The Core Question of Indiscernibles: The Presence of Goodman's Theory of Representation in Arthur Danto's The Transfiguration of the Commonplace
Walter Menon Junior

Brillo boxes or Brillo boxes?

As we know, the problem of indiscernibles was suggested to Danto as a problem of philosophical relevance by Andy Warhol’s exhibition at the Stable Gallery in New York in 1964. With the help of Gerard Malanga and Billy Linich, Plywood boxes identical to cardboard boxes of supermarket products were painted and printed with the silkscreen logos of the brands Kellogg’s, Brillo, Mott juices, Del Monte peaches, Heinz ketchup. The boxes were stacked in the gallery. As it is very well known, in an interview with a Canadian TV network at the opening of the exhibition, in which he was asked about the originality of these boxes, assuming that they were sculptures, Warhol replied that they were, in fact, not original sculptures.1 Although the exhibition and the boxes were designed by Warhol and created by him with his assistants within what could be understood as a process, or set of processes and procedures for the execution of sculptures, Warhol’s response points the consequences of its exhibition: pushing even further the limits of the definition of a work of art, already frayed by Duchamp and part of the artistic production of the 20th century.

It is not relevant to the scope of this text to know to what extent this was Warhol’s main purpose with his facsimile boxes. That is, whether he had intended to provoke a question about the conceptual limits that define a work of art or whether, based on the observation that these limits were already in constant crisis since the beginning of the 20th century, a broader horizon of possibilities left space for the kind of interest Warhol demonstrated in objects and icons of American mass culture. The issue concerns the limits between objects that are not works of art and identical objects that are. What criteria did Warhol use to determine this difference? A plausible answer could be the absence of such criteria, at least as they were conceived in 20th-century artistic production. Perhaps for Warhol, the aesthetic interest aroused by the packaging of mass consumer products was no different from that produced by a Picasso painting or any other renowned work of art. From then on, anything could be a work of art if it aroused such interest. However, although boxes of soap on supermarket shelves may suggest some aesthetic interest, even similar to what a Picasso painting arouses, this does not eliminate the question of what is the reason for the boxes of Brillo steel wool in the supermarket not being objects of art, unlike those created by Warhol.

I will not discuss Danto’s 1964 essay “The Artworld”, but just focus on his simple explanation of the difference between Brillo’s steel wool boxes and Warhol’s Brillo Box, and why the former is not a work of art, but the latter is. Both objects have identical visual properties; that is, they are visually indiscernible. Therefore, visual properties alone are neither necessary nor sufficient to define a work of art. In other words, works of art do not have specific visual properties that distinguish them from other things that are not works of art. Furthermore, knowing that Warhol’s Brillo Boxes are not Brillo steel wool packaging is not sufficient to make them works of art. It is possible that other facsimiles of Brillo steel wool boxes, like Warhol’s, could also be toys or decorative objects. Thus, neither aesthetic nor perceptual properties, nor just declaring an object a work of art could be sufficient criteria to define what art is. Therefore, a work of art must have another set of defining properties. According to Danto, these properties arise in the special use of the verb “to be” to designate works of art in what he calls the World of Art. Therefore, we can say that this use depends on its relevance within the world of art, that is, whether it makes sense to call such or such an object a work of art, taking the world of art as contextual support. Basically, Danto says if something x could be named a “work of art,” it is because x represents a work of art, meaning it is a case of.

The problem of indiscernibles points to the fact that works of art do not have intrinsic aesthetic properties that define them, nor does knowing how to “recognize” a work of art imply having good descriptions of works of art. Warhol’s operation reveals the fragility of some of the notions that, until then, seemed to constitute what is unquestionably understood as a good description of a work of art, such as originality, authorship, creativity, talent, meaningful form, and so on. The interest sparked by the Brillo Box evidently results from its consequences for the definition of art. However, these consequences are not limited to artistic production or art criticism. By forcing the limits between what is art and what is not, the question of indiscernibles appears to be epistemological and ontological, insofar as it implies notions of definition, intentionality, identity relationships, similarity, and, notably, the notion of representation, which is central in the history of art and philosophy. In this sense, the problem of indiscernibles, as thought of in philosophy of art, becomes significant for philosophical discussions beyond art criticism and theory.

Throughout his 1981 masterwork, The Transfiguration of the Commonplace: A Philosophy of Art, Danto creates a series of thought experiments in which he elaborates on possible variations of the same problem to answer why, amongst visually identical objects, some are art and others are not. In seeking to explain the problem of indiscernibles in relation to works of art, Danto is simultaneously developing a theory of art as representation. His examples are not limited to the world of visual arts, as can be seen in the case of Menard’s Quixote. Danto’s entire effort consists in demonstrating that similarities between entities, in general, are not limited to sensory properties, nor is similarity a criterion for representation. The issue is knowing what two similar entities represent when they mean different things, more specifically, when one of those things is a work of art. In my understanding, what lies behind this perspective is Danto’s acceptance, not without restrictions, of Nelson Goodman’s thesis that representing is denoting and the attempt to understand how this idea of representation works in relation to arts in general. In other words, on the one hand, we have the idea that representing is to denote or name. On the other hand, Danto refines the meaning of denoting and naming in the context of visual representations. Therefore, his dialogue with Goodman’s theory of representation seeks to understand what type of representation characterizes works of art. Despite partially agreeing with his theory, Danto rejects Goodman’s conventionalist view of art’s nature.

Representation as transfiguration

To answer the question about indiscernibles, Danto creates a series of thought experiments in the Transfiguration of Commonplace. Although they are not the starting point nor the only “cases” addressed by Danto throughout the book, I will briefly describe both the Don Quixote experiment by Pierre Menard and the Picasso blue tie experiment. The conclusions in both cases illustrate how his theses are closer to Goodman’s theory on representation.

In Borges’ short story, the character Pierre Menard writes an original literary work identical to Cervantes’ novel Don Quixote, sharing the same properties. However, a copy of Cervantes’s novel differs from a copy of Menard’s because they are not copies of the same work. What makes them examples of different works if they are indistinguishable in every way? Danto suggests a second situation in which there are three identical blue ties: one made by Picasso, another by a child, and the third by a forger, but signed by Picasso. The rule of identity elaborated by Leibniz and evoked by Danto applies to both situations, the one that illustrates Menard’s work and the one with ties: for each property F, a is identical to b, and when a is F, b is also the same. It is.2 In the case of Menard’s and Cervantes’ Quixotes, there is a clear disobedience of this rule. Although the works share all visible properties, they are not identical. The error lies in taking perceptible properties as properties that define identity. In fact, from this angle, the works are similar, and the difference between them would be in the order of the absurdity created by Borges’ story. But this is not the meaning that Danto sees in Borges’ words. According to Danto, the writer points to a set of properties that make a plausible difference between the two works. Firstly, Borges states that Menard’s Quixote is infinitely more subtle than Cervantes’, while Cervantes’ is immeasurably cruder than his counterpart. Cervantes’ Quixote reveals the poor reality of his country in contrast to the chivalric fiction of the 17th century. Menard speaks of “the land of Carmen during the century of Lepanto and Lope Vega”, using the exact same words as Cervantes. They describe the same place and time, but the way of referring to them belongs to different historical moments.3

The referent is the same in both novels, but not the meaning. Cervantes could not refer to Spain as the “land of Carmen,” a character in Bizet’s story written in the 19th century, and, according to Danto, “the poor provincial reality of his country” is something that does not apply to Menard’s novel, since the designated country is Spain and Menard was French. Menard’s novel is historical. Cervantes’ could not have this meaning, considering that the story described in his Don Quixote is contemporary to him. Borges observes that Menard’s Spanish style is “archaizing”, it suffers from “affectation”, attempting to mimic 16th-century Spanish, while Cervantes’ style is “without inhibitions”, naturally handling the Spanish of his time.

Danto concludes that Borges’ contribution to the ontology of art is extraordinary, as it demonstrates that it is impossible to isolate factors that permeate the essence of the work. Despite their graphic congruences, these works are profoundly different.4 The precedence of Cervantes’ work is necessary for understanding Menard’s work and for it to be, therefore, a work of art. Menard’s objective, the meaning of his work, was not simply to copy Cervantes’ Quixote but to recreate Quixote “word for word.” A copy of Cervantes’ Don Quixote is a copy of the same. One only needs a “xerox machine” to copy Cervantes’ Don Quixote or its PDF file. Menard’s work also cannot be considered a quote from Cervantes’ work because it does not refer to Cervantes’ book but rather, as Borges states, to the “land of Carmen during the century of Lepanto and Lope Vega”. Danto insists that Menard does not imitate Cervantes since his original work is a true Don Quixote. The emphasis, Danto points out, must be the connection between the identity of a work and its time, place, and provenance, even more so as neither Menard’s style nor theme can be identified if we abstract completely from history.5 Although the two are perfectly identical, these factors and their histories must be considered when examining the difference between any object and a work of art.

In the case of blue ties, Danto creates a fiction to show how much more complex the problem of indiscernibles is when it comes to a context in which physically indistinguishable objects, with different origins and ends, are confused in a game of exchange of references. The context emerges from a trend within North American artistic production of the sixties to appropriate men’s clothing items, such as the tie. If Danto is to be believed, the first artist to use this object in a work of art was Jim Dine in an etching called The Universal Tie, followed by countless other artists, such as Claes Oldenburg and John Duff, culminating in a 1975 exhibition “entirely devoted to ties in a Madison Avenue gallery”.6 This proliferation of “artwork ties” suggested to Danto an imaginary situation akin to a comedy of errors, which serves as a thought experiment to explore the problem of indiscernibles. Firstly, it is worth highlighting that the object which one disputes the status of a work of art is such that has its meaning derived from its use, better said from its function, which is eminently aesthetic. These are three examples of a piece typically identified as men’s clothing: the tie. From then on, a plot of quirks unfolds in which three instances of ties, due to interference with each one and a coincidence between the results of these interferences, become indistinguishable artifacts. In a circumstance imagined by Danto, Picasso would have painted, shortly before his death, one of his old, disused ties with light blue paint. He painted it in such a way that the brushstrokes were, as far as possible, invisible, that is, in such a way that the surface of the tie was painted uniformly and without leaving brush marks.

The “uniform touch of the brush” must, according to Danto, be seen as part of the content of the work. This approach suggests a repudiation of the physicality of painting (la peinture) or that apotheosis of paint and brushwork that characterized New York painting movement of the 1950s.7 This type of operation carried out by Picasso would not be new, considering that his work contains, among other everyday objects, a bicycle handlebar that imitates a bull’s head, a fish, etc. Picasso’s tie, La cravate however, would be more radical, as it breaks the limit of allusion that characterizes its antecedents:, when displayed, it does not allude to the tie or any other object. Suppose, continues Danto, that a child takes his father’s tie, identical to Picasso’s, and paints it light blue, using the same brand of paint as Picasso and intending to do it uniformly, without leaving brushstroke marks. Right away, we know this second painted tie is not a work of art. Something prevents it from joining the confederation of authorized works of art in which Picasso’s tie is easily accepted, although without great enthusiasm.8 Note that this is not institutional for Danto, and does not arise from the fact that the child is not considered an artist; Picasso’s blue tie is a work of art by what Danto called the world of art. Further, Danto introduces a forger who copies Picasso’s work, that is, who produces a tie like Picasso’s and therefore similar to the one made by the child.

As we saw with Goodman, similarity relations are transitive: if a is similar to b and b is similar to c, then a must be similar to c. However, the issue here, as in Menard’s Quixote, is knowing what each of these ties represents, which is different from knowing the intention behind them. Picasso could answer a question about the meaning of La cravate by saying that it “means nothing”; when asked why he painted the blue tie, the child may respond that it was an impulse and that his gesture “doesn’t mean anything”. Similar answers are different by holding different meanings. In other words, the similarity, whether visual or descriptive, “doesn’t mean anything”; it doesn’t count when deciding which of the ties is a work of art.

When Picasso created La cravate, he intended it to have no meaning. In contrast, the child‘s blue-painted tie’s “saying nothing” carries another connotation: the child simply wanted to do “a cool thing”, with no other purpose than it being just a joke. Rather than thinking about the difference between the two ties in terms of intentionality (i.e., that Picasso’s intention would be the reason for his Cravate to be a work of art), Danto argues that the difference is found in the meaning of each tie and that these meanings have a history. The relevance of the intention is conditioned by the historical context in which the work was produced. The case of the forged Picasso is exemplary in this sense. To what extent does knowing the forger’s intention count for the meaning and value of the fake La cravate? Both have the same name and appearance; however, one was made from the original to be similar to it in every way – to blend with it. In other words, there is a historical precedence of one over the other, and the original is the necessary condition for its counterfeiting. If the forger’s intention counts, it only reinforces the evidence of the required relationship between the two. The falsified La cravate is not a citation of Picasso’s work because it does not follow citation rules, nor is it an original work, a “rereading” of Picasso’s creation. Following Danto’s argument, the fake tie’s “saying nothing” does not refer to the “saying nothing” that constitutes the meaning of Picasso’s work, which itself has meaning within the context of the history of painting. False means “to be La cravate”; certain types of images, such as mirror images, lack autonomous existence. Their meaning is directly derived from Picasso’s work.

It is possible for specialists, in some instances, to verify the difference between a work and its copy by closely examining perceptual aspects of the work, such as brushstrokes and other “clues.” However, irrefutable proof of their difference is provided by the history of the two works – their origin, time, and context of production. To insist on this point, Danto introduces a series of identity confusions to further reinforce his argument that the properties that differentiate the works are not those of “similarity.” The three ties are “comically interchanged” and confused among the dealers.

The sequence in which the confusions occur is essential because it raises the question of authenticity conditioned by authorship. The tie made by the child is on display “at the Palais de Beaux-Arts, in Luxembourg, protected by a large insurance”. Picasso disputes its authenticity by stating that the tie displayed at the Palais de Beaux-Arts is not his Cravate and therefore refuses to sign it. Instead, he signs the fake La cravate. The original La cravate, crafted by Picasso, ends up in the Defraud Department’s warehouse, where it “lies forgotten” with other forgeries. The problem is that the artist’s visual acuity and signature – in this case, Picasso’s – do not alone bestow authenticity upon the object. The artist’s signature on a work could result from a mistaken belief on his part. Thus, to give value to a piece, even the signature must be integral to it; that is, it must be part of the work as completed by the artist. Here, again, the condition is historical. Picasso’s signature on a copy of La cravate does not render it the original La cravate. Once the mistake was discovered, value would reside not in the object but in the authentic signature. Supposing that Picasso had signed his work and did not recognize his signature years later, there would still be ways to prove its authenticity. Picasso is by no means the authority on Picasso’s signatures.9

Necessary to this discussion, is Danto‘s refutation of the argument that we can know how to differentiate what is art from what is not through an acquired competence in recognizing specific patterns and certain types of characteristics of works of art. To illustrate this, we will briefly describe the philosopher Willian Kennick‘s thought experience, which Danto analyses. He suggests that in Kennick’s text “Does Traditional Aesthetics Rest on a Mistake?”, the problem of definition may simply be a matter of cognitive skills.10 Kennick envisions a scenario where, in a warehouse full of the most varied objects, some unknown to us are works of art. He posits that we would be able to distinguish among them which objects are art without having any satisfactory definition of a work of art. The critical question is: what kind of competence allows us to visually recognize works of art without having a definition of what a work of art is? Danto answers that this competence does not exist. To show that Kennick’s experiment is mainly wrong, he proposes the existence of a second warehouse, where everything that was a work of art in the first warehouse is not so in the second, and vice versa. A bicycle in the first warehouse is just a bicycle. In the second, it is a work of Pop art, for example.

The inversion proposed by Danto undermines any analysis of the concept of art that presupposes the absolute relevance of the capacity for recognition. Since this capacity is acquired through a process that we can broadly consider inductive, by simple enumeration or repetition of uses of terms in given circumstances, generally in the face of physical properties considered relevant to classify works of art, by people considered experts, for example, it seems plausible to affirm that the capacity for recognition does not systematically eliminate all cases of works of art. “Although many artists and periods in the history of art have clear recognition characteristics, there is nothing to prevent, for example, an artist from having works that are very different from those that are recognizably recognized as his”.11

Again, the problem is that more perceptual properties are needed to define or learn to recognize the difference between a work of art and a similar object that is not art. Following Danto, we can finally formulate this fundamental clause as follows: if a is not identical to b, there must exist a property P such that a is P, but b is not P. And P is not necessarily a perceptual property. What differentiates between a work of art, a forgery, a child’s play, or any object of everyday use are not perceptual predicates. If no visible properties differentiate works of art from non-art objects, we must admit thatno description of sensory, perceptual experiences is sufficient and necessary to define works of art per se.

Differentiating art from non-art presupposes knowledge of a certain non-perceptual property P that distinguishes a work of art from a non- art object. This type of difference is intrinsically conditioned by how we relate to certain objects. The way of being of certain objects determines how we relate to them. According to Danto’s historical argument, to understand the difference between art and non-art, how we relate non-art to art objects, implies knowing a specific conceptual model that founded this relation in the West: the imitation paradigm. Understanding art history basically means understanding the relationship between copy and original works; our engagement with this type of relation is a way to understand what art is.

Let us return to the first chapter of The Transfiguration of the Commonplace. Here, Danto reviews what he believes to be some of the fundamental episodes in the history of the idea of imitation in philosophy, especially concerning art. The primary source is book X of Plato‘s Republic, where Socrates reduces imitation to the metaphor of the mirror. Against this perspective, Danto highlights Nietzsche’s theory on the origins of Greek theater. It contains a historical explanation of the origin of representation, or the idea of representation, which would be simultaneous with the conception of Greek theater. According to Danto, Nietzsche states in The Birth of Tragedy that Greek theater originated in the cults of the god Dionysus.12 Gradually, the choir distinguishes itself from other officiants, who later divide into actors – those who represent – and the audience – those for whom the representation is directed. Thus, there would be two regimes of representation. One that takes place before the emergence of theater, inherent to the sacred space of the Dionysian ritual. At this moment, the idea of representation as a representation “of something” had not yet formed because there is neither an object – “something” to be represented – nor spectators – someone “to whom” the representation is directed, external to the ritual. That means everyone is involved in a trance induced by music, wine, and dance, emulating the deity’s characteristics; thereby, Dionysus becomes present and “presents” himself.

A consequence that Danto draws from Nietzsche’s ideas is that with the establishment of representation by designation, the dual concept of reality and fiction is also established. The concept of real only exists because there is a concept of illusion or unreal. The case of visual representation illustrates well Danto’s intention in conceiving the fields of real and fictional as structurally linked and established by the second type of representation. However, he argues that this second sense of representation kept the first in its wake: presentation. In Western tradition, the image of something is considered fictional. This “false” entity represents something “real”, whatever real means. However the unreal thing shares common properties with the real thing: visual properties, for example. Danto’s interest in this theory of imitation is to develop the hypothesis that the divides between representation and reality, art and life, or better yet, between a work of art and an everyday object, are of the same kind. Despite Western artistic representation increasingly incorporating the “real” from a certain point in its history, an ontological separation persists between representation and represented, that cannot be overcome. We can perceive a similar distinction between a Brillo box – a packaging of this steel wool brand – and Warhol’s Brillo Box. Such separation is not due to the false idea that the Brillo Box is a “representation” of a Brillo packaging.

The idea that something of the represented is preserved within representation stems from a confusion that occurred during the development of mimetic art theory and mimetic art itself: the first sense of representation as non-denoting presentation became confused with the second sense of representation, which is a dual concept. It is as if, in the representation of something, some properties of that something must necessarily be present, notably perceptual properties. In this way, the limits between the two worlds, that of fiction (the image) and the real world (the object), end up blurring. This extends to predicates: an unary predicate is confused with a dual one when similarity is considered a criterion for defining imitation. In this sense, considering Warhol’s Brillo Boxes as imitations or copies of Brillo steel wool boxes is to apply the wrong type of predicate concerning representation. This also occurs with the idea that Warhol’s work is innovative within the tradition of mimetic art solely because it imitates or reproduces banal things like steel wool packaging, as if its originality would consist in the production of sculptures that resemble the packaging of commercial goods found in supermarkets, enhancing their visual properties, and pointing to new conceptions of which properties of the real are relevant for the art world. However, this interpretation needs to consider the problem of indiscernibles, which arises precisely from misunderstanding the type of predicate that generally characterizes representations. At this point, we can say Goodman’s theory of representation gains importance for Danto.

Which representation, Mr. Goodman?

For the purpose of this text, I will just present a synthesis of the representation process as seen by Goodman in Languages of Art from 1968. Firstly, as in the case of maps and diagrams, it is important to understand that similarity relations are neither necessary nor sufficient to define a representation. Similarity presupposes symmetry, reflection, and transitivity. If A=B, then B=A. However, A can represent B without B representing A. If A=B and B=C, then A=C. For example, blue 2009 Fiat Palio cars may all be similar, but none represents the others. To represent is to denote or name. Images function like words or other signs: they are denotative, descriptive, or classificatory. Thus, if the portrait of x represents x, it is because it denotes, describes, or names x according to a set of rules of a symbolic system, in which this portrait is classified as a portrait of x.

An ID card photo represents the person in the photo, but the person does not represent the photo on the ID card. In general: If i is an image of an object o, i symbolizes that object. Symbolizing means “to be in the place of.” An image i is in place of o if and only if it refers to o, denotes the referent o. It is not necessary for i to have properties in common with o to represent it. The image i represents what x is or what makes being x. For example, a painting that represents Napoleon as a general in battle does so by employing everything that makes us recognize Napoleon as a military genius. Every representation depends on what one wants to represent and is conventional. It follows that the idea of a “faithful copy” is meaningless for Goodman. A “faithful” image, one that represents something exactly as it is, presupposes the idea of a neutral, innocent look. However, the gaze does not mirror the object but takes possession of it and makes it something.13 This could serve any purpose. The fact that some representations have extensions does not necessarily imply a similarity between representation and represented. An image, if it has an extension, can refer to a single object or denote the members of the class to which it refers. This is the case with a dictionary entry. For example, the definition of the eagle in a dictionary represents or refers not to a specific eagle but to eagles in general.

There is a more problematic type of representation than the denotative one: one without extension, that is, lacking a referent. That is the case with fictional beings such as unicorns, Capitu and Pinocchio. The analysis of this type of representation contributes to a theory of representation that will be of interest to the question of indiscernibles, as we will see later. For Goodman, this type of representation predicates something despite not denoting anything: they have no referent, that is, they do not point to anything outside the representation itself. They are not relational predicates but unary. A unicorn image does not function as a relational predicate; it does not point to anything outside the unicorn image but functions as a description of a unicorn or a denomination of a unicorn such as a “unicorn image.” A “unicorn image” is not an image of a unicorn animal. That is, it is rather the name of a representation of a unicorn. A significant property of representations follows from this: the fact that R is a representation of o does not necessarily imply that o is an entity in the world, that is, that o is the referent of R. If we take the example of Capitu, or Bentinho, characters from the Machado de Assis’ novel Dom Casmurro, although there are no references in the world for Capitu and Bentinho, they represent a woman and a man, respectively. Goodman’s solution is the unary predicate, the name, or rather the label of the hyphenated representation. Woman-image is the abbreviation, so to speak, of the phrase: “the image that represents a woman”, just as a square-image is not necessarily a square image. Capitu is an image-of-woman that does not necessarily need to be like a woman. Similarity was a determination by the rules of the 19th-century novel. The image-of-woman is the description of a woman, it can be Capitu, Evita Perón, etc. The image of a woman is the description of a woman, such as: “Evita Perón”, “Perón’s wife”, or “the woman who governed Argentina.” We can state that: to denote an object o, an image i must represent o, but an image-of-a-woman does not need to denote anything to be a representation-of-a-woman.

Goodman’s analyzis of the representation of fictional beings leads to a nominalist theory of images. Representations can refer to something, denote something, or just represent a certain use of the verb “represent”. An image representing Eva Perón as a woman, “the woman who governed Argentina”, for example, is a woman-image that denotes Eva Perón. In other words, an image representing an object denotes it; an image i representing a fictitious o is an o-image; an image i representing an object o as o is an image-of-o that denotes it.

The emphasis lies on the “manner” of representation indicated by the conjunction “as”. In the first case, the “how” refers to what the image denotes; in the second, it refers to the “type” of the image; while in the third, it refers to the denotation itself and, therefore, the classification of the referent. There are always two questions when defining a representation: what is being represented and what type of representation it is. Hence, we have predicates that can be objects of prediction. Classifications with verbal or pictorial labels, that is, descriptions or images, are labeled. The name “table” is a label (predicate) of an object o, and this label (predicate) may have an image or a description: a table-image or a table-description. Therefore, a representation, such as an image, identifies a class and belongs to a specific class of images.

From Narcissus back to Brillo Box

As the myth of Narcissus seems to teach Danto in The Transfiguration of the Commonplace, Narcissus learned to recognize the boy reflected in the water not as another boy, a real boy in a world of water with its own properties, but as an image, a boy-of-water-reflection. Note that Danto uses the same method as Goodman to deal with representations of fictional characters. As we have seen, Goodman uses hyphenations for predicates that lack extension; in this sense, the reflection boy is the name, label, and predicate that indicates the type of image that attracts Narcissus. Or better yet, “of-reflection” would be the predicate that characterizes this image of a boy. Danto further illustrates this with the example of dreaming about a cat. This cat is a dream cat, and “dream” predicates it. It is not the predicate of any cat, but the name of an image of a cat, cat-of-dream, just as “of-reflection” is not a predicate of any boy but indicates a type of image of a boy. The use of this type of predicate – “dream”, “reflection”, etc – ontologically isolates the subjects of the predicates by indicating that they do not refer to possible referents. In other words, it classifies them as images of certain types and as types belonging to fiction.

Like the example of Narcissus or the cat’s dream, works of art have the same kind of predicate: an unary one. They are ontologically heterogeneous when they refer to unary predicates. They belong to a separate, fictional world, with no necessary relation to the real world. However, they are ontologically homogeneous if they refer to relational predicates, that is, if they necessarily have an extension. Therefore, something is a work of art within a particular representational system, a specific historical context, but not within another. Just as “Narcissus’-reflection” applies to Narcissus, Warhol’s Brillo Box exemplifies a type of representation characteristic of art, specifically North American pop art, a case of the works of Warhol, and so on.

Consider the statement: the Brillo Box is art. The predicate “is art”, if understood in terms of a type of representational relation intrinsic to every work of art, as Danto wants, must necessarily exclude any similarity criterion, as we had seen in Goodman’s representation concept. Furthermore, the representation that every work of art is does not dependent of any referent and, therefore, is not necessarily denotative. “Warhol’s Brillo Box” is a type of representation that does not necessarily need a referent. It can denote, for instance, a “Pop artwork of art”, the Brillo brand steel wool boxes, or some description of a work that is part of the history of art. But denoting a referent only becomes necessary when the representation belongs to a semantic system that implies the notion of truth. In the descriptive sense, an accurate representation of something is in no way necessary to works of art.

A work of art may be an approximate image or description of its referent and even a representation that is entirely identical to what is represented. Still, this similarity does not make it a representation or a work of art. Following Goodman’s representation theory, a Brillo-steel-straw-box-image does not necessarily need to be visually identical to a Brillo steel-straw box, nor was it necessary for Warholto have used a copy of these boxes for his work. This is because, as we have seen, its status as a work of art does not arise from a relationship of denotation or similarity. The work of art Brillo Box is not an imitation of Brillo steel wool packaging because, precisely, its meaning is not “to be similar”, in any sense, to Brillo steel-straw boxes. As I have highlighted, Warhol’s work does not denote or designate soap packaging. It is not a work of art on Brillo product packaging; Brillo steel wool packaging does not give meaning to Warhol’s work of art.

The Brillo Boxes are works of art not because they accurately describe something but because meaning and representing are synonymous here. They describe their meaning; in the sense given to the term representation in the non-denoting sense, they present their meaning. In this, the work of art differs from the type of imitation of the mirror, as Danto explores in the example of Plato’s Republic: imitation that “describes” the reflected object to the extent that it depends on that object. Despite the similarity between Warhol’s Brillo Box and Brillo packaging, the former is not a description or a denotation of the latter. This is the difference highlighted by Goodman between the type of representation that works of art are and representations that necessarily denote, such as maps, diagrams, graphs, medical reports, etc. Works of art may have a documentary character, but documents are not necessarily works of art and, therefore, do not define a work of art. Menard’s work does not denote Cervantes’ work, does not represent it, and is not a citation of it, just as the meaning of Warhol’s Brillo Box does not denote, describe, quote, or exemplify Brillo Box product boxes. Warhol’s creations are Brillo-box images, where Brillo-Box is the name or label on an image which is a work of art.

Also, we can imagine a situation in which similarity would be a necessary requirement of representations or for which they would only need to denote other representations. If we take the following representations, Danto as baby, “Danto-as-baby”, Danto as philosopher, “Danto-as-philosopher”, we will see that they have different meanings despite referring to Arthur Danto. Danto as a baby could be a photo of Danto when he was a baby. “Danto-as-baby” could be a caricature of Danto in the form of a baby. Danto as a philosopher could be a video of him philosophizing about indiscernibles, and “Danto-as-philosopher” could be a painting portraying Danto as Socrates. None of these images describes or denotes the others but typifies them. Brillo steel wool boxes are ordinary objects; Brillo-Boxes are instead works of art; theyare images and predicates. These unary predicates do not replace, as representations, steel wool boxes, but typify a type of image: “like-Brillo-steel-straw-box,”. Their meaning is not to replace, or necessarily resemble “real” soap boxes, even though they realistically reach them. As Danto exhaustively argued, the meaning of Warhol’s work does not lie in its appearance. It must be sought in its history – the history of its conception and making –in the place that they occupy in the narrative of art made from everyday objects, that is, in the narrative of the incorporation of such objects into works of art or their “transformation” into a work of art. And this applies to any work that uses the same method characterized by Brillo Box.


Danto thinks about representation based on three meanings. The last two are very close to Goodman’s types of predicates apart from the one in which representation has the meaning of presentation, in the sense that Danto recovers from the analysis of Nietzsche’s text. In this case, there is no ontological distance between the representation and what it represents. The image is not in place of the referent; it is indistinguishable from it. In the second sense, representation denotes the referent, contrasting with reality. It is a “representation of...”, “image of...”, “description of...”. Truth values can be applied to it. The last sense of representing contrasts with reality, or at least with the necessity of a referent. It can denote, but not necessarily. It points to representation as a type of image or description. A “representation of representation” names, that is, typifies a representation: “image-of-x as y”, where y may be, or not, identical to x.

Works of art fit into this last type of representation. In this sense, we can risk saying that Warhol’s Brillo Boxes are close to Picasso’s portrait of Ambroise Vollard. Just as the “subject” of Warhol’s Brillo Boxes is not about the Boxes of the mark Brillo, the portrait of Vollard painted by Picasso is not about Vollard, although we can recognize some features of this dealer’s face. This painting “resembles” any other Cubist painting much more than it resembles Vollard. Therefore, it deals with, points out, designates, and has the following meanings, such as, for example, the deconstruction of the three-dimensional illusion of space in Western painting, the spatial theory of Cubism, the valuing of the pictorial plane over the traditional “window” scheme in Western painting and so on.

The same can be said of Warhol’s Brillo Box. It refers, for example, to the difference between art and reality, that is, to objects or fragments of everyday life that enter into art and become part of the reference that works of art make to reality; to a notion that reality, everyday objects or gestures can be introduced into the history of art since, we can say, cubist and dada collages since modern music incorporated machine noises, familiar sounds, or even silence. It can also refer to American consumer culture, etc. In this sense, Warhol’s Brillo Box is more like other pop artworks than the Brillo brand’s steel wool box.

Finally, far from intending to exhaust the argument of The Transfiguration of the Commonplace, this article was limited to discussing the idea of representation in the problem of indiscernibles and the importance of Nelson Goodman’s theory of representation for it. My aim was not to outline any reductive synthesis of Danto’s theory of art either. This would go beyond the limits of the problem which originated this paper, that is, the problem of indiscernibles and the idea of representation as a fundamental characteristic of works of art.

I believe this conception of representation is directly linked to Danto’s use of the idea of intension. I was interested in this article only to raise the hypothesis that Goodman’s conception of representation is central to the problem of indiscernibles, as proposed by Danto. However, if we take the idea of name, label, or description as representation, we can think that intension as an intensional context must be closely related with what Danto adopts from the Fregean tradition, thus bringing Goodman closer to Frege in terms of the conceptual content of representation, or better, by defining representation as the conceptual content. Frege’s influence on Danto’s thought is multiple. However, it seems to me that, fundamentally, to understand the contents incorporated in the work of art, what it is about – which legitimizes certain hermeneutic procedures, as Bogéa well demonstrates14 – it is essential to take Frege’s theory of meaning into account, as Bogéa shows, since it is directly linked to the notion of intensionality, or contextual intensionality, in Danto. As I stated above, it is not the purpose of this text to extend considerations about the presence of Goodman’s conceptions of representation to the question of intention in Danto, especially as it appears in chapter seven of The Transfiguration of the Commonplace, but it was important to point to this possible relationship.

Bibliographical references

BOGÉA, Anderson. “A influência fregeana na filosofia da arte de Arthur Danto”. Discurso, v. 53, n. 2 (2023).

COHNITIZ, Daniel; ROSSBERG, Marcus. Nelson Goodman. Montreal: McGill-Queen’s University Press, 2006.

COSTA, Rachel. “O mundo da arte como proposta”. Paralaxe, v. 2, n.1 (2014),p. 81-98.

DANTO, Arthur C. A transfiguração do lugar-comum. São Paulo: Cosac Naify, 2010.

GOODMAN, Nelson. Linguagens da arte. