoneme), theories of optics are only a prerequisite to the understanding of visual language. They have to be complemented by a definition of visual variables out of which visual semiotics can be constructed.

In regrouping a finite and specific ensemble of visual variables, the coloreme offers itself as an autonomous—and by its structure, a differentiated—unity, constructed by perception. Its global characteristics are quite different from those of the individual visual variables that constitute it. When one variable is modified, the coloreme is itself entirely transformed along with its interrelations with other coloremes. Visual variables themselves offer a wide potential for differentiation, but no single one among them can play the role of a basic element, since they are all inextricably bound up in any given percept.

If one can consider and analyze them as abstract categories, their behavior and functions within the perceptual process can only be known through the observation of what is concretely qualified and quantified in the visual field. It is the concrete and simultaneous conjunction of variables which constitutes the material structure of each of the coloremes. Once regrouped and energized by perception, these plastic units will enter into interrelations with others, according to a specific group of organizational laws, still totally dependent on the nature of the variables perceived.

In the past, there have been attempts to elevate one of these visual variables over another or to conjugate two of them, as the basic element of visual language. This very temptation has, in fact, and for a considerable time, slowed down the development of visual semiotics.

The historical record is replete with arbitrary choices made in this domain. One may recall the efforts of André Lhote (1967; 91) to determine the constants in the midst of visual variables he called the “plastic invariants.” He first listed: “the design, color and value,” but added subsequently: “rhythm, decorative character, reversal on the plane and monumentality.” In a pioneering work, R. Passeron (1962) identified the “technical elements” as being color, value, line, and “la pate.” For his part, Kandinsky (1976) made the point, the line, and the plane the basic elements of his “grammar of creation.” We will return to this latter proposition in our study of the Basic Plane.

More recently, Felix Thürlemann (1982; 20), while denying the possibility of determining the primary elements of pictorial language, did not hesitate to identify them with the duality of color-form, which Vasarely (1970) had previously identified as the very basis of his system of production of visual representations. But the indecision as to the clear definition of what should stand as color or form caused Thürlemann to retain as minimal units of the plastic discourse only the “well-formed” aggregates. These “color-endowed forms” will harmonize eventually with a list of verbal labels, since both proceed from the same reservoir of mental iconic images.

One could ask, given the omnipresence of the visual variable of color, as any line, form, or texture, possesses a color, why it has not been retained as the fundamental element of visual language, superseding all others. In effect, visual perception can be realized only by the mediation of colors which correspond to different quantities of reflected light. Essentially an energetic phenomenon, the chromatic dimension, as we will see further on, is also totally dependent (within the perceptual process) upon the dynamics of the other visual variables. For this reason, the term ‘coloreme’, which underlines in its formation the importance of color in visual reality, is used as a means of designating the minimal ensemble which regroups, together with the diverse qualities of color, all the other visual components in every point of the visual field.

The attempt to isolate a given visual variable, or even a handful, as basic elements cannot be fruitful. It would be equivalent to a decision in verbal linguistics to undertake the study of a chosen acoustic variable among those which constitute the phoneme (the sound’s pitch, tonality, timbre, harmony, etc.) as the basic unit of verbal language. Furthermore, in contrast to the sequential and often isolated mode of presentation of each phoneme, visual variables are immersed in a simultaneous predicament where their perceptual aspect is constantly being modified.

The case is the same within a coloreme or between coloremes. In regrouping these, one should not minimize or ignore the action of some visual variables to the benefit of others. One region cannot totally oppose another by the sole specificity of one variable, its size, color or texture, and so on. Sometimes separated by contrasts within one visual variable, regions may be joined by liaisons produced by other variables.

One should be particularly careful not to isolate from the visual context those regions which can be endowed with the characteristics of a “good form,” without due consideration of the nature of those variables which constitute it besides that of form.

One should be constantly aware that abstract considerations relating to the nature of the visual variables can be misleading, as none can exist independently of the others and none can be apprehended in an identical way in two places of the visual text or at different moments. As each coloreme, by necessity, regroups the totality of a set of visual variables, any process of junction/disjunction in visual language is a dialectic equilibrium realized between two or several coloremes and not between different aspects of a unique visual variable.

By definition—the coloreme being the correlative of foveal fixation on a small zone of the visual field—it is impossible to perceive several coloremes through one single centration. As a consequence, given the unique resources for vision of color of the foveal or macular instruments, it is deemed impossible to perceive adequately a large expanse of the visual field through a unique ocular fixation.

