PRINT October 1973


On Quality in Art

Jacob Rosenberg, On Quality in Art (Princeton, N.J.: Bollingen Series, Princeton University Press, 1967).


The problem is that I am, in a sense, speaking to Jacob Rosenberg, but if he is speaking to me, he doesn’t know it, or he didn’t know it until now. And of course there is a certain unfairness to this sort of dialogue, all of it in my favor; for in a sense, I choose what he says, though my choice of what I may have him say is restricted to the 232 pages of his text (plus his introduction).

1. “Artistic value” or “quality” in a work of art is not merely a matter of personal opinion but to a high degree also a matter of common agreement among artistically sensitive and trained observers and to a high degree objectively traceable.

Our value judgment is a composite of “subjective” and “objective” elements. “Subjective” I call those which are the purely personal responses of an individual; “objective,” those which are agreed upon by trained observers and therefore meet with general acceptance. The composite nature of value judgment leaves room for the strengthening of the objective aspect, thus allowing us to attain a measure of validity. This, then, is what we are striving for in our investigation (p. xxiv).

Rosenberg has laid out the “consensus of experts” argument that became familiar on the letters page of Artforum last spring, with some additional complications. There are several questions to be asked which may sound picky, and involve certain philosophical problems, but which are nevertheless serious for such a line of argument. Rosenberg acknowledges in his introduction that these questions lead to philosophy or esthetics, but promises not to linger long in those areas, and he doesn’t. Within the notion of a consensus of experts or the common agreement of trained observers is the presupposition that a significant percentage of “trained observers” agree that a certain artwork is of high quality. But how do we know that they still agree, and how do we account for the changing of opinion within a specific consensus group, and more important, for the change in value judgments from one consensus group to the next over a span of generations?

As Rosenberg points out, by the common agreement of trained observers from the Renaissance until the late 19th century, medieval art was considered of little artistic value; now there is common agreement by our current crop of trained observers that much of medieval art is of high quality. Within this span of time, a lot of trained observers lived and died who agreed that medieval art was of little artistic value, and within the terms of Rosenberg’s argument, this critical judgment, by so much agreement, was objective and, therefore, valid. But it isn’t anymore, at least not according to the theory. Now the critical judgment, supposedly objective because commonly agreed upon, is just the reverse. Do we really mean by “objective” that which can change at a subjective whim? Essential to this notion that consensus of judgment equals objective judgment is the equation: 1 subjective response or judgment + 1 subjective response or judgment = 1 objective response or judgment; while this seems like 1 apple + 1 apple = 1 peach, the word “objective” could be so defined within the scope of a discussion. But there is the great danger, inevitably realized, of confusion and error of the word’s maintaining its more conventional connotations and reverting to a more conventional meaning midway through the argument.

Rosenberg clearly desires the strength added to his argument by the conventional connotations of “objective,” meaning beyond the subjective, as when we speak of physical facts. He does not use the word simply as a quicker way of saying “the common agreement of trained observers,” but to put these critical judgments on the same footing as statements of fact. In these terms, to say “this painting has quality” is the same kind of sentence as “this piece of wood with paint on it exists.” But when Rosenberg shows us a drawing, or two drawings for comparison, he doesn’t spend a page and a half of writing convincing us of the drawing’s existence, while on the other hand, he clearly feels he needs a page or so to convince us of its quality. Rosenberg and Roger de Piles agree that Giovanni Bellini made paintings, and that the paintings look a certain way, but they do not agree on whether the paintings have quality. The point is that if a consensus of critical judgment constitutes an objective judgment, “objective” does not mean in the same way as when we speak of a set of physical facts or state of affairs. This is not to say that speaking about a set of physical facts is other than subjective, but merely to point to the distinction between speaking of a set of physical facts and ascribing quality. No one reading this sentence will disagree that there are certain letters and words in it arranged in a certain order, but there may be general disagreement on what the sentence says, whether what it says is correct, and whether what the sentence achieves is good.

2. Finally, as for Vasari’s occasional glimpses beyond the standards of his own time—one is almost tempted to call this his “free” judgment (we noticed it in Vasari’s reaction to Donatello and to the late Titian, and there are other instances)—it proves that his intuitive feeling for quality could not be fully suppressed by the prevailing theories. This occasional tension between the rationalized and the standardized judgment, on the one hand, and the merely intuitive on the other, is an interesting phenomena which we shall also notice in the work of the gifted critics of later times. It speaks, I believe, for Vasari’s innate power of discrimination, and is not the least reason why the record of the validity of his value judgment is a very impressive one (p. 28).

