Chapter Three

Syntax of Visual Language

3.1.Syntactic Rules

One of the prime objectives of a grammar is to describe the greatest number of possible enunciations of a language with a limited number of syntactic rules established at a generative level of its structure.

The syntactic rules of the visual language are constituted by the set of operations and functions through which perceptual mechanisms establish interrelations among the basic elements in diverse visual fields. Their application results in the construction of specific spatial totalities.

The visual field is defined as a force-field where given energies produce particular effects leading to different types of spaces. These energies are provided by the aggregates of reflected light first differentiated in the perceptual units of the coloremes. These are the locales of incessant transformations activated by the subjective processes of perception as well as by interactions generated by the coloremes themselves in a visual field.

In other words, the syntactic rules of visual language are regulators of energy and visual transformations, and not merely a description of relations between some sort of substantial and stable elements. The coloremes themselves, as mentioned above, are not units that predate the emergence of visual variables, but they are produced by and transformed coextensively with them.

Coloremes could be compared to corpuscles only if these were defined, as Bachelard (1951) did for the basic units of physical matter, as particles having neither assignable form nor dimension, that are flexible and penetrable and that can exist, as nuclei of energy, even under a very little quality of individualization.

Coloremes are defined as nuclei of energy since they are made up of matter. As Bachelard has pointed out, Einstein has “attributed an energy level to the resting mass of an inert body” (1951; 110). Coloremes are even more so since they are, as percepts, made up of vibrations of reflected light.

To define the syntax of the visual language as the ensemble of regular modes of production of visual movements that constitute dynamic invariances is to imply a conception of movement different from that of classical mechanics which only considers the sort of movement produced by the collision between two bodies. Bachelard has reminded us that even Nietzsche still carried over this limited intuition of the dynamics of movement in his philosophy:

In truth, Nietzsche only lived for the will to strike. The shock and the stimulus, the attraction and repulsion, constituted for him dynamic, fundamental images. He did not understand those trembling movements, those hesitant movements, this whole phenomenology of vibrations, oscillations, rhythms; a phenomenology which gives a unity to complexity, which institutes the notion of frequency at the level of a primitive concept, of a simple concept . . . Nietzsche did not either have access to this enormous problematic of a movement which, born in a center, sets in motion the whole of a milieu, dynamizes the whole of a space, calls on multiple hypotheses to explain the reactivity of space. (1951; 52)

While geometric concepts are necessary to describe any visual spaces, the geometry which is needed should incorporate dynamism and movement. Again, as Bachelard recommended to physicists, semioticians should reject “the ultra-realist intuition of the movement as systematically attached to a defined mobile at a physical point. . . Instead of a determined mobile it is a dynamic environment that one must study” (1951; 184).

The energetic character of basic units and of syntactic laws between elements is also increased by the mobility inherent in the processes of perception which alone can constitute the visual text as such. Indeed one can analyze a visual statement only after having actively constructed it according to syntactic structures proper to this type of language.

If N. Chomsky (1980) could contend that competence in verbal grammar belongs to the order of unconscious knowledge, the intuitive use of visual syntax by producers belongs to the same category, the more so since no overt, conscious theory about it has ever been evolved.

Analysis of visual language has been so much more difficult to achieve since no external criteria could establish that a proposition about its structure was adequate, or that one given visual statement corresponded to a valid enunciation more than another. As we have mentioned earlier, the visual speaker, that is, an artist or a producer of a visual representation, does not enjoy a theoretical authority as to what belongs or does not belong to this language. On a great many occasions, visual representations produced by other groups (or of one school compared to another, like the Academicists vis-à-vis the impressionists) have been denied the property of belonging to “visual language.” But these deviant forms of representation were duly recognized as legitimate by cultural collectivities at a later date.

Paradoxically, the semioticians who study visual language are not generally “users” of this language and thus would not possess a linguistic competence acquired from practice, analogous to that of the grammarian of verbal language who is himself a user.

Visual semiotics will therefore be able to invoke for its validation mainly internal criteria of deductibility and coherence, as well as the fact that its propositions on the grammatical structure of visual language allow for the actual functioning of visual language in its diversified forms. It should not exclude any of them, past or present, without the consequence of offering an inadequate description of the foundations of that language. In addition, it must be able to describe the modes of production of future types of visual statements, which will result from transformational rules applied to the same basic units.

The syntactic rules that allow the formation of particular spatial fields will necessarily contribute to later semantic interpretations of these forms of representation, since they determine the linguistic functions and preside over the interrelation of elements in the global text. But their semantic function may be even more important, as suggested by R. Jakobson: “All languages are founded on a system of grammatical categories and the meanings of these categories are characterized by being binding for the speakers” (1980; 119). It follows that “grammatical meanings play a fundamental role in the linguistic thought and verbal communication” (110). We suggest that the case is similar for visual language, making the elucidation of the nature of its syntactic structures more urgent.

By definition, visual grammar is required to describe the operations by which basic elements are determined and interlinked through basic rules so as to form more complex ensembles. These rules are interrelated in an orderly fashion and in such a way that the first level operations are necessarily accomplished before the second level ones, and constitute their indispensable foundation. The first level rules provide a basis for the higher level operations which regroup a greater number of elements in more and more complex perceptual schemas.

The grouping of basic elements in visual language is carried out according to certain laws and with the aid of perceptive operations which in turn are influenced by the structure and organization of each particular visual field. Thus the syntactic interrelations between the first correlatives of perception and subsequent percepts constitute agglomerates or sub-assemblies generated by certain operators. Regroupings of regions or superregions result from other operations, up to the elaboration of systems of regions, of subregions and of superregions, that will be in turn organized in “systems of systems” by higher level dialectical operators.

The syntactic rules that regulate the joining of coloremes and transform their internal/external functions call into play a variety of operators and may be regrouped on three principal levels as follows:

A. The rules of regrouping of coloremes among themselves that depend on:

(a) topological relations,

(b) gestaltian relations,

(c) the laws of interaction of colors.

B. The rules generated by the insertions of coloremes within the energetic infrastructure proper to each visual medium; that is, the Basic Plane for the pictorial, the Virtual Cube for the sculptural and its environmental extension for the architectural.

C. The modal rules that preside over effects of distance and which are inscribed in various codes or systems of perspectives. We will comment presently on the first group of syntactic rules (A), while full chapters will be devoted to the second and third groups.

3.1.1.Topological relations

The first group of syntactic rules that govern the relations and functions of the coloremes is based on the topological relationships that have been recognized as primordial in the “construction of the real,” In the 1940s, the genetic epistemology of Piaget established that the first geometrical model of space used by human beings is not Euclidian geometry but topology. This spatial model of the organization of perceptual experience remains throughout human life the basic means by which one constructs his notions of reality. As we have devoted an earlier work to the study of topological functions in pictorial representation (Saint-Martin, 1980), we will summarize the essential features of the more important of these notions.

Acting as a qualitative intuition in advanced mathematics, topology relates in common experience to a “pure perception” of spatial relations as forming a homogeneous, continuous, syncretic totality. This continuum of space is based on the sensation of the proximity of stimuli in the visual field and leads to the intuition of a space without holes, ruptures, or voids. This spatial continuum is not a given, but is the result of an organization of the perceptual field into a number of fundamental relations.

Topological relations are established in the visual field on the basis of the tensions animating the visual variables themselves, activated and expanded by the perceptive activity of the observer. The interrelations established between diverse points in the visual field, called ‘paths’, can be of two types: a) locomotions, resulting from a de-centration or changes in the points of viewing, or in the position taken by the perceiver in relation to the visual field; b) communications, which are occurring as a result of the interrelations arising from the various dynamisms proper to the material energies constituting this field. Both are fundamentally subjected to and describable by notions belonging to topology.

K. Lewin (1935; 80) has defined topology as “a non-quantitative discipline which deals with the possible modes of connexions between various spaces and their elements.” Any other modes of structuring or modelling spaces have to be built on such topological stratas, to which are added other perceptual mechanisms or conceptual and logical organizations. Properly speaking, many of the latter are not spatial as they lack the quality of continuity.

The relationship of neighboring is the most important topological notion by which the function of continuity is constructed in any spatial field, whether physical or perceptual. Its importance to physical sciences was underlined by Bachelard (1951; 6) when he stated that any force in the continuity of a field “presents itself as determined by the condition of neighboring. This term, vague in the everyday language, acquires all of the desirable conciseness in mathematical expressions” (1951; 6).

This intuitive relation is produced in the perceptual field through a subjective mechanism operating on external data. It establishes a proximity or a most immediate connection among elements that could otherwise be described as existing independently from one another. It may stem from a homogeneous contiguity, from an attraction, a fusion, an assimilation, or a similarity sufficiently dynamic to force a binding between elements. The strong continuity it produces is basic to the constitution of topological masses as well as in the production of a space conceived as a plenum.

The antagonistic relation of separation is mostly brought into being by an external causality or the estimation of a difference, and consequently of a disjunction, between groups of coloremes in terms of some visual variables. By definition, in this context, the notions of contiguity or juxtaposition between diverse aggregates of coloremes imply that they have been perceived as separate.

The fundamental relationships of neighboring or separation are not assigned to visual regions because of the consideration of a single characteristic among visual variables, since none as such can determine the level of energetic intensity of a region. They are the result of a perceptual estimate and a synthesis integrating all visual variables in a region, along with their interaction with the ambient field. Both notions are the source of the emergence of the topological notion of limits, boundaries or frontiers, where the nature of the connections or passages between separate regions can be recognized.

The order of succession is the type of relationship set up among coloremes when some of the visual variables are repeated in the three dimensions. It implies a similarity between one or more visual variables forming the coloremes or in their dynamic interrelations within the field. It occurs under a variety of organizations, such as:

A regular alternance in depth produces a type of “woven” space, while a “folded” space results from the superimposition of interrelated layers of groups of coloremes in the third dimension.

By definition, groups of coloremes made part of a succession differ among themselves, given their different points of implantation in the plane. They will be further modified by their structural function in a succession.

Properly speaking, a serialization is an order of succession imposed upon a limited number of separate elements through a similarity between very few of the visual variables. The term also applies to a repetition of these elements, but in changed positions in the two dimensional plane and consequently in the third dimension, offering endless variations of internal structures:

And so on.

The orders of succession, established by perception between distant regions, constitute the basic function of rhythm, specific to any individual space of representation.

The notion of envelopment corresponds to the interrelation among several coloremes when some totally surround others. A semi-envelopment partially regroups some coloremes around others. In a topological context, this relationship occurs only within the notion of continuity, creating the strongest bonds of connection and dependence between regions and adding to them a new range of dynamisms, such as:

Topological envelopment is realized only when the differences among visual variables pertaining to the enveloping and the enveloped are neutralized by the establishment of a neighborhood relation, producing continuity in depth. It is also felt as an asymmetrical group given the contrasts between the vectorial energies defining the container and the content.

Encasing or emboxing is the phenomenon by which some regions are perceived as if they were “inserted” between other regions, similar or dissimilar, such as the position of B in ABA or ABC. It can be a global or a partial characteristic as in the case of envelopment, from which it differs by virtue of the stronger separations caused by contrasts between their respective constitutents, as well as by the firmness of their boundaries. As the region encased is estimated as a part of a particular system, this relation isolates both elements inside the global field.

The topological relations are to be understood in their potentiality for the constitution of continuity. This seminal notion is a function that results from the state of the basic elements being perceived as sufficiently similar, close, dense, and compact, so that no disjunction can separate them to the point where a breach, or a void of energy disrupts them.

Spatial continuity must not be confused with temporal continuity, or the duration continuity, which may or may not accompany the consciousness of any experience, perception included. The feeling of duration put forward by Bergson cannot establish a spatial continuity, as Bachelard already observed: “For Bergson, continuity is an immediate given of consciousness. It is this intimate continuity that we jilt in discontinuous grasps of external experience” (1951; 55).

According to Bachelard, should a new ontology today be based not on common sense but rather on more objective information about the nature of reality, the idea of continuity would become very problematic, since “the most compact matter is taken by common sense as an assembly of very dispersed corpuscles” (1951; 107), that is, a discontinuity.

For semiotics of the visual language, discontinuity is a primal occurrence arising from perception itself, given the multiplicity of heterogeneous stimuli encountered; it also arises from the logical levels where semiotics is assumed to operate as a science explaining effects by causes: “Any causal phenomenology is necessarily discontinuous because one speaks of an effect that follows a cause only for an effect that differs from the cause” (Bachelard, 1951; 206).

But this logical type of discontinuity does not affect the continuous/ discontinuous spatial relationship which perceptive activity establishes at the level of topological relationship. Continuity appears here as a constructed intuition reorganizing the discontinuous aspects of material events. In the same way, a spatial continuum is not a given datum but a construction of perception itself.

We will finally consider the notion of vectoriality, which is differently interpreted in various geometrical contexts. Many systems of geometry afford a progressive transition between a non-Euclidian geometry, such as topology, and the Euclidian model, and they differ greatly among themselves.

Projective geometry reorganizes the spatial and topological field by introducing the vectorial point of view of a straight line extended to infinity. It thus offers a system of generation of forms from peripheral rather than focal forces. This non-Euclidean geometry has been a major factor in the elaboration of Suprematist and Constructivist forms of representation, as well as in Mondrian’s Neoplasticism, in the first decades of this century. The terms “a-logical” or “irrational” were used to describe this intuition of a space as infinite. Not only could the objects inserted in that space be interpreted as infinitely close or distant, but they would be described from various simultaneous points of views (El Lissitzky, 1968).

The notion of vectoriality is fundamental in projective geometry, where it is defined as a linear oriented progression. In topology, it corresponds more to a planar or mass thrust animating internal volumes. It characterizes in particular the centripetal-centrifugal energy animating the central and peripheral layers of a topological region, as well as its capacity to react to the vectorial forces of the ambient field.

The vectorial elements of projective geometry lead to the construction of Euclidean space. This spatial model restricts its constitutive coordinates to the perpendicular meetings of straight lines in three dimensions. The first two coordinates, meeting at 90 degrees, form the so-called bidimensional plane, while the third axis, also perpendicular to the vertical/ horizontal lines, represents the dimension of depth. This constant mathematical grid determines the position of any point in space, as well as a nonvarying distance between two points. Useful at the macroscopic level of experience, this Euclidean grid cannot be used to map more proximate perceptual spaces (tactile, thermal, and so on). Nor can it be used to map or represent internal, psychological or emotional relations which are unquantifiable dynamic phenomena.

3.1.2.Gestaltian relationships

At the level of aggregates of coloremes, the visual variables are governed by the Gestaltian “laws of perception.” While the findings on visual perception achieved by the experimental psychologists regrouped under the name of the Gestalt school are recognized as constituting laws, analogous to Newton’s laws of motion in physics (Rock, 1975; 281), they have not been systematically applied as structural means to describe the organization of the visual field.

We have already tried to establish that, both in a restricted and in a larger part of the visual field, movements among visual variables are achieved in conformity with the laws of visual movements proposed by the Gestalt theory (Saint-Martin; 1990).

From the summary offered by Kohler (1940; 134-135), we recall briefly the more important of these perceptual structural processes:

(a) The visual field is first organized by the constitution of the figure-ground relationship. When a denser agglomerate of visual variables is present, it is endowed by perception with the characteristics of a figure and the surrounding areas acquire those of a background.

As it establishes among others a superimposition of the figure in front of the ground, this process is instrumental in creating fictional space, that is, the percept of a distance in depth between the two regions.

(b) The visual variables are above all regrouped according to the factor of their proximity in the three dimensions.

(c) Visual variables are perceptually regrouped so as to present an approximation of a more regular, simpler, relatively closed and symmetrical totality, called a gestalt. The visual movements involved in this process are said to respond to the pressure of the good form, that is a perceptual tendency to actually “misinterpret” more or less the actual data so that they would correspond more to geometric patterns possessing a strong internal structure.

We assimilate to this process the tendency to associate familiar remembered schemas to aggregates of visual variables, so as to achieve a rapid recognition of “objects” in the visual field. The pressure of the “already known” in the interpretation of visual aggregates is the basis of the iconic recognitive function.

(d) Similarly regions, lines, or vectors, that can be joined in a virtual good curve will do so, that is, a continuous and simple arabesque.

The same happens with elements that lend themselves to the making of a good angle, that is, a right, acute or obtuse angle.

(e) Strong regroupings are produced under the influence of the factor of similarity between some or most of the visual variables, the most effective being the similarity in color. It may be of importance to point out that in the two dimensional expanse, there cannot be total similarity between two elements given their differences of position in the plane. Besides bringing regions closer together, the factor of similarity tends to establish a homogeneity of the field.

(f) As most Gestaltian researchers have acknowledged the influence of acquired habits, of knowledge, and of experience on perceptual processes (Gabar, 1968; 55), one may rightly add as common factors in perception the affective and conceptual structures of the individuals, their needs and abilities, which affect the nature and extent of their involvement in perceptual activities. As shown by Rorschach’s conclusions to his renowned test on perception (1947), emotional factors play a major role in an individual’s aptitude to apply to the entire visual text the gestaltian factors of groupings (Saint-Martin, 1968). While these appear basic and universal, they are not necessarily made use of by every viewer confronted with a visual representation.

As much as a verbal text would not be adequately apprehended if some grammatical categories, be they pronouns, genders, or tense of verbs, were not recognized, visual semiotics proposes that a description of any visual representation is not adequate without a recognition of how the various syntactic categories are applied.

Both topological and gestaltian operators are at work in any viewing of a visual representation. While gestaltian processes are applied when separations are recognized, topological intuitions still act upon the internal constituents of these regions and can contribute to minimalize the contrasts between some visual variables in view of attaining some measure of neighboring.

The regroupings produced by both topological and gestaltian operators are perceived as more or less stable totalities, dependent on larger regroupings. Any grouping can thus become a subgestalt for a potentially larger gestalt. Under the effect of prolonged and multiplied centrations, these very regroupings may dissolve and their constituents appear as isolated or regrouped differently with other regions of the field.

3.1.3.The laws of interaction of color

The principal laws of interaction of color are those of equalization, simultaneous and successive chromatic and tonal contrasts, the optical mixture, and other regular effects not yet adequately categorized. While submitted to gestaltian factors of junction/disjunction, they produce phenomena that cannot be entirely explained by them.

These interactions between colors can produce intimate junctions between contrasting elements, dissolve boundaries, produce new virtual forms, modify the dimensions, positions and vectorialities of regions, and so on.

Given the complexity of behavior of the various chromatic events and to avoid repetition, we refer the reader to our previous chapter dealing with the specific mechanisms of the interaction of colors.

3.2.Specificity of Visual Linguistic Fields

The various types of visual language (painting/photography, sculpture, architecture, etc.) do not differ by their use of visual variables that are common to all, but rather they are differentiated in their syntactic and semantic structure by analytical hypotheses that have given them birth. These provide permanent and different infrastructures that we call the Basic Plane, in the case of painting or photography; the Virtual Cube, in the case of sculpture; and the environmental Cube, in architecture.

The concept of the Basic Plane was developed by the artist Wassily Kandinsky in his efforts to produce a grammar of pictorial creation. This expression can be used in a general way to refer to those grammatical structures that establish the specificity of visual linguistic fields. In this respect, the proposition of Kandinsky is as fundamental to the development of visual semiotics as was Troubetskoy’s phonology to the development of verbal linguistics. It is indeed only by the structure of the Basic Plane that the visual variables used in various forms of visual language cease being pure material data in order to acquire linguistic functions.

It is important to distinguish sharply between the Basic Plane of painting, of sculpture, of architecture, and the pictorial plane or the sculptural plane, at the level of effective creation of an artwork in these sectors.

The Basic Plane is an infrastructure antecedent to any production of visual discourse. It presents a potential spatial energy allowing visual variables, which are physical quantities common to any visual reality to be inserted in a context where they can be given linguistic functions.

By contrast, the pictorial plane or the sculptural plane designates a level of energy resulting from the use by the artist of a group of visual variables that possess the potentiality of self-organization according to the topological and gestaltian syntactic rules. But they can be defined spatially only through their interactions with the structural energies of the Basic Plane.

The pictorial Basic Plane or the sculptural Virtual Cube correspond to the dynamic functions linked to the support/format that carries the pictural or sculptural message. It necessarily antedates production of any pictural plane and conditions the movement of plastic energies produced by the artist. The infrastructure of the Basic Plane, independent of all strategies of artistic production, organizes a level of constant energetic constraints that impose specific grammatical linkages/transformations to coloremes that are incorporated within.

Thus, in addition to specific syntactic operations that result from the perceptive recognition of visual variables in coloremes and the functions that are acquired by regroupings of coloremes, these linkages are determined and transformed by their insertion into the infrastructure of the Basic Plane. These operations are regrouped in an ensemble of specific energetic trajectories that we will describe with respect to the pictorial and sculptural Basic Plane. Grammar of visual language will be completed by the study of particular syntactic modes corresponding to the various systems of perspectives.