PRINT November 1966

Frank Stella’s New Paintings

The craving for simplicity. People would like to say: “What really matters is only the colors.” You say this mostly because you wish it to be the case. If your explanation is complicated, it is disagreeable, especially if you don’t have strong feel­ings about the thing itself.


FRANK STELLA’S NEW PAINTINGS INVESTIGATE the via­bility of shape as such. By shape as such I mean not merely the silhouette of the support (which I will call literal shape), nor merely that of the outlines of elements in a given picture (which I will call depicted shape), but shape as a medium within which choices about both literal and de­picted shapes are made, and made mutually re­sponsive. And by the viability of shape, I mean its power to hold, to stamp itself out, and in—as verisimilitude and narrative and symbolism used to impress themselves—compelling conviction. Stella’s undertaking in these paintings is therapeu­tic: to restore shape to health, at least temporarily, though of course its implied "sickness’’ is simply the other face of the unprecedented importance shape has assumed in the finest modernist paint­ing of the past several years—most notably, in the work of Kenneth Noland and Jules Olitski. It is only in their work that shape as such can be said to have become capable of holding, or stamp­ing itself out, or compelling conviction—as well as, so to speak, capable of failing to do so. These are powers or potentialities—not to say respon­sibilities—which shape never until now possessed, and which have been conferred upon it by the development of modernist painting itself. In this sense shape has become something different from what it was in traditional painting or, for that matter, in modernist painting until recently. It has become, one might say, an object of convic­tion, whereas before it was merely . . . a kind of object. Stella’s new pictures are a response to the recognition that shape itself may be lost to the art of painting as a resource able to compel con­viction, precisely because—as never before—it is being called upon to do just that.

The way in which this has come about is, in the fullest sense of the word, dialectical, and I will not try to do justice to its enormous complexity in these rough notes. An adequate account of the developments leading up to Stella’s new paintings would, however, deal with the following:

1. The emergence of a new, exclusively visual mode of illusionism in the work of Pollock, Newman and Louis. No single issue has been as continuously fundamental to the development of modernist painting as the need to acknowledge the literal character of the picture-support. Above all this has tended to mean acknowledging its flatness or two-dimensionality. There is a sense in which a new illusionism was implicit in this development all along. As Clement Greenberg has remarked:

The flatness toward which Modernist paint­ing orients itself can never be an utter flat­ness. The heightened sensitivity of the picture plane may no longer permit sculptural illu­sion, or trompe-l’oeil, but it does and must permit optical illusion. The first mark on a surface destroys its virtual flatness, and the configurations of a Mondrian still suggest a kind of illusion of a kind of third dimension. Only now it is a strictly pictorial, strictly op­tical third dimension.1

But the universal power of any mark to suggest something like depth belongs not so much to the art of painting as to the eye itself; it is, one might say, not something that has had to be established so much as something—a perceptual limitation—that cannot be escaped,2 whereas the dissolu­tion of traditional drawing in Pollock’s work, the reliance on large and generally rather warm ex­panses of barely fluctuating color in Newman’s, and the staining of thinned (acrylic) pigment into mostly unsized canvas in Louis’s were instrumen­tal in the creation of a depth or space accessible to eyesight alone which, so to speak, specifically belongs to the art of painting.3

2. The neutralizing of the flatness of the picture-support by the new, exclusively optical illusionism. In the work of Pollock and Newman, but even more in that of Louis, Noland and Olitski, the new illusionism both subsumes and dissolves the picture-surface—opening it, as Greenberg has said, from the rear4—while simultaneously pre­serving its integrity. More accurately, it is the flatness of the picture-surface, and not that sur­face itself, that is dissolved, or anyway neutralized, by the illusion in question. The literalness of the picture-surface is not denied; but one’s experience of that literalness is an experience of the properties of different pigments, of foreign substances applied to the surface of the painting, of the weave of the canvas, above all of color—but not, or not in particular, of the flatness of the support. (One could say that here the literalness of the picture­surface is not an aspect of the literalness of the support.) Not that literalness here is experienced as competing in any way with the illusionistic presence of the painting as a whole; on the con­trary, one somehow constitutes the other. And in fact there is no distinction one can make between attending to the surface of the painting and to the illusion it generates: to be gripped by one is to be held, and moved, by the other.

3. The discovery shortly before 1960 of a new mode of pictorial structure based on the shape, rather than the flatness, of the support. With the dissolution or neutralizing of the flatness of the support by the new optical illusionism, the shape of the support––including its proportions and ex­act dimensions—came to assume a more active, more explicit importance than ever before. The crucial figures in this development are Frank Stella and Kenneth Noland. In Stella’s aluminum stripe paintings of 1960, for example, 2 ½-inch wide stripes begin at the framing-edge and reiterate the shape of that edge until the entire picture is filled; moreover, by actually shaping each picture—the canvases are rectangles with shallow (one stripe deep) notches at two corners or along the sides or both—Stella was able to make the fact that the literal shape determines the structure of the entire painting completely perspicuous. That is, in each painting the stripes appear to have been generated by the framing-edge and, starting there, to have taken possession of the rest of the canvas, as though the whole painting self-evidently fol­lowed from, not merely the shape of the support but its actual physical limits. Noland, on the other hand, cannot be said to have come into contact with the physical limits of the support until his first chevron paintings of 1962. His initial break­through to major achievement in the late 1950s came when he began to locate the center of concentric or radiating motifs at the exact center of square canvases. This related depicted shape to literal shape through a shared focus of symmetry. Whether or not Noland recognized that this was the significance of centering his rings and armatures of color is less important than that he ex­perienced the centering itself as a discovery: a constraint in whose necessity he could believe, and in submission to which his magnificent gifts as a colorist were liberated. His shift to chevron motifs a few years later was, I believe, inspired in part by the need to achieve a more active or explicit relation between depicted and literal shape than the use of concentric rings, none of which actually made contact with the framing-edge, allowed. Within a few months Noland discovered that suspending his chevrons from the upper cor­ners of the support (the bottom edge of the lowest chevron running into each corner) empowered him, first, to prize loose the point of the bottom­-most chevron from the mid-point of the bottom framing-edge, and second, to pull all the chevrons away from the central axis of the painting—besides enabling him to work with rectangular formats other than the square. In these paintings—the asymmetrical chevrons of 1964—the exact dimen­sions of the support become important in this sense: that if the edge of the bottom-most chevron did not exactly intersect the upper corners of the canvas, the relation of all the chevrons—that is, of depicted shape—to the shape of the support would be acutely problematic and the ability of the painting as a whole to compel conviction would be called into question. Since that time, apparently in an attempt to make depicted shape relate more generally to the shape of the support in its entirety, Noland too has shaped his pictures. (His recent work includes a number of narrow diamond-shaped pictures which I will discuss fur­ther on.) It cannot be emphasized too strongly, however, that Noland’s chief concern throughout his career has been with color—or rather, with feeling through color—and not with structure: which makes the role that structural decisions and alterations have played in his development all the more significant. This is not to say that Noland’s colorism has had to maintain itself in the teeth of his forced involvement with structural concerns. On the contrary, it is precisely his deep and impassioned commitment to making color yield major painting that has compelled him to dis­cover structures in which the shape of the support is acknowledged lucidly and explicitly enough to compel conviction.

4. The primacy of literal over depicted shape. In both Noland’s and Stella’s (stripe) paintings the burden of acknowledging the shape of the support is borne by the depicted shape, or perhaps more accurately, by the relation between it and the literal shape—a relation that declares the primacy of the latter. And in general the development of modernist painting during the past six years can be described as having involved the progressive assumption by literal shape of a greater—that is, more active, more explicit—importance than ever before, and the consequent subordination of de­picted shape. It is as though depicted shape has become less and less capable of venturing on its own, of pursuing its own ends; as though unless, in a given painting, depicted shape manages to participate in—by helping to establish—the au­thority of the shape of the support, conviction is aborted and the painting fails. In this sense de­picted shape may be said to have become depend­ent upon literal shape—and indeed unable to make itself felt as shape except by acknowledging that dependence.


Let this stand as the general background of con­cerns from which Stella’s new paintings emerge. A fuller delineation of their immediate context is still required, however, if the concentrated and radical exploration of shape which they undertake is meaningfully to be described.

Although Noland has found it necessary to de­velop structures in which the shape of the support plays a determining role, his continuing ambition to liberate feeling through color has made him reluctant to call attention to the physical limits of the support—the way, for example, Stella’s stripe paintings call attention to them. In the latter Stella identifies the shape of a given picture with its framing-edge, thereby assimilating the first to the second. Noland, on the other hand, is anxious to keep this from happening; or rather, the same con­cerns that, in effect, compel him to acknowledge the shape of the support also compel him to try to keep our awareness of its physical limits to an absolute minimum. Above all Noland is anxious to keep us from experiencing the shape of his paintings as edge, hence as something literal and non-illusive; and in order to make sure this does not happen, he tries to keep us from experiencing the shape at all. It is as though, for Noland, to experience the shape of a painting is inescapably to experience the painting itself as something literal, as a kind of object; and this would compro­mise its presence as visual illusion. And in general the shapes of his paintings are never experienced as acutely as the limits of and boundaries between the depicted elements within them.

That Noland’s paintings avoid calling attention to their physical limits does not mean that those limits are not still there—and there to be felt. What, so to speak, put them there to be felt is the acknowledgment of literal shape which the paintings themselves make—and which, as it were, exerts upon the edge a kind of pressure, or inqui­sition, from which it cannot escape. If Noland’s paintings offered some alternative to our expe­riencing their shapes as an aspect of their literalness—either by positively identifying literal shape with illusion, or by repudiating it altogether—their effi­cacy as illusion or presence would not be, as I sometimes find them to be, threatened by, or at, the edges. (The suggested alternatives are those explored in Olitski’s spray paintings and Stella’s new pictures respectively.) This is not to deny that throughout Noland’s masterful paintings of the past several years the literal shape of the support is made to seem the outcome, or result, of the depicted shapes—rather than, as in Stella’s stripe paintings, the other way round. But the fact remains that a painting by Noland cannot be said to hold as shape—it cannot be said to need to either—but merely to have one, like any solid object in the world. Or rather, it is as though the shape were itself a kind of object in the world—an ob­ject that has been prized loose from the illusion­istic presence of the painting by its very importance to the structure of that painting. One is made to feel, that is, that in these paintings the distinction between depicted and literal shape marks a dif­ference, not simply between two kinds of shape—each, so to speak, conceived of as a pictorial entity—but between two utterly distinct and different kinds of entities. The first of these, depicted shape, is powerless to make itself felt except by acknowledging the primacy of the other, while that other—literal shape—does not hold as shape. It is a shape, but what this suddenly seems to mean is only that it is an object in the world—an object whose relevance to our experience of the painting is not clear.

The fact that in some of his recent paintings Noland has not been content simply to minimize the shape of the support, but has instead begun actively to subvert it, suggests that his previous paintings may have come to seem problematic to him for the sorts of reasons I have been discussing. For example, his last show at the Emmerich Gallery included four 8-by-2-foot diamond-shaped paint­ings in each of which four relatively broad bands of color run parallel to one or the other pair of sides, thereby acknowledging the shape of the support. At the same time, however, the extreme attenuation of these pictures makes them unable to contain within the limits of the support their own extraordinary presences as color and illusion. In the grip of the sheerly visual illusion generated by the interaction of the colored bands, the acute-­angled corners of the supports appear to vibrate and shimmer, to erode both from within and with­out, to become even more attenuated and needle­like than they are, while the obtuse-angled corners tend to round off, to appear dull or blunt. The result is that the physical limits of the supports are overrun, indeed all but dissolved, by the paint­ing’s illusionistic presences. At the same time an effect like that of simultaneous contrast between the colored bands makes them appear to overlap one another physically, like shingles. So that, while the physical limits of the support are assaulted by illusion, the (depicted) boundaries between the bands are the more acutely felt—as if absorbing the literalness or objecthood given up by the support. Moreover, because some sort of progressive sequence (e.g., of value) among the bands appears to be required for the illusionistic overlapping I have just described, one’s actual experience, or sensation, of these paintings is directional. One is aware, that is, of being held and moved by a pro­gression or sequence—a resource until now for­eign to modernist painting—and this further intensifies the assault these paintings make on their own static, literal shapes. In several other paintings in the same show—long horizontal rectangles with a few parallel bands of color, again arranged pro­gressively, running their entire lengths—Noland achieved an equal subversion by somewhat simpler means: the rectangles are too long, and propor­tionally too narrow, to be experienced as discrete shapes. Instead, confronted head on, they seem to extend almost beyond the limits of our field of vision, to become nothing but extension, to only end up being rectangular; approached from the side (their length makes this inviting) what is striking is not their rectangularity but the speed with which that rectangle—or rather, the speed with which the colored bands—appears to dimin­ish in perspective recession.5 Here, again, although the relation of depicted to literal shape within each painting acknowledges in the simplest pos­sible way the primacy of the latter, the actual limits of the support do not quite manage to constitute a single, definite shape, while the boundaries between the colored bands seem almost tactile or stepped by comparison.

I have argued elsewhere that the desire to oppose the kind of structure at work in Noland’s and Stella’s paintings provided much of the moti­vation behind Jules Olitski’s first spray paintings of 1965.6 These pictures are completely devoid of depicted shape, and in fact represent what is al­most certainly the most radical and thoroughgoing attempt in the history of modernism to make major art out of nothing but color. At the same time, no paintings have ever depended so completely or so nakedly for their success on the shape (and in particular the proportions) of their supports, experienced, one might say, in relation both to nothing particular within each painting and to everything it contains.7 It is, I think, true of these paintings—and of no others—that they succeed as paintings just so far as they succeed, or hold, or stamp themselves out, as shapes. And in fact no shapes, depicted or literal, have ever stamped themselves out more compellingly or more feelingly. In the sense in which I have been using the word, it is not true to say that these paintings acknowledge the shape of the support; but their relative quality depends even more intimately upon it. (In this respect they differ sharply from Noland’s paintings, whose success or failure as art does not depend on their efficacy as shapes.) So that while they were made in opposition to a mode of pictorial organization which established the primacy of literal over depicted shape, in these paintings literal shape’s assumption of authority has become not merely relative but absolute: as though it alone were capable of performing the office of shape, of being felt as shape.

The very success of Olitski’s paintings as shapes lays bare the conditions that make this success pos­sible—conditions it is hard to imagine any paintings but Olitski’s being able to fulfill. It is, to begin with, clearly central to their potency as shapes that they are wholly devoid of depicted shape; but it is also clear that two paintings equally devoid of depicted shape may succeed unequally as shapes—and, therewith, as works of art. More­over, virtually all the best early spray paintings belong to a single format—the narrow vertical rectangle—and the more any painting departs from this format toward the horizontal or square the more likely it is to fail. This is connected with the fact that when the early spray paintings fail—relatively speaking—we tend to see the framing-edge as marking the limits of a spatial container, and the sprayed canvas itself as some­thing like background in traditional painting. The narrow vertical format somehow keeps this from happening: not by denying the illusion but, so to speak, by making it self-sufficient, a presence, like that of a human figure, instead of a void waiting to be filled. In the best narrow vertical paintings the framing-edge does not appear to contain the illusion; on the contrary, the illusion itself contains the limits of the support. So that whereas the relatively square paintings can often be seen as receptacles which may happen to be empty but which could be filled, could contain objects, the best of the narrow vertical pictures already contain their object, namely, the edges of the painting, its outermost and tactile limits. (In this connection it is significant that, in the paintings in question, all relatively well-defined bursts of color and variations in value are restricted to the vicinity of the edges and corners of the canvas.) One might say that whereas in traditional painting the illusion of a tactile space commences at the inside of the framing-edge, in the best early spray paintings the illusion of something like depth or space accessible to eyesight alone ends at the outside of that edge. And that whereas traditional illusionism begins at the surface of the canvas, the strictly visual mode of illusionism of the Olitskis in question ends there.

In recent paintings, such as those exhibited at the Biennale, Olitski has taken to masking out all but thin bands around two or three sides of the sprayed canvas, spraying some more and then removing the masking. The result is a clear dif­ference between the previously masked and un­masked areas, a difference that can be subtle or blatant, and can vary enormously from place to place along the boundary between the bands and the rest of the picture which they partly frame. Further, this internal “frame” is not strictly para­llel to the edges of the canvas; sometimes its long vertical component is inflected slightly away from the perpendicular. Both these developments can be understood, at least in part, as undermining or mitigating the absoluteness of the primacy which literal shape assumes in his first spray paintings. To begin with, the partial internal “frame” amounts to something like depicted shape; and this in itself means that the quality of individual paintings no longer solely depends on the almost unanalyzable relation between the sprayed canvas and the shape of the support that apparently governs the success or failure of the first spray paintings. But because the boundary between the framing bands and the rest of the painting consists of the same pictorial stuff—the same sprayed color—as the areas it delimits, the role of the internal “frame” as a kind of middle term between the shape of the support and the rest of the painting is far more complex than that played by depicted shape in Noland’s paintings or Stella’s stripe paintings. To be sure, the internal “frame”—or, more accurately, the boundary between the “framing”-bands and the rest of the painting—relates structurally to the shape of the support. But it also establishes an extraordinary, and indeed unprecedented, conti­nuity across that boundary. This enables the paint­ings in question both to contain depicted shape, or something like it, and yet be seen as pictorially seamless and integral—like the early spray paintings. Moreover, the fact that the long vertical component of the internal “frame”—or the ver­tical boundary between that “frame” and the rest of the picture—is sometimes inflected away from the perpendicular further reduces the perspicuity of literal shape’s primacy at the same time that it acknowledges, or anyway is made possible by, this primacy. That is, in these paintings the primacy of literal shape is such that even a slight departure from verticality within the painting makes itself felt with an intensity of expression I, for one, find astonishing. But it is precisely the strength of this primacy that enables the paintings in question both to tolerate the departure and to move us by it. The very acuteness, even poignancy, of our ex­perience of what is, after all, an extremely slight inflection, acknowledges the strength, and more than that, the depth, of the norm from which that inflection departed—in this case, the shape of the support. But the fact remains that what we actually feel, and are inexplicably moved by, is the inflec­tion from the norm rather than the norm itself. All these differences between his early and later spray paintings have enabled Olitski to realize his ambitions across a considerably wider range of formats. And if it is true, as I believe it is, that none of the later spray paintings (none that I have seen at any rate) stamps itself out as shape quite as powerfully as the best of the early ones, part of what these differences have meant is that the quality of a given picture no longer depends en­tirely on its success or failure as shape.

There is, then, a sense in which the conflict between a sheerly visual or optical mode of illu­sionism and the literal character of the support is central both to Noland’s and Olitski’s paintings. In Olitski’s pictures—at any rate, the early spray pictures—the conflict is naked and direct. It is, for example, felt in the threat that the illusion will, so to speak, come detached from the framing-edge, leave the literal shape hanging on the wall and situate itself indefinitely further back. This is not to say that when this does not happen the illusion is properly described as attached to the edge of the support. Rather, the physical limits of the sup­port mark, or declare, or simply are, the limits of the illusion itself. We become aware of the con­flict in question only when, in relatively less suc­cessful paintings, illusion and literal shape actually part company—despite the fact that when this happens, the illusion can no longer be described as sheerly visual, any more than background in traditional painting can be characterized in these terms. In Noland’s paintings, on the other hand, opticality and the physical limits of the support are not juxtaposed against one another as in Olit­ski’s paintings. Instead, it is the structure of his paintings—the relation between depicted and literal shape in them—that, so to speak, brings the two into conflict with one another. This is what makes the fact that his paintings do not stamp themselves out as shapes feel like a failure or refusal to do so—a failure or refusal that, espe­cially in the light of Olitski’s spray paintings, leaves the literalness, or objecthood, of the limits of the support there to be felt. I said earlier that No­land himself seems to have become increasingly troubled by this, and in his recent narrow diamond and long horizontal rectangle paintings appears to have tried to subvert their shapes. But it should be remarked that this does not resolve the conflict between opticality and the literal character of the support that, I have claimed, is central to both No­land’s and Olitski’s work; if anything, it intensifies it.


It is only in the presence of this conflict that the question of whether or not a given painting holds or stamps itself out as shape makes full sense—or rather, only here that the issue of “the viability of shape as such” characterizes a specific stage in resolving, or unfolding, problems of ac­knowledgment, literalness and illusion which, as I said at the beginning of these notes, have been the issues of modernism from its beginning. In Stella’s stripe paintings, for example, the reitera­tion by the stripes of the irregular shapes of the support makes the dependence of depicted on literal shape far more explicit than Noland’s paint­ings ever allow it to seem. But if one asks whether Stella’s paintings hold better or make themselves felt more acutely as shapes than Noland’s paintings, the answer, I think, is not just that they do not, but that the whole issue of holding or failing to hold is much less relevant to them. That is, because they are not illusive in anything like the way Noland’s and Olitski’s paintings are, there is nothing for them to hold as shapes against.8a

I must emphasize that in defining this conflict between visual illusionism and literal shape in No­land’s and Olitski’s paintings I have not meant to imply an adverse criticism either of the quality of their best paintings or of the general level of their respective achievements. This is worth stressing precisely because there are certain younger artists to whose sensibilities all conflict between the lit­eral character of the support and illusion of any kind is intolerable, and for whom, accordingly, the future of art lies in the creation of works that, more than anything else, are wholly literal—in this respect going beyond painting. It should be evident that literalist sensibility is itself a product, or by-product, of the development of modernist painting itself—more accurately, by the increas­ingly explicit acknowledgment of the literal charac­ter of the support that has been central to that development. But it ought also to be observed that the literalness isolated and hypostatized in the work of artists like Donald Judd and Larry Bell is by no means the same literalness as that acknowl­edged by advanced painting throughout the past century: it is not the literalness of the support. Moreover, hypostatization is not acknowledgment. The continuing problem of how to acknowledge the literal character of the support—of what counts as that acknowledgment—has been at least as cru­cial to the development of modernist painting as the fact of its literalness; and this problem has been eliminated, not solved, by the artists in ques­tion. Their pieces cannot be said to acknowledge literalness; they simply are literal. And it is hard to see how literalness as such, divorced from the conventions which, from Manet to Noland, Olitski and Stella, have given literalness value and have made it a bearer of conviction, can be experienced as a source of both of these—and what is more, one powerful enough to generate new conven­tions, a new art.8

Because Frank Stella’s stripe paintings, especially those executed in metallic paint, represent the most unequivocal and conflictless acknowledg­ment of literal shape in the history of modernism, they have been crucial to the literalist view I have just adumbrated, both because they are seen as extreme instances of a putative development within modernist painting—i.e., the increasingly explicit acknowledgment of literalness per se—and because they help to make that development visi­ble, or anyway arguable, in the first place. They are among the last paintings that literalists like Judd are able to endorse more or less without reservation: largely because the ambition to go beyond them—to pursue their apparent implica­tions—was instrumental in the abandonment of painting altogether by these same artists.

In Stella’s new paintings, however, the relation between depicted and literal shape seems nowhere near as straightforward in its declaration of the latter as in the stripe paintings—or, for that matter, in Noland’s work. Rather, there is a new and even somewhat startling freedom both in the variety of shapes used in a given picture and in their disposition relative to one another and to the support. This is not to say that the shape of the support is either ignored or denied. On the contrary, it is very clearly taken into account; but the way in which this is accomplished does not affirm the de­pendence of depicted on literal shape so much as it establishes an unprecedented continuity between them. In Moultonboro III, for example, the shape of the support is an irregular polygon formed by superimposing a triangle and a square, the first apparently having come slanting down from the upper right to wedge itself deeply into the second. (In Chocoura III a triangle is superimposed on a rectangle; the same is true of Tuftonboro III except that the rectangle is missing its upper right corner; while in Conway III a parallelogram is superimposed on another, this time more horizontal, rec­tangle. These are the only formats among the eleven Stella has used for his new paintings which have been arrived at by superimposition, pure and simple.9) The triangle itself comprises two elements—an 8-inch wide light yellow band around its perimeter and the smaller triangle, in Day-Glo yel­low, bounded by that band—both of which seem to be acknowledging, by repeating, the shape of the support. For that reason it is almost startling to realize that only a relatively small segment of the triangle coincides with, is part of, the shape of the support. Most of the triangle lies wholly inside the picture and, in the terms proposed at the out­set, exists only as depicted shape. Even more sur­prising, however, is the fact that realizing this does not in itself undermine the triangle’s efficacy as shape. It is as though that segment which coincides with the literal shape of the painting somehow implies the rest of the triangle—the merely de­picted portion of it—strongly enough for the latter to succeed as shape despite its failure to relate self-evidently to any other segment of the framing­edge. But it would, I think, be just as true to one’s experience of Moultonboro III to claim that what enables the relatively small segment of the triangle that coincides with the shape of the support to make itself felt as shape is what might be called the implicative power, in this context, of the merely depicted portion of the triangle. The yel­low triangular band and the Day-Glo triangle within it are, after all, what make that segment intelligible: without them, and without another largely internal shape—the blue Z-form in which the triangular band (and hence the triangle as a whole) rests—the upper righthand segment of the support would not be part of a triangle but would belong instead to the literal shape of the painting perceived in its entirety as an irregular seven-sided polygon, whereas in the painting as it stands roughly the opposite is the case. The beholder is, in effect, compelled not to experience the literal shape in its entirety—as a single entity—but rather to perceive it segment by segment each of which is felt to belong to one or another of the smaller shapes that constitute the painting as a whole.

This last point is important. For one thing it indi­cates a crucial difference between Stella’s new paintings on the one hand and Noland’s and Olit­ski’s pictures, as well as Stella’s own previous work, on the other. In this respect Noland’s paint­ings in general are closer to Olitski’s spray pictures than to Stella’s new work, despite the fact that­—unlike Olitski—both he and Stella work with non­rectangular supports and discrete areas of color. It also suggests that, confronted by Stella’s new paintings, the distinction between depicted and literal shape becomes nugatory. It is as though in a painting like Moultonboro III there is no literal shape and, therefore, no depicted shape either; more accurately, because none of the shapes that we experience in that painting is wholly literal, there is none that we are tempted to call merely depicted. There are shapes that lie entirely inside the picture limits—that do not make contact with those limits—just as there are others that partly coincide with the edge of the support. But neither enjoys precedence over the other—in particular, neither sponsors nor guarantees the other’s efficacy as shape—any more than either the depicted or the literal limits of a shape that partly coincides with the edge of the support is experienced as more fundamental to that shape’s efficacy than the other. Both types of shape succeed or fail on exactly the same grounds—grounds that do not concern the relation of a given shape to the shape of the support seen in its entirely. Each, one might say, is implicated in the other’s failure and strength­ened by the other’s success. But the failure and success of individual shapes cannot be understood in terms of the distinction between depicted and literal shape with which I have been working until now.

The relation between depicted and literal shape that holds in the stripe paintings no longer holds in these, not because the relation has been altered or defied but because the distinction is defeated by the paintings themselves. Nothing, apparently, is more central to their conception than the desire to establish all shapes on an equal footing—to make pictures which comprise nothing but indi­vidual shapes each of which is felt to stand or fall without reference, or appeal, to a single master shape, a support seen as a single entity. In fact, because in most of the new pictures the physical limits of the support are not perceived as constituting a single shape, there is even a sense in which—despite the non-rectangularity of their supports—the pictures in question are not shaped: if being shaped implies having an enclosing shape, the term is less applicable to Stella’s non-rectangular pic­tures than, for example, to Olitski’s rectangular ones. (In this same sense the physical limits of the support can be said not to constitute a framing edge.) It should be remarked, however, that Stella could not have made paintings of which this is true except by using irregular supports—that is, by avoiding not only the rectangle but geometri­cally regular figures of any kind—in order to pre­vent the eye from instantly perceiving the shape of the support as a single entity. Moreover, the fact that in perhaps the three most successful sub-series of the new paintings—the Union, Effingham and Wolfeboro pictures—Stella has not used regular geometric shapes at all seems to me to have some­thing to do with their success. In certain other of the new paintings the eye pounces on a shape of this kind and only then takes in the rest of the painting. When this happens the latter is experienced both as some kind of remainder or appen­dage to the initial shape, and as in some sense inferior as shape in comparison with it. More ac­curately, the rest of the painting is put under enor­mous pressure by the geometrically regular shape to match its own sheer perspicuousness—which, inevitably, it cannot do. One might say that the regular geometric shape is a shape—a square, a rectangle, a triangle, a pentagon—while the other is not;10 this in itself seems to be enough to disturb the parity among shapes on which the success of Stella’s new pictures seems largely to depend. In Moultonboro III—as in the Chocoura and Tuftonboro paintings mentioned earlier—the desire for parity manifests itself in the implied juxtaposition of two equally regular and hence equally perspicu­ous shapes (i.e, a triangle and a rectangle). But in each painting the two shapes compete for one’s attention, almost as though they were juxtaposed to one another within a larger conventional paint­ing—with the result that one tends to pull back, to distance the pictures in question and, as it were, to surround each of them with an imaginary rec­tangular frame large enough to contain the paint­ing and some space around it besides, whereas in the Union, Effingham and Wolfeboro paintings there is no competition for one’s attention. None of the elements they comprise is in any way per­spicuous, or even particularly interesting in itself; one does not, so to speak, recognize any of them—except perhaps the trapezoid at the bottom of Wolfeboro (and then, as we shall see, it is an open question what one recognizes it as). And far from being inclined to distance or frame these pictures, I for one feel strongly that—more than any pictures I have ever seen—they ought not to be framed at all. ­

Moreover, the fact that the physical limits of the support do not make themselves felt as a single entity but, in effect, belong segment by segment to individual shapes the remainder of whose limits do not coincide with those of the support implies a strong and, I think, unprecedented continuity between the “outside” of a given painting (i.e., its physical limits) and its “inside” (i.e., everything else). The 8-inch wide colored bands deployed throughout the new paintings are a kind of para­digm for this continuity. In general one such band begins by running along at least one side of the support—in Union III the same band runs along four of five sides—until, at some point or other, it encounters another shape whose “merely” depicted portion it follows into the heart of the canvas, taking the beholder with it. That is, a par­ticular stretch of the edge of the painting is, in effect, first isolated from the rest of that edge—the band, as it were, broadens and usurps the of­fice of the edge—and is then carried into the inte­rior of the painting. The result is both that the paintings are infused with an extraordinary and compelling directionality, and that one is made to feel that the important difference in them is not between “inside” and “outside” but between open and closed. The side or sides along which the bands run are experienced as closed (or closed off) while the others are felt as open—and when, as in Union III and Effingham III, the open side or sides are at the top of the painting, the effect can be one of an astonishing vertical acceleration, or soar­ing, or release. There is, one might say, no more “outside” or “inside” to the best of Stella’s new paintings than to the individual shapes they com­prise; and to the extent that a given shape can be said to have an “outside” and “inside” the rela­tion between the two is closer to that, say, be­tween the edge of a table-top and the rest of that table-top than to the relation between the edge of a Noland or an Olitski or even a Stella stripe painting and the rest of that painting. This is not to say that Stella’s new pictures are nothing more than objects. Unlike Judd’s constructions, for exam­ple, or Bell’s glass boxes, they do not isolate and hypostatize literalness as such. At the same time, however, literalness in them is no longer experi­enced as the exclusive property of the support. Rather, it is suffused more generally and, as it were, more deeply throughout them. It is as though literalness in these pictures does not be­long to the support at all except by coincidence—specifically, the coincidence between the limits of the individual shapes that constitute a given paint­ing and the physical limits of the support; as though, that is, one’s experience of literalness is above all an experience of the literalness of the individual shapes themselves. Though of course what I have just called their literalness is identical with their success as shapes—and that, while not a direct function of the literalness of the support, is at any rate inconceivable apart from that literal­ness.

The dissociation of literalness from the support that I have just tried to describe is intimately re­lated to another aspect of Stella’s new paintings, namely, their extraordinary, and sheerly visual, illusiveness. This is not to say that, in a given pic­ture, each shape seems to lie in a definite or specifiable depth-relation to every other. On the contrary, nothing is more fundamental to the na­ture of the new paintings’s illusiveness than the extreme ambiguity, indeterminacy and multival­ence of the relations that appear to obtain among the individual shapes, as well as between those shapes and the surface of the picture (or, at any rate, the plane of that surface). In Moultonboro III, for example, although one is not made to feel that the light yellow triangular band stands in any sin­gle or definite spatial relation to the turquoise blue Z-shaped band into which it fits, one neverthe­less experiences their juxtaposition somewhat as though both were objects in the world, not simply nor even chiefly shapes on a flat surface—objects, moreover, whose relation to one another, and in­deed whose actual character, are ineluctably am­biguous. This is most salient in the case of the Z-shaped turquoise band, largely because—or so it seems—its top and bottom segments are not parallel to one another. (The first, running as it does along the upper edge of the square, is hori­zontal, while the second, flush with the lowest side of the triangle, slants from the lower left toward the upper right.) That is, one tends to see the bot­tom segment, or the bottom two segments, as though somewhat from above and in perspective—while at the same time one is not given enough data to locate them in a definite spatial context, in relation either to contiguous shapes or to some ground plane. Moreover, because the top segment of the Z-form runs across the upper edge of the square and is therefore horizontal, one tends to experience that segment as frontal. But this would mean that the Z-form is not only irregular in two dimensions but bent or warped in three—though it is not at all clear which segment or segments are bent or warped and which, if any, are to be taken as normative. The bevelled ends of the Z-form, each parallel to nothing else in the painting, com­pound the ambiguity by implying that the respec­tive planes of both the bottom and top segments are warped away from, or anyway are oblique to, that of the picture-surface—though, of course, they might not be. (Almost all the bands in Stella’s new paintings are bevelled in this way—a brilliant stroke that adds immeasurably to the illusionistic power, and general complexity, of the paintings in question. In fact its absence from Conway III is partly responsible for the relatively flat and con­ventional appearance of that picture.) The result is that the Z-form is seen as participating in a wide range of equally ambiguous and indeterminate spatial situations—more accurately, an entire gamut of such situations each of which is simultaneously not merely compatible with but continuous with or transparent to every other. But it is not just the situations in which the Z-form finds itself or the relationships into which it enters that continually escape one but—more than anything else—its actual shape. (Similarly, when one “recognizes” the shape at the bottom of Wolfeboro III, does one recognize it as a trapezoid—its configuration on the surface of the canvas—or as a rectangle seen in perspective?) It is as though across the entire gamut of illusionistic possibilities the real Z-form—flat or warped, regular or irregular, partly or wholly parallel or oblique to the picture-surface—lies somewhere out there, beyond the painting, waiting to be known. There is, of course, a “real” Z-form on the surface of the canvas. But the con­figuration on that surface of the individual shapes that constitute a given picture is no more definitive in this regard than their possible configurations in illusionistic space: above all because, as I have claimed, literalness in these paintings is primarily experienced as the property, not of the support, but of the shapes themselves. All this makes Stella’s new paintings the most radically illusive and irre­ducibly ambiguous in the history of modernism. Radically illusive in that what is rendered illusive in them is nothing less than literalness itself; and irreducibly ambiguous in that the shapes they com­prise are experienced as embracing an entire gamut of existential possibilities—including their juxtaposition on the surface of the canvas—each of which is simultaneously continuous with every other, and none of which is sufficiently privileged to make one feel that it, at any rate, is really there. There is, one might say, no it at all.

Stella’s new paintings, then, depart from his stripe paintings in two general respects—first, by not acknowledging literal shape, and second, by resorting to illusion—both of which ought to make them unpalatable to literalist sensibility. And in­deed I want to suggest that it is one of the most significant facts about his new pictures that Stella seeks in them to repudiate—not literalist taste or sensibility exactly—but the literalist implications which, in the grip of a particular conception of the nature of modernist painting, his stripe paint­ings appear to carry. This is not to claim that his new pictures are chiefly a response to the drawing of those implications by others—Judd, for exam­ple. Rather, I am suggesting that it was in his own unwillingness, even inability, to pursue beyond painting what were to him as well—if not indeed before anyone—his stripe paintings’s apparent implications in that direction that Stella discovered both the depth of his commitment to the enter­prise of painting, and the irreconcilability with that commitment of what may be called a reduc­tionist conception of the nature of that enterprise.11 At the same time it is hard not to see their relation to Noland’s and Olitski’s paintings as issuing, at least in part, from a dissatisfaction, or anyway an uneasiness, with their work that—to my mind, at any rate—has much in common with that which literalist sensibility appears to feel. Moreover, it is tempting to regard this in turn as evidence in favor of the suggestion that the impulse behind the work of literalists like Judd and Bell is anything but alien to Stella. Because if it is true that, unlike Noland and Olitski, Stella has actually felt a reductionist conception of his undertaking urge toward the iso­lation and hypostatization of literalness, it would be surprising if there were not at least some agree­ment between his response to painting other than his own and the literalist attitude toward that same painting. And in fact Stella’s new paintings can, I believe, be seen as responding critically to the same aspect of Noland’s and Olitski’s paintings that, I suggested earlier, literalist taste finds unacceptable, though here again the differences between Stella and the literalists lie deeper than their apparent agreement. From a literalist point of view the aspect in question is experienced as a conflict between pictorial illusion of any kind on the one hand and literalness as such on the other; this conflict is unacceptable because it compromises the latter; and its elimination entails making works of art (or putative works of art) that are nothing but literal—works in which illusion, to the extent that it may be said to exist at all, is itself literal. Whereas Stella’s new paintings, by making literalness illusive, not only come to grips with but actually resolve what I characterized earlier as the conflict in Noland’s and Olitski’s paintings between a particular kind of pictorial illusionism—i.e., addressed to eyesight alone—and the literal character of the support. And by so doing they unmake—at least in the event and for the moment—the distinction between shape as a fundamental property of objects and shape as an entity belonging to painting alone that emerges for the first time in Noland’s and Olitski’s paint­ings.


In closing I want merely to touch on another aspect of Stella’s new paintings—namely, what seems to me their intimate and profoundly sig­nificant relation to the finest modernist sculpture of the recent past. (I am thinking chiefly of the work of English sculptor Anthony Caro.) Almost any of the remarks and observations I have made about the new pictures could, I think, lead to an obvious comparison with Caro’s sculptures: what, for example, do pieces like Bennington and Yellow Swing do if not make literalness illusive? Moreover, the relation between Stella and contemporary sculpture is far from superficial or coincidental. Rather, it has to do with the problematic character of shape in the most advanced painting of our time—even, I want to say, with the nature of shape itself, with what shape is. In any case, I am suggesting that one result of the development within modernist painting discussed in these notes is that for the first time since the late 18th century sculpture is in a position to inspire painting; and that in Stella’s recent paintings this has actually begun to happen. At the same time, however, painting is in a position, not simply to be inspired by advanced sculpture, but in certain respects fundamental to that sculpture actually to have an advantage—though not of quality—over it. I will mention three: (1) The irreducible ambiguity of the visual illusionism in Stella’s new pictures goes beyond advanced sculpture in the direction of the opticality and illusiveness—of seeming a kind of mirage—that, as Greenberg was the first to remark, is basic to it.12 Because sculpture is literal it can, in the end, be known; whereas the shapes that constitute Stella’s new paintings, and the new paintings as experienced wholes, cannot. (2) The fact, or the convention, that paintings hang on a wall means that Stella’s new paintings begin off the ground; whereas advanced sculpture—which, as Greenberg has again remarked, is illusively weightless—has to begin at ground level and liter­ally climb to whatever height it reaches. This “ad­vantage” is perhaps most strikingly evident in Effingham III, largely because that painting as a whole is most like a ground-plan. Union III, as well, profits from it immensely. And in general Stella can float or suspend elements as though without visible means of support. (3) There is no general difficulty about the use of color in Stella’s paintings; but the problem of color in contempor­ary sculpture seems—to me, anyway—to be acute. And by this I mean not simply the propriety of applied color but the fact that all sculpture—like all solid, opaque objects—is colored, or has color, or anyway has surface. It is as though, finally, the opticality toward which advanced sculpture aspires brings one up short, not against its literal­ness exactly, but against the fact that when we perceive a solid object eyesight makes contact with no more than its surface (and then only part of that). That is to say, advanced sculpture, such as Caro’s, makes this fact a disturbing one, and in effect thrusts it into our awareness. It makes us note it, whereas painting, one wants to say, in comparison with sculpture, is all surface.13 (Which is not at all the same as saying that it is done on a flat and very thin surface; an element of equal thinness in a Caro is experienced as solid.) Stella’s paintings, by the very closeness of their relation to advanced sculpture, make this difference more salient than it has ever been.

––Michael Fried



1. “Modernist Painting,” The New Art: A Critical Anthology, edited by Gregory Battcock, New York, 1966, p.107.

2. Mondrian, in his paintings of the 1920s and after, often seems to be attempting to combat just this minimal illusionism. Some­times, for example, he stops his black lines short of the framing­edge, thereby emphasizing their paintedness, i.e., the fact that they are marks on a flat surface. In other paintings he takes the more radical step of continuing the black lines and even the blocks of color past the edge onto the sides of the canvas (which appears to have been meant to be exhibited with its sides visible). The result is that one tends to see these paintings as solid slabs, which helps to counteract—thought it cannot efface—their minimal illusionism.

3. For discussions of these developments see Greenberg’s essay “Louis and Noland,” Art International, IV:5, pp. 26–29, and my Three American Painters, Cambridge, 1965. The latter also dis­cusses in some detail the emergence of what I have called “de­ductive structure,” a development I adumbrate here.

4. Greenberg says this of Noland’s paintings in his “Louis and Noland,” p.28.

5. That Noland’s long horizontal paintings make their own shapes ungraspable in this way was observed by Rosalind Krauss in her article "Allusion and Illusion in Donald Judd,’’ Artforum, IV:9, p. 26. In the same issue Mrs. Krauss discusses Stella’s new paint­ings, pp. 47–49.

6. In my essay “Jules Olitski’s New Paintings,” Artforum, IV:3, pp. 36–40.

7. In his brief remarks on Olitski’s work, published in the cata­log to the United States pavilion at the last Biennale, Clement Greenberg wrote, “The degree to which the success of Olitski’s paintings depends on proportion of height to width in their enclosing shapes is, I feel, unprecedented.” Greenberg goes on to note the relative superiority of the pictures with tall narrow formats. If Greenberg’s remarks have not been republished in this country—and I do not think they have been—they ought to be. See also Olitski’s statement in the same catalog.

8a. The aluminum paintings of 1960 are an exception to this. Although not illusionistic, they can, I think, be said to hold as shape—chiefly by virtue of the fact that their supports depart from the rectangular only by a few shallow notches at the corners and sides. As a result the paintings are seen as restrained or held back by these notches by completing the rectangles they all but occupy. This gives the shapes of these paintings something to hold against—i.e., the pressure from within each painting toward the rectangle it almost is—and, in effect, makes the question of whether or not they make them­selves felt as shapes a real one.

8. My own feeling is that the (extremely limited) ability of liter­alist work in general to incite and sustain interest derives from its relation to the most advanced painting of the present and recent past—a relation which may be wholly unintended, wholly fortuitous. For example, what seems to me to give certain pieces by Judd such effectiveness as they have is the contrast between the laconic self-evidence with which they present them­selves in all their literalness as shapes and volumes and the way in which literal shape is minimized throughout Noland’s oeuvre and even subverted in some of his recent paintings. Similarly, Bell’s glass boxes hold my interest as long as they do because they can be seen as literalizations of the illusion of something like space or depth accessible to eyesight alone created by Olitski’s spray paintings. (Because the illusion is literally con­tained by the metal edges and glass walls it cannot come de­tached from these the way that, in Olitski’s paintings, the illusion sometimes parts company from the framing-edge. It is, so to speak, already detached, just as the edges of his glass boxes are already those of a container.) But I cannot believe in the quality of either Judd’s or Bell’s work.