In my original notes taken while reading On Quality in Art, this passage, and a few like it which preceded, left me a bit stunned; as I continued to encounter similar passages throughout the remaining 200 pages, I got used to them. There is a circularity in Rosenberg’s argument that is so blatant at times it is almost difficult to get hold of. But this passage is not uncommon. What Rosenberg is saying here is that Vasari’s ascribing high quality to late Titian proves his intuitive “feeling for quality,” a feeling or capacity in Vasari so strong as to override even his theoretical position and his prejudices. Beyond the difficulty of determining what an “intuitive feeling for quality” is, there are two problems here: the most important, the notion that by recognizing quality in Titian Vasari proves his capacity to recognize quality, as if Titian were born with the stamp of “Great Master” on his ass, and anyone who can identify quality in a Titian painting can therefore identify quality—on the condition he or she has never peeked. This first problem takes the form: If you can recognize quality in Titian, then you can recognize quality; Vasari recognized quality in Titian, therefore, Vasari can recognize quality. The second problem is very close to the first, but the emphasis is on the use of the word “proves.” Surely nothing has been proven, and certainly proof or proving is a more rigorous operation than what we have here. The quarrel with “proves” could be what people like to call semantic quibbling, but the word carries a similar persuasive, connotational overload to “objective,” and its strength in that direction certainly partly motivated its use.

“The record of the validity of his value judgment is a very impressive one,” beyond the question of what “valid” or “validity” could mean in the context of value judgments, reminds me of the daytime TV quiz shows which ask competing contestants questions of the type: It is better to spank a child for disobedience than to tweak the child’s ear, true or false? One contestant is said to answer more of these questions correctly, compiles a better record than the opponent, and wins a vacation for two to Hawaii. But there is something odd here. Questions of this type have no correct answers in the usual sense, usual in terms of questions asking about the state of facts. A correct answer in this case, is correct only insofar as it agrees with the answer of the quiz show’s resident expert; thus, one is not striving to answer correctly, for correctness is simply not relevant; one is striving rather to guess how the resident expert answered. This is exactly what the “impressiveness” of Vasari’s “record” amounts to, and, in this case, the resident expert is Jacob Rosenberg. Vasari was very lucky with his Titian answer, for unlike most quiz show contestants, he didn’t know whose answers he was going to have to match. It adds up very nicely: Vasari has a special feeling for quality (“gifted critics”) and sees quality in Titian; Rosenberg sees quality in Titian also, Rosenberg also has a feeling for quality.

3. From our present day point of view, de Piles’s quality judgment, as expressed in this list, is certainly disappointing after what we have read of his broadening of the classicist approach, his method of analysis and appreciation, and his great effort to reach a higher level of objectivity. We are disappointed that he rates Michaelangelo (37) much lower than Andrea Del Sarto (45) or Giulio Romano (49); and that Giorgione (39) falls far below Titian (511 and Tintoretto (49). We cannot understand why Dürer receives a grade of only 36, when a second rate mannerist like Taddeo Zuccaro gets a total of 46. Our admiration for the relatively high rating of Rembrandt (50) cools when we see Teniers follow fairly closely (with 46). We shake our heads at the curious underestimation of Caravaggio (28) when the Carracci and the Bolognese in general get high marks (58, etc.) And we are by no means happy about Giovanni Bell ini’s very low grade of 24. In this case, it is certainly small consolation that Lucas Van Leyden suffers the same fate. On the other hand, why does Charles Lebrun (56), this driest of the academicians, rank higher than Poussin (53), Leonardo (49), Correggio (53), Titian (51), and Rembrandt (50)? Is there any sense to de Piles’s criteria when they lead to such failures in his value judgment, at least as seen from our own time and point of view, which is backed by the accumulated experience of three more centuries? (p. 46).

No, there is no sense to de Piles’ criteria, but not because they lead to failures in value judgment, for the terms “success,” “failure,” “valid,” “invalid,” “true,” “false,” have no meaning or application with reference to value judgments. If de Piles’ criteria do not make sense, it is because no criteria for quality or excellence in art ultimately make any sense. We can say a work of art should be a certain way, but. we cannot support why an artwork should be that way, except to say we like works that are that way. But even if the notion of the possibility of quality and meaningful evaluation were accepted, and the notion of possible success or failure of a given evaluation accepted as well, where is there any evidence of a failure in de Piles’ value judgment? De Piles’ value judgments failed only insofar as they failed to coincide with those of Rosenberg, and other of de Piles’ judgments have succeeded on precisely the same terms—by coinciding with those of Rosenberg. But what sort of “accumulated experience of three more centuries” does Rosenberg mean that can somehow “back” current value judgments, sounding much like an insurance company backed by 75 years of experience? Clearly there is an accumulation of information, research, and information-gathering methodologies, which adds to our knowledge of the nature and sequence of historical events, attributions, etc., but there is no sort of accumulated experience of value judgments. Each individual has an evolving set of values of some sort, and each culture has, similarly, a set of evolving values. The individual takes on the values of the culture, and the culture’s values are an aggregate of the values of constituent individuals. However, values in an individual or a culture do not accumulate in the way individuals and cultures accumulate information; when certain values are added, other values are dropped, often of the necessity of new and old being mutually exclusive.

If de Piles’ numerical rating system seems a bit silly, so does Rosenberg’s agitation over his favorites not being picked, like a regular in the bleachers at the stadium who knows Munson deserved to be the All-Star catcher and finds little consolation in the fact that Matty Alou didn’t make it either.

4. Whatever our final judgment of de Piles’s critical ability, there are many good things in his writings which we should take to heart when we ourselves make the effort to work out valid criteria of excellence. These are: his insistence that we derive our standards from the study of as many great masters as possible, that we evaluate by comparison, that we keep our minds critical and independent, that we form our judgments from originals only. (p. 47).

The problem here is blatant: how can we derive our standards for evaluation “from the study of as many great masters as possible” when who are the great masters can only be determined by evaluation on the basis of the standards we are supposed to be deriving? Even if we overlook this circularity, which is something that is not reasonably overlooked as it is basic to the entire problem, the question of which are the “great masters” remains. De Piles would have us study Romano, Lebrun, and the Carracci, which, for Rosenberg, would inevitably cause a certain warp in the standards we derived from those masters. But undoubtedly, Rosenberg’s recommendation would be that we study those masters so determined by current conventional wisdom, which is to say, those agreed upon by today’s trained observers. Must there be a unanimity of all current trained observers, and if not, what percentage of trained observers agreeing on an artist is sufficient to determine a great master worthy of study? Further, the history of connoisseurship, even as Rosenberg presents it, is but a history of changing opinion and attitude, which not only throttles the notion of “objective critical judgment,” but should leave us with some skepticism about taking the word of current trained observers and studying the masters they recommend only to have next year’s batch of trained observers telling us we were wasting our time; and, of course, next year’s batch will be backed by another year of “accumulated experience.” Once we have somehow arrived at who are the great masters, the standards that can possibly be derived from them cannot in any sense be meaningful as standards for anything; at worst, the standards derived will be a specific description of the master’s work, at best, a generalized description. In any case, those “standards” will amount to a list of adjectives, not unlike the list Rosenberg arrives at toward the end of the book, and most problematic, those adjectives by being necessarily generalized will also be necessarily vague. The end product is not so much a set of standards as a vague description of what an artwork should be in terms of what artworks were, which is to say, artworks should be a certain way because that’s the way they used to be.

5. As for a systematic approach, there remains one big question: Do we not move in a vicious circle in assuming that a great master is a great master because his fame is tested by the approval of the centuries? We certainly ought to test and to experience again this value judgment. But I think it does no harm if, in our choice of examples, we go first on the generally accepted assumption that Rembrandt and Raphael, Rubens and Watteau, Cézanne and Manet are great masters, and then prove it again by our analysis and comparison. In this way the circulus vitiosus will be broken (p. 126).

How is the vicious circle broken when the standards by which we are to reassess the generally accepted great masters were derived from the study of those masters in the first place. If the standards derived from the study of the generally accepted great masters consist of a generalized description of their works, (see the criteria Rosenberg arrives at in his conclusions), then it is inevitable that these works should conform to the description on reassessment. In this way their quality is “proved again.” When the standards are at least in part derived from Manet, and are then reapplied to Manet in order to retest his greatness,naturally Manet will be seen to match the standards and be proven a great master again. Vicious circles are not so easily broken.

6. In Schongauer’s work the line is finer, more sensitive, and more graded in its accents [than the work of his pupil, with whom Rosenberg compares Schongauer]. It models with a more sure feeling for roundness, for texture, for organic structure. . .Schongauer’s line is also marked by a calligraphic quality, by an extraordinary graphic charm. A small detail like the pigtail of the oriental makes us aware of that (p. 131).

Rosenberg, probably unconsciously, uses a technique quite common to evaluative comparisons and subtly persuasive. He describes the “master’s” work (for whom he is rooting to come out best) in rich, glowing, sweeping language while withholding such language from the opponent, though in most cases, the terms apply equally well to either drawing. But there is more here. To say one drawing is better for sensitive calligraphic lines, a feeling for roundness, for texture, for organic structure, and graded accents, is only to say one prefers drawings which have those attributes (if it can be determined that the drawing in question has them, and if it can be specified, even, what sorts of attributes these words denote) to drawings which do not have these attributes, and consequently, have other attributes. What this comes to is a hierarchy of attributes without any sort of justification for placing more value on one than the other, beyond the fact that an individual, Rosenberg in this case, prefers one set of attributes to another. The point is, in ascribing a characteristic to a work, there is nothing to support why works should partake of that characteristic, however implicit the fact that one or more individuals prefer works that do. If four individuals or five billion individuals over a span of three centuries prefer works with that characteristic, they remain individuals with necessarily subjective responses and preferences. There is no argument with Rosenberg’s having certain preferences, but there is an argument with his attempt to project those preferences beyond himself by saying implicitly that these preferences are objectively valid: “my preferences are the ones that count for everyone.” The problem is not that I resent others telling me what I should prefer, but that there is simply no support for their doing so.

7. Therefore, to sum up the results of these two chapters, we suggest the following list of criteria of high quality as being more or less valid for the whole field of master drawings from the fifteenth through the nineteenth century:

line and tone: sensitivity, articulateness, flexibility, ease, surety, spontaneity, rhythmical quality, suggestiveness.

form: solidity, organic character, coherence, clear distinction of planes.

space: an articulate, continuous, comprehensive treatment (“comprehensive” is a Baroque addition) composition or perhaps formal organization: integration (unification), gradation, density, sense of balance, and. richness of formal relationships.

Finally the following more general characteristics, which cannot be properly confined to any one of the above categories: range of accents, artistic economy, selectiveness, feeling for the media, sense for the significant, consistency, vitality, originality, inventiveness, clarity, expressiveness (pp. 203, 204).

Like a statement made with great confidence early on in a Greek tragedy, it seemed inevitable that Rosenberg’s statement on page 124, “We should, in general, avoid judgment based on too narrow a set of principles or on vaguely or ill-defined criteria,” would return in some ironic form to discredit the author’s accomplishment. “Too narrow a set of principles” seems to be no problem for Rosenberg’s criteria, but “vaguely or ill-defined” is a problem. And it’s a big one. To be sure, he uses the words that form his criteria frequently in the body of his text, especially in the comparative evaluations of drawings. Often, frequent use of terms will tend to clarify how an author means them, but if that is the case here, I’m afraid someone will have to show me. After puzzling over his tossing out of the terms throughout the book, always in the direction of his favorites, I simply cannot grasp any clear meaning or rules of application for these luxurious adjectives, which now become criteria for high quality. As stated earlier (#5), Rosenberg has formed his criteria for high quality from a list of adjectives he likes to apply to the works of his favorite artists. The list of criteria is but a list of vague, ill-defined adjectives which can be applied to just about anything at all. To hislist we may as well add Beauty, Truth, and Goodness. How do we recognize sensitivity, ease, organic character, richness of formal relationships, etc.? The simplest thing would be to remove connoisseurish disguises, and lay bare exactly what the theory says: The criteria for high quality is high quality. But even if we accept the theory totally, what have we got? We can now say with assurance (as we could have before if we cared to) that those works we prefer are “good,” and now we can even show that those works we prefer are good because they meet the criteria; is there any work that cannot be described in the terms of this list? More importantly, what has any of this to do with the artworks themselves, our experience of them, and whatever pleasure we clerive from them? Do we go to galleries and museums, read art magazines, think, to enable us to have an opportunity to pass judgment? The point is precisely, what’s the point?

8. The difficulty of making valid quality distinctions greatly increases when we come to the art of the present. There are good reasons for this. We lack the perspective and we do not have the benefit of time-tested judgments (p. 207).

All that “time-tested judgments” means is that we still agree with something someone thought a long time ago. The supposition here is: If people think something for a long time, it must be true. It should be fairly clear that a belief is not magically transformed into Truth with the passage of a specified duration of time. That medieval art was for four centuries considered of little or no artistic value, within the terms of thethesis, qualifies that consideration as a time-tested judgment. Yet, despite that judgment’s standing the test of time for four centuries, our current crop of trained observers no longer agrees with it. Of how much benefit, then, was the time-testedness of that judgment? By the necessary infiniteness of time, any judgment, concept, or theory which is considered as time-tested today is highly vulnerable to being considered as not having passed the test of time tomorrow, which is to say, vulnerable to being discarded as nonsense. The one judgment, more accurately a prediction, that seems properly considered as having passed a test of time came ironically from Vasari. Vasari wrote that the leaning tower of Pisa would never fall, and the tower is still up and leaning 400 years later. The bell tower may someday bite the dust, but 400 years is a pretty good “never.” For most of us, what we thought we would never do, or never would happen, occurred within only a few years or sooner.

If there is frustration that we can “not at least approach an absolute scale of quality,” it is because the absolute brooks neither approach nor any concept of proximity. All critical evaluations are necessarily subjective, and there are no conditions under which these subjective responses can mysteriously sneak into the realm of the objective or the absolute. Thus the concept of quality in art is incoherent; it is the product of confusion and hope. With the concept of quality go concepts of master and masterwork. What we are left with are our preferences, interests, thoughts—in short, with the possibility of experience. If this logic brings a certain amount of insecurity, there is no help for it.

Bruce Boice