In order to perceive such a large expanse adequately, one must effect several successive centrations of the visual field (Saint-Martin; 1988). The different percepts resulting from this circuitous action in time and space, and stored in memory, will act one upon the other at the cortical levels. These will influence the nature of subsequent percepts which will, in turn, be integrated together with the first ones. The very first ocular fixations, which have not been re-equilibrated by subsequent centrations, are particularly deformative of the objective reality of the visual field. In general, as confirmed by K. Lewin (1936; 157): “The first reactions to stimuli are the nondifferentiated reactions of the global organism.” In other words, these first reactions result less from sensorial relations with objective stimuli than from reflexes linked to acquired schemas, to anterior structuring habits of the organism, assimilating the new to the already known, in order to secure conceptual or emotional equilibrium.

It is nevertheless true that the spontaneous conduct of most perceivers in front of a visual representation seems to be an active searching, rapidly scanning the entire surface in order to “recognize” some of its parts and obtain a global reaction to the ensemble. This hasty mixing of peripheric vision with a few foveal and macular liaisons leaves most of the visual field literally unperceived.

This spontaneous trajectory using mainly the scanning of peripheric vision, and recognizing groups of coloremes which lend themselves easily to iconic interpretations, is not to be confused with the sort of scanning advocated by A. Ehrenzweig (1967) which tends more to counteract the hegemony of closed forms within common perceptive habits. Through a deliberate use of intense foveal fixations and peripheric vagueness, Ehrenzweig tries to counteract the pressure toward the “good form” which Gestalt psychology has observed as being a common feature of perception.

What can be called the “normal” perceptive behavior of most subjects in front of a visual field of representation corroborates, in fact, the conclusions of Rorschach (1947) in the application of his noted test on perception. Global responses to a visual field are the result not of an extensive sensorial perception, but rather of a process of interpretation which extrapolates to the totality the meanings extracted from major or minor details, in liaison with mnemonic visual images.

Visual semiotics may also retain from Rorschach’s work the fundamental distinction between what can be revealed from observation—namely the description and classification of perceived events according to certain clear analytical categories—and what is added as an interpretation.

Given the inadequate and arbitrary character of most spontaneous behavior in visual perception, as revealed by Rorschach, visual semiotics will propose that any representation be observed in a systematic way and according to all parameters constituted by visual variables. It will thus aim at an awareness of the dynamics of each centration and will enlarge them sufficiently so as to take into account the effective interactions between all the parts.

In the course of this constitution of the visual text, through a perceptual investment in one coloreme after another and their innumerable interrelations and regroupings in supercoloremes, it becomes apparent that any description of a visual text as a whole refers to a specific moment in time, following a decision to put an end to the perceptual process. This temporal closure is deemed legitimate only if the observer has related himself in an adequate fashion to the multiple components of the visual field. The temporary synthesis he achieves should be an integration of the results of this attentive perceptual experience. But this pause is also illusory since the observer, already transformed by his perceptual syntheses, would be able, at the very next moment, to perceive the work anew through new equilibria and felt intensities. This would render him susceptible to experience the dynamic factors at play in the text very differently.

Semiotic analysis is thus conceived as an intense subjective experience calling for the conceptual and emotive resources of the individual and not as a distribution of verbal labels. It does not immobilize or tear apart the visual representation, but rather gives it flesh and blood. It is in no way reductive as, say the work of the physician is, of whom Lewin (1951; 157) speaks, who cuts up a radiography of a fractured leg into minute pieces and classes them according to their degree of gray tonality. Visual semiotics has to relate to an eminently performative language which is not only dynamic in its material constituents but which cannot be constituted as a text without constant changes in the internal experience of the observer.

1.2.Spatialization and Topology

The intuitions of topological geometry seem to be most fruitful for the definition of the spatializing process proper to visual language. But other conceptual means must be called upon, such as the notions of vectors, intensities, equilibria, and so on, in order to describe the energetic behaviors of visual variables and the perceptual process itself.

This point has been recognized as essential by the topological psychology of Kurt Lewin (1936), which has wed to topological concepts a hodo-logical geometry allowing some calculations, even if it remains nonmetric. The importance of topology has been demonstrated by its possibility of being used efficiently in mathematical applications to vectorial, set theories, and metric spaces (Berge, 1966). Even in the context of tridimensional Euclidean volumes, topology seems eminently successful by virtue of its fidelity to its own intuitions:

Topology cannot, in fact resolve equations. What it offers is a mathematical vocabulary—adjectives and nouns—which allows solutions to be discussed on a general plane, without them actually being specified, under the condition that the theoretic manipulation does not imply any laceration, cut, excavation or cavities in the volume under consideration. (Thurston, 1984; 108)

In the same way, we have seen the necessity to interpret the concrete visual space as a continuous and undulating mass constituted as a field of forces, the orientations or vectors of which are spread out in three dimensions: height, width, and depth.

These characteristics are present in each and all of its constituents, the coloremes, defined as masses possessing a general nonspecific configuration, as well as a certain thickness or interior volume. But as a dimension of a topological object, interior volume has characteristics quite different from its definition in an Euclidean context, where it is wholly dependent on the stable metricization of the “external volume.” We will try presently to clarify this point, which may represent one of the most fundamental intuitions of topology.

1.2.1.The topological volume

It is fundamental for visual semiotics to understand the distinction between the notions of internal and external volume, which was established in Jean Piaget’s early works, notably in The Child’s Conception of Geometry (1960). The internal volume is defined as the quantitative perception of the matter that constitutes the object. As for the external volume, it corresponds to the “place occupied in the environment” (470), measured from a point of view exterior to the object. Knowledge of the external volume necessitates a relation through metric measurements between the object and that which surrounds it. This point of view tends to make the object “an undeformable solid, in the sense of being uncompressible and undilatable” (449).

The internal volume has opposite characteristics. It is perceived as a nonmetric quantification, as the “filling up of envelopments” which constitute the object. If one considers any visible object as offering external surfaces, acting as frontiers or envelopments for the internal mass, the internal volume is perceived as “that which is enveloped by an ensemble of frontiers constituted by the visible surfaces of the exterior” (458). This type of volume can be transformed in its dimensions without losing its quantification.

It can change its external frontiers or contours without modifying the quantity of matter constituting its internal volume. For instance, a quantity of modeling clay always maintains its internal volume, corresponding to the given quantity of clay, even if its external volume adopts the form of a sphere, a cylinder or a surface plane.

There is, therefore, an invariance of the internal volume even if there is a modification of the external one, and this is the basis for the topological notion of a polymorphic, extensible mass, not definable by its external contour, as is the case in Euclidean geometry. In the genetic evolution of the child, as well as in the general culture, the opposition between these two notions of volume has generated a persistent conflict and many misunderstandings. In the domain of visual representations, this conflict was constantly active as seen, for instance, in the difficulty to understand the introduction by baroque art of the notion of a fluid matter and of “deformations” of external volumes (Deleuze, 1988).

From Piaget’s research (1952, 1960), it is suggested that topological intuitions are a priority in the experience of external reality and that the ulterior Euclidean notions will evolve from that basis without eliminating them. But the Euclidean intuition of space will attempt to substitute for a dynamic intuition of matter, one of “stable objects” that are substantial, autonomous, isolated each from all others, and unchangeable in time.

In a correlative manner, the topological notion of surface will diverge from that provided by Euclidean geometry, emphazing its function of envelopment: “A surface, in effect, from the point of view of elementary topological relations, is only a part of space enveloped by a line closed on itself and a volume is, from the same point of view, only a part of space enveloped by the frontiers of surfaces, equally closed on themselves” (Piaget, 1960; 490). The surface understood as a boundary will envelop a mass then considered as a full and dense volume, and not a hollow, empty volume or an unfilled expanse of space.

On its smaller scale, the coloreme, being a part of a space enveloped by surfaces/frontiers, always possesses a density of matter corresponding to its internal volume. This forbids that its perception, as a unit or in agglomeration, would be that of a kind of abstract plane endowed with only two dimensions. Indeed no material reality can have solely two dimensions. Only a logical fiction can conceive of a point without dimension or of a plane in two dimensions, as suggested by Euclidean definitions. But a logical construct cannot be equivalent to the physical matter serving as the semiotic instrumentality of visual language.

Both surfaces and volumes are here defined in terms of the topological properties of the relation of envelopment. It is a relation by which an object is inserted in a whole which becomes in part or in total its immediate environment in a global proximity. This notion opposes the intuitions of Euclidean geometry, as Piaget explains: “Topological space ignores, indeed, any distinction between container and content. It is, in contradistinction, proper to Euclidian space to imply it continuously” (1952; 31).

Topology considers, therefore, the envelopments and the enveloped, the container and the content, as sharing the same properties. The first term is not immobile and the second one mobile, in relation to one another. The two elements maintain the same possibilities of expansion or contraction, typical of all topological quantities.

This phenomenon explains another fundamental characteristic of the coloreme, namely that, while it is delimited by an ocular centration, it remains essentially linked with other regions that surround it. Separation between coloremes does not produce emptinesses or voids, similar to the silences arising when the sounds forming phonemes, morphemes, or propositions in the verbal language are no longer produced.

The coloremes are always bound to other visual regions in an unbroken perceptual contiguity which radically distinguishes them from all other sensorial stimuli, tactile, kinesthetic, etc. This characteristic distinguishes them from the verbal linguistic phenomenon where, as Prieto (1964; 16) observes “the speech acts present themselves in the form of blocks or cross-sections, of different lengths, separated by silences, that is, those moments when one does not speak.”

The elements of visual language are presented in a fundamental physical continuity which is at the same time the characteristic of the visible world itself: for the eye, everything touches everything, as Piaget’s young interlocutors had already observed. Within perception, there are no chromatic silences between coloremes in a visual text. The notion of a possible void between two groups of coloremes is not one revealed by perception, but is the introduction of an abstract, nonvisual point of view foreign to the visual field itself.

The set of relations that are established between different juxtaposed coloremes influences in an essential way the structure of each. In other words, the coloreme, which is always situated in a diversified context, can be known only through those interrelations it establishes with the elements of that context. These interrelations, arising from the multiplicity of centrations effected on regions of the visual field, constantly modify those established previously by virtue of the dynamisms of both the visual variables and the perceptual process.

While having its own distinct internal structure, the coloreme is, at the same time, a region which acts on and transforms its environment, only to be in turn transformed by it. This fundamental relational function affects its perceptual appearance in a manner unknown to the linguistic verbal unit, even if we agree on the definition given to the latter by Prieto (1964; 154), as an object which “does not precede, as such, the relations of which it is recognizably the terminal point.”

The consequence of this situation is that the visual variables which constitute a coloreme cannot be “described” as identical to some previously established repertoire, but are completely new events born through their interrelations with other perceived coloremes in the visual field. They always appear in a specific constellation which transforms their quality/ quantity through a process of dynamic interaction by which they diverge and converge, or are integrated one into another, producing complex models of continuous topological spaces.

In the past, certain researchers, conscious of the dependence of visual semiotics on perceptual processes, have alluded to the necessity to establish the first elements of visual language on topological notions. However, they have not taken any decisive action in this regard. We refer particularly to the interesting research undertaken by R. Lindekens (1971) into the semiotics of photography.

He recognized that “perception depends on the previous learning of the parts of an object, a process encompassing a series of visual fixations and depending on the visualization of an amorphous mass containing certain foci, to then succeed in seeing a distinct figure” (53). According to him, a verbal or iconic identification of this “amorphous mass” is not at all required in order for perception to be effectuated. He quotes numerous experiences which “attempts to prove that, effectively, the unit resting on separate physiological basis, and independent psychologically, can be perfectly perceived without identification” (51).

It seems that the ideological refusal to admit as perceived objects the zones of a visual field which remain unnamed influenced nevertheless Lindekens. He did not develop the results of his demonstration of the fact that a slight modification of distinctive traits (or in our terminology, visual variables) for instance, in typographical variations of the alphabet, involves an important modification in the “meaning” ascribed to each letter by a group of perceivers. This meaning was expressed in terms of diversified aesthetic and emotive connotations, such as “nice,” “elegant,” “delicate,” “aristocratic,” “strong,” and so on. But Lindekens also defined a correlation with another level of meaning, qualified as “infra-verbal” or more “intralived” than either thought or verbalized experience (198). The pursuit of the ever evasive function of meaning has prevented this researcher from exploring further the notion of contrast in the functional mechanisms of visual signifiers.

This last step seems to us essential in the study of visual language because the visual variables constitute the material sine qua non of reality as visible. The analysis of interconnected visual variables within the coloreme affords access to the first semiotic structures of this mode of representation. As such, it permits the description of dynamic units which are the very “building-blocks” of various visual texts. In effect, the energetic intensities and interactions of coloremes sampled in the works of Cezanne, Matisse or Pollock, present entirely different dynamisms which already give information on each artist’s basic manipulation of visual language. This diversified structure of coloremes betrays organizing mechanisms at play in smaller areas, susceptible to repetition or contradiction on larger syntactic levels.

Still more important, given that ocular fixation allows for continued recourse to foveal/macular liaisons in the visual field, experimentation with coloremes will play a continuous role in the perceptual relations to a visual text, interfering at the end with reactions provoked by larger agglomerates of visual regions.

As the observation and description of coloremes require a very particular attention, we feel that it is necessary to provide additional information on the very nature of the visual variables that constitute them.