Judd, almost certainly the foremost ideologist of the literalist position, has claimed—in “Specific Objects,” Arts Yearbook, No. 8, 1965—that “a work needs only to be interesting.” It is hard to know exactly what this means, because some work, such as Noland’s, Olitski’s and Stella’s paintings, is more than just interesting. It is, I want to say, good—more accurately, good painting. And in fact—despite the proliferation of work that is neither painting nor sculpture, and despite the pervasiveness of the facile notion (hawked recently by Susan Sontag) has the arts in our time are at last heading toward synthesis—what modernism has come increasingly to mean is that, more than ever, value or quality can persuasively be predicated of work that lies only within, not between, the individual arts. (Though it has also come to mean that that work must chal­lenge, in characteristic ways, what we are prepared to count as belonging more than trivially to the art in question.) The cir­cularity of this stale of affairs will be repugnant to many, and it is certainly harrowing, but I do not think that il is self­condemning. The crucial question, after all, is not so much whether anything artistically valuable lies outside the circle, as whether a meaningful concept of artistic value or a significant experience of it can reside anywhere but in its coils.

This, however, is not to say that the category of interest is irrelevant to art. I want to quote here at some length from Stanley Cavell’s essay “Music Discomposed,” both because of its obvious relevance to the present discussion and because the essay has not yet been published:

What is composition, what is it to compose? It seems all right to say, It is to make something, an object of a partic­ular sort. The question then is, What sort? One direction of reply would be, An object of art. And what we need to know is just what an object of art is. Suppose we give a minimal answer: It is an object in which human beings will or can take an interest, one which will or can absorb or involve them. But we can be absorbed by lots of things people make: toys, puzzles, riddles, scandals . . . Still, some­thing is said, because not everything people make is an object of this son. It is a problem, an artistic problem—an experimental problem, one could say—to discover what will have the capacity to absorb us the way art does. Could someone be interested and become absorbed in a pin, or a crumpled handkerchief? Suppose someone did. Shall we say, It’s a mailer of taste? We might dismiss him as mad (or suppose he is pretending), or, alternatively, ask ourselves what lie can possibly be seeing in it. That these are our alternatives is what I wish to emphasize. The situation de­mands an explanation, the way watching someone listening intently to Mozart, or working a puzzle, or, for that matter, watching a game of baseball, does not. The forced choice between the two responses—“He’s mad” (or pretending, or hypnotized, etc.) or else “What’s in it?”—are the im­perative choices we have when confronted with a new development in art . . . But objects of art not merely inter­est and absorb; they move us; we are not merely involved with them, but concerned with them, and care about them; we treat them in special ways, invest them with a value which normal people otherwise reserve only for other people—and with the same kind of scorn and outrage. They mean something to us, not just the way statements do, but the way people do . . .

My own impulse is to say that interest is basic to art—but not to either making or judging it. And if it is objected that what we ought to try to do is enjoy art rather than judge it, I would simply say that that may have been possible once but isn’t any more. This, however, is not to contrast enjoyment with judging—it is rather to insist that there is no real enjoyment, or no enjoyment of what is really there, apart from judging. One can still enjoy Olitski’s paintings simply as color, if one wants, but that is not to enjoy them, or be moved by them, or see them as paintings. And this means that there is an important sense in which one is not seeing them at all. But to experience painting as painting is inescapably to engage with the question of quality. This, too, is the work of modernism, and if one does not like it one ought to face the fact that what one does not like is painting.

9. Stella made four paintings in each of the eleven formats, There are, then, eleven sub-series within which not only the shape of the support but the configurations on the surface of the canvas are identical. Only the colors are different—but say this while someone is looking at two or more pictures in the same series and he will probably laugh, it is so inadequate to the experience of seeing the two at once. What seems nearer the truth of that experience is to say that the only things the two pictures have in common are the shapes of their supports and configurations of shapes on their surfaces—which, in front of the actual paintings, seems rather trivial.

10. I owe this distinction to Stanley Cavell, who has also made numerous other suggestions throughout the text. My indebtedness to Cavell is far deeper and more encompassing than this ac­knowledgment suggests, and indeed it would be impossible fully to acknowledge it. I also wish to record the fact that in writing this essay I have put to my own uses several remarks and observations made in my presence by Noland, Olitski and Stella themselves.

11. I take a reductionist conception of modernist painting to mean this: that painting roughly since Manet is seen as a kind of cognitive enterprise in which a certain quality (e.g., literal­ness), set of norms (e.g., flatness and the delimiting of flatness) or core of problems (e.g., how to acknowledge the literal char­acter of the support) is progressively revealed as constituting the essence of painting—and, by implication, of having done so all along. This seems to me gravely mistaken, not on the grounds that modernist painting is not a cognitive enterprise, but because it radically misconstrues the kind of cognitive enterprise modern­ist painting is. What the modernist painter can be said to discover in his work—what can be said to be revealed to him in it—is not the irreducible essence of all painting, but rather that which, at the present moment in painting’s history, is cap­able of convincing him that it can stand comparison with the painting of both the modernist and pre-modernist past whose quality seems to him beyond question. (In this sense one might say that modernist painting discovers the essence of all painting to be quality.) The object of his enterprise is therefore both knowledge and conviction—knowledge through, or better still, in, conviction. And this knowledge is simultaneously knowledge of painting (i.e., what it must be in order to elicit conviction) and of himself (i.e., what he finds himself convinced by)—apprehended not as two distinct entities, but in a single, in­extricable fruition. It should be clear that the conception of modernist painting which I have just adumbrated is not only anti-reductionist, but anti-positivist; in this respect I believe it has deep affinities with the persuasive account of the enterprise of science put forward by Thomas S. Kuhn in The Structure of Scientific Revolutions (Chicago, 1962). The further exploration of these affinities would, I am sure, prove rewarding. But a foot­note is not the best place to begin.

12. See "The New Sculpture,’’ Art and Culture, Boston, 1961, p.144. I also want to call attention to Greenberg’s essay on the sculpture of Anthony Caro in Arts Yearbook, No. 8, 1965. The relevance to Stella’s paintings of many of Greenberg’s observa­tions about Caro’s work seems to me striking.

13. See Thompson Clarke’s essay, “Seeing Surfaces and Physical Objects,” in Philosophy in America, edited by Max Black, London, 1966. The fact that eyesight touches only the surface of solid objects, and then only part of that surface, has traditionally played an important role in philosophical skepticism.