PRINT March 1978

The Body Language of Pictures

TRADITIONAL COMPOSITIONS ARE ANALYZED, in terms of form, as part/whole relationships. Untraditional compositions tend to be described as grids, wholes, modular systems or “collage,”1 and they are discussed in terms of the artists’ intentions. Maybe the implication is that esthetic theory now does the job of formal analysis, or that contemporary art transcends mere visibility. Both artists and critics treat composition like a throwaway when the disposition of component units is anonymous and unstressed—for example, in Duane Michals’ text-and-image pieces. Still, visual organization can be ignored only as long as it works reasonably well. Axial symmetry, for one thing, is all over the place. It is the dominant device of what one might call underground composition. Like the traditional organization of book pages, it reinforces planarity, even though small, evenly articulated surfaces require little in the way of control. Such “compositions” have become so familiar that they escape notice, so commonplace that they look natural. Yet they reflect the contemporary preference for forms of visual organization that seem nonarbitrary and consistent with the flatness of the surface. A rather limited number of structures serves a wide variety of artistic rationales.

In the remarks that follow, axial symmetry is discussed as the axis of symmetry established by the support. Images may or may not be symmetrical, but formats, including shaped canvases, usually are. Even when the vertical axis of the format is unmarked, its implied presence guides the viewer in centering attention on the work. The image’s own axis of symmetry (if it has one) will strengthen or play against the axis of the support. My attention was drawn to this element by Dominique Fourcade’s provocative essay in the catalogue to the Matisse cutout exhibition. Under the title “Something Else,” Fourcade claims that with a few exceptions (L’Escargot and Souvenir d’Océanie) Matisse’s large, ambitious cutouts show a departure from chromatic organization and its replacement by a principle of symmetry, “. . . a symmetry which was completed and even worsened by still another need he [Matisse] seems to have felt at the time—that of uniformly filling a surface.”

Now Fourcade is a passionate and intelligent critic, but Europeans simply have not seen as many grid structures in the last ten years as Americans have. Actually, Matisse’s late works are predominantly grids rather than mere symmetries. The formal properties of the late cutouts as Fourcade describes them—the symmetry, the uniformly filled surface, the discreteness of the forms (their separation from the surrounding ground)—are normal characteristics of simple grid structures. Fourcade remarks indignantly that the taut white field underlying the cutouts (which he calls “the unpainted”) is no space at all, but an antithesis to the painted forms set upon it. But Americans are familiar with the flat physicality of that ground. They have accepted its refusal to envelop the shapes it separates since it first appeared in Stella’s black paintings.

Stella’s black paintings offered a remarkably full exploration of more conventional symmetries, where multiple axes of orientation are complications of the pictorial gestalt rather than of perceived space. Usually discussed as deductive structures, the variety of “deductions” in the black paintings depends on the fact that each figure presents simultaneous and alternative readings, depending on whether structural axes are perceived as diagonals or as linear coordinates. Arbeit Macht Frei, for example, can be seen as “deduced” from a centralized cross (incomplete rectangles opening each of the corners and stressing a centrifugal movement); or centripetally, as the conjunction of paired, pointed figures, one horizontally oriented, one vertically; or as rectangular part-to-part mirroring along either a horizontal or a vertical axis of symmetry. The succession of perceivable symmetries offers viewers a way of engaging themselves with the work; it gives them an active role in reorganizing perception.

The “space” which underlies linear grids in two-plane lattice structures normally presents itself to us as a physical surface and not a constructed one. What distinguishes lattices from other versions of grids is that lattices usually read as dualistic figure-ground systems, though the “unpainted” areas can play either role. Sol LeWitt’s wall drawings, Agnes Martin’s pencil drawings on paper or canvas and Valerie Jaudon’s denser interlaces are all examples of the simple lattice, the same disjunctive system that Fourcade deplores in Matisse’s Decoration with Masks.

The separateness is most subtle in Martin’s work. Her grids lie on the surface as lightly as spider webs, drawing us into an appreciation of the physical texture of her support. (It is worth noticing how frequently Martin’s work is reproduced in details, the relationship between the grid and the surface being vital to her effects.) This most sensuous artist was excruciatingly careful about adjusting the limited contrast between the linear veil and the surface. That contrast is what gives each work its specific color. LeWitt was drawn to use silverpoint and colored pencils in order to achieve a comparable disjunction in the coarser visual field of his walls.

The spatial disjunction between a linear grid and the surface it measures is spelled out when we are allowed to see that their axes of symmetry do not coincide. Initially, Martin’s grids were given a unit shape of their own. In a smaller size, the grid departed ever so slightly from the squareness of the format. Just as some grid lines went faintly askew, the grid itself was seldom exactly centered in the area it .occupied. Often the grid’s horizontal axis was displaced upward, giving a faint buoyancy to the image. When there is no border, the slippage of grid plane against ground plane becomes invisible and the opposition of a horizontally or vertically oriented oblong grid against Martin’s square format is carried by the shape of the module itself. The nonidentifiability of the grid plane and the surface it rests upon was deliberately chosen for the faint mobility it yields. Martin has called attention to this factor, saying: “My formats are square but the grids are never absolutely square, they are rectangles a little cut off the square, making a sort of contradiction, a dissonance. The rectangle lightens the weight of the square, destroys its power.” Formally, the grid and its support are given slightly divergent axes of symmetry.

The axis of symmetry has come to signal a new interpretation of the viewer/object relation, offering the chief resource for controlling that relationship. Instead of locking the viewer into a specific (imagined) viewing position in relation to the created space of the image, as Renaissance perspective does, modern compositions often encourage the viewer’s mobility in a scanning pattern rather than a determinate focus of vision. The format’s axis of symmetry is the anchor that moors us to the site of artistic action. Unstressed, it holds us ever so lightly, as in Noland’s targets; insisted upon, as in the chevrons, it can become the major key to pictorial structure. We may be led to acknowledge the persistence of the created image despite the contingencies of our physical position in relation to it. Instead of being drawn into the fantasy of imagined space, we face an immediate, independent presence whose intentions require investigation. By fixing our initial position, axes of symmetry announce our newly contingent orientation to what is seen.

The viewer’s spatial mobility is linked to the 20th century’s escape from gravity. Like it or not, we are all less earthbound than we used to be. Nature seems feeble and vulnerable, while we see ourselves as the victims of powerful social and historical circumstances. Now any statement about reality is perceived as reflecting the intellectual, moral or political position of the speaker. A multifaceted, equivocal world can be understood only by those prepared to grasp different “points of view.” In art, mechanism and contrivance are acceptable for new reasons. Systems and photographic images are not supplementary to nature, but the only way an artist can signal that he’s “going public,” avoiding merely personal expression. The importance of pictorial flatness is not that it’s a special novelty of modernism which rejects illusion. The flat, bounded surface accepts the need for narrow sharp-focus attention, foregoing unattainable completeness. Far from reflecting a simplistic preference for the literal over the ideal, flatness is the stigma of art’s special language. The work’s physical surface is a deliberately neutral ground that does not assign a viewpoint so much as make the viewer aware of the fact that he’s using one.

The new “rule” about flatness is what accounts for the triviality of depicted shape (as if drawing were nothing but rendered shape!). “Composition” or pictorial organization is more important than ever,and its new conventions are as significant as the old ones used to be. A small, rarely seen Agnes Martin called Mountain looks anomalous, probably because of its implications of perspective space. Dibbets’ “Dutch Mountains” series, like his perspective corrections, is more readily acceptable because it supports a modern focus on the contribution of the mobile eye/I.

To claim that modernist composition involves dealing with the axis of the format sounds vague, but it has the virtue of including both reinforcing and dissolving strategies. An artist can “lose” the vertical axis by choosing a very wide format, or by adopting a shaped canvas whose atectonic silhouette leaves us suspended among various axial possibilities (Stella’s procedure in the irregular polygons and some works in the protractor series). Alternatively, an artist can suppress the format’s vertical axis by introducing multiple symmetries (as in Stella’s black paintings). But regular, literally handled shapes usually present us with an implicit dominant axis which guarantees a horizontal or vertical orientation for the work as a whole, and reinforced axes do predominate.

Achieving spatial orientation within a world of objects implies or projects a shared gravitational field. We do not take the experience of gravity as a private experience, but assume that it affects all the bodies and objects around us. This faith underlies our ability to interpret the perceived forces of the visual field as acting upon us, the observers. When a visual field is presented on a horizontal plane (floors or ceilings), that innocent tendency results in a more or less severe sense of dislocation or disembodiment. Artistically elaborated vertical surfaces usually reinforce the stable vertical axis and mobilize horizontal or diagonal axes—this is characteristic of pictures and wall treatments. Artistic elaborations of horizontal planes, like carpets and ceilings, on the other hand, tend to invite displacements of direction and imagined depth of field. The classic Persian vase carpet can in fact be read as a lattice with five shallow planes. Carl Andre’s floor pieces ideally generate new valences, boundaries and scale relationships in the rooms where they appear. Neither his procedures nor his results are as interesting as these revisions of placement that his work engenders.

Symmetry and the axis that locates it cannot be investigated adequately as long as we think of them merely as properties of objects and images. As a prime element in the viewer’s relationship to the work, the axis of symmetry inevitably affects our interpretation of the visual field. Whatever “space” the artist proposes, some definition of the viewer’s position in relation to the work cannot be denied. As an experience rather than a characteristic of the object’s gestalt, vertical axial symmetry aligns the viewer’s axis with that of the object, its “face.” The sharpened awareness of this parallelism intensifies intellectual or psychological response, for that face-to-face parallelism is the physical paradigm for interpersonal intimacy, the posture of dialogue. On the other hand, seated and leaning over a picture in a book, we easily forget physical constraints. Thus manuscript illustrators, East and West alike, have rarely worried much about spatial consistency.

Early Conceptual art, like Dibbets’ perspective pieces and Bochner’s measurement series, often focused on the difference between subjective and objective knowledge of real space. Such works were generally interpreted as investigations of the peculiarities of sign systems or as documentation, but they depended equally on nonsymbolic visual experience, which set for the viewer the mild dilemma of deciding where he “stood”—an issue that is also important with Barry Le Va. The continuity between Bochner’s conceptual pieces, using verbal or mathematical sign systems, and his permutations of geometric shape—the polygons—is marked by a persistent manipulation of the viewer’s (imagined) position in relation to the work. Multiple readings are demanded by forcing the viewer to face the work from different positions in order to decipher it. Two, sometimes even four or more (for the polygons) different orientations might have to be taken in order to “see” the construction of the whole—a perfect paradigm for the elementary mental acrobatics that the works embody. The imaginary repositioning of the spectator is the major stimulus to participation that such work of Bochner’s affords. It is given an extra fillip in the Axiom of Indifference, where the viewer’s imaginary placement in relation to the work is physically blocked by a wall, forcing us to read the piece from a position we are forbidden to assume.

Pictures with a single, strongly marked vertical axis can also suggest a centripetal effect, fading out toward the edges and weakening the impact of the painting as a self-contained physical object. Robert Mangold said, “I liked the idea of a section [of a circle] implying more and yet being a complete thing.” And Mangold’s juxtapositions of sectional arcs were intended to minimize the importance of the boundary. (Shaped canvases are still too unconventional and too reminiscent of sculpture to support that suppression easily; rectangular formats do it better). The same sense of a complete yet potentially expandable painting showed up in Johns’ 1975 exhibition, where Scent, a work with a dominant vertical axis, strongly suggested a regular vertical mirroring which did not, in fact, exist (the repetitions actually involved the reflection and transposition of smaller, horizontal units). The most successful example of the axis serving to suppress boundaries is still Jawlensky’s last series of contemplative heads, where the edges of the face and the edges of the canvas exactly coincide. There a rectangular format, a gridlike linear pattern, and an allover blackness yield an eerie, St.-Veronica’s-veil-type image, a kind of apparition whose psychological force overwhelms the object-hood of a painted canvas.

So far I have discussed the role of compositional axes in paintings whose unitary character was strongly marked, either by keeping the picture small enough so that its boundaries are all always visible within the normal range of viewing, or by emphasizing the central axis of the pictorial field. Axes of symmetry have completely different functions when the artist’s surface is so large that it cannot be seen as a whole or is subject to perspective effects. The late 17th and early 18th centuries delighted in such vacuum-cleaner effects. Artists specialized in whooshing the viewer down an imaginary tunnel into nothingness. The illusion was repeated ad nauseam in infinite vistas of landscape architecture, especially in stage design. In medieval Islam as well as contemporary America, ideology and taste enjoin the contrary effect. We prefer to deny diagonal, space-piercing vectors in favor of a stabilization of intervals that lets us feel we can handle an extended, potentially overwhelming situation. This allows us to preserve an allover grandness of scale while dealing with one thing at a time.

Valerie Jaudon’s new ceiling in Philadelphia consists of only three repetitions of a bilaterally symmetrical motif. The great size of the design units has the merit of lowering the ceiling visually and of undercutting the inhuman scale of the lobby space beneath without adding an unmanageable amount of detail. The rush of the dominant longitudinal axis is slowed down by a sequence of four transversely oriented motifs. Because any viewing position makes us perceive the transverse elements as diminishing in width, the short axes look more varied than the single unchanging, dominant longitudinal one, an effect that is reinforced by a series of crescendos of light from the widely spaced lighting fixtures. Oddly similar effects occur in the abrash of Eastern carpets, those streaky irregular variations in color which are traditionally introduced into the weft, completely independent of the pattern.

Extremely wide pictures are usually seen in rooms that frustrate any attempt to see the image as a whole. That frustration can be insured by eliminating axial clues (as in many of Noland’s stripe paintings) or by introducing one or more patently false axes (Newman). Swamped by color fields of indeterminate extension, we are exposed to an ungraspable environment and open to suggestions of infinity, of ungovernable space or uncontrolled velocity.

More “kindly” artists, or artists willing to stay on the wall, provide us with a sequence of symmetries to pace an orderly progress through a large work. Ellsworth Kelly’s panels of color are linked to each other as a series of symmetrical shapes. His units develop a sequential, narrative effect, each contributing its voice to an ultimately apprehended chord. This is fundamentally the traditional narrative use of long horizontal formats. In the Bayeux tapestry and in many Japanese narrative scrolls (like the 13th-century Chōjū Giga, the one with the rabbits), symmetrical trees or buildings both link and separate juxtaposed scenes. Such symmetries keep us from seeing the story as a simple addition of units—they define neighboring sections as segments of a whole. Matisse’s Thousand and One Nights pays explicit tribute to the scenic treatment of very wide formats.

Pictures whose details read part by part avoid the single-axis characteristic of icons and marks. Emphasis is laid on the separate areas: in highly unified compositions such areas become zones. Parakeet and Mermaid is certainly nonaxial, and it is not, strictly speaking, a grid. Matissean leaf forms in different colors, shapes and sizes swirl over the surface in a multitude of directions. It does not require too much imagination to see the work as thematic and organized by zones. The left half, punctuated by the parakeet, pertains to air, while the right half, which features the mermaid, pertains to water. The sea/jungle of vegetation that envelops them is a part of both worlds at once. As a pastoral image, the theme is the unity of normally disparate realms (like those of the court and the forest in As You Like It). Here the absence of axial division becomes altogether appropriate, and the technique of collage, an integral element of meaning. As cutout representational images, collage implies that some original context of the chosen motif has been discarded and that a new one—with new meanings—has been developed, within which the original motif functions differently. For iconographic purposes, the new context is not the total image but a zone, the immediate adjoining area with its figures. This is often the way one reads a Rosenquist, for example. In older painting, pictorial zoning laws separated human and divine figures, while Persian miniatures often assign separate zones to upper- and lower-class figures.

Iconographic systems do not really concern themselves with credibility (which is a critical concern outside the system), and most of Matisse’s decorative compositions use motifs so remote and so conventional that symmetries serve purely artistic purposes. Matisse’s heraldic Apollo has a single vertical axis, but the much wider Decoration with Masks presents a multiplicity of symmetries. There two horizontally oriented grids flank a central vertical panel. Each of the three areas has a marked axis of symmetry—the mask in each horizontal panel, set above the horizontal axis, and the arch-framed vertical line-up of floral motifs in the center. The contours of the motif are certainly important here, although I don’t agree with John Neff’s claim that the drawing “saves” Matisse from decoration. Shifting from the four-petalled flower motif of the grids to the five-petalled flower of the panel introduces the hint of a secondary diagonal grid, carried through in the orientation of the cloves and the line of surprised-looking diagonals at the top border. As Fourcade says, the whole is not completely successful. In particular, the blue columns at the sides look cramped and awkward. Yet the organization is essentially chromatic. The central panel takes its closure and direction primarily from its unmodified blueness, within which white takes on a positive brilliance that is supplied in the grids by the oranges and greens. Only in the vicinity of the brown cloves does the ground white begin to match the luminosity generated by the continuous blue edge in the center.

Although Matisse died before he could resolve all the artistic problems of his monumental decorations, his work remains the most sophisticated use of symmetry and color we have seen in modern times. We have done very little compared to the subtle and intelligent masterpieces of medieval Persia. In the west iwan of the Friday mosque in Isfahan, the central axis of the great calligraphic mosaic is spelled out in the multi-directional Kufic script, but it does not force the alignment of the door below it or the decorative panels above. Symmetries abound, but they never lock us into a single closed visual field. All boundaries are abrogated by inclusion into greater wholes, so that we are finally drawn to see the entire mosque precinct as a total environment. This unity does not depend on the pervasiveness of a single historical style, even less on the consistency of a single artistic personality or “hand.” The unity is essentially formal, dependent on esthetic principles so powerful that they can sweep gardens, architecture and decoration into a single vision. Without an appreciation of the possibilities of formal principles that can integrate discontinuous areas, disjunctive patterns and materials, we can see nothing but mechanical symmetries and trivial collage.

Formal principles are not mere technical matters, relevant to the artistic process but meaningless in the final experience of work. The conventions of artistic structure are central to the pact between the artist and the audience. Form is the grammar of artistic utterance. Unless artist and audience share an understanding of how the artistic experience is set up, no sophisticated or extended expressiveness is possible. Wherever that common focus is missing, the reception of artistic intention goes awry. Thus art continues to be sold as celestial transport while artists produce it for its immediacy. Such matters are now dealt with philosophically in a long overdue recognition of the social and ideological components of perception. But formal principles have a peculiar status, half empirical and half theoretical. As supra-stylistic emergents within contemporary art, new conventions of form link “pure” perception to other components of visual experience.

Amy Goldin



1. See Jeff Perrone, “Approaching the Decorative.” Artforum, December 1976, pp. 26–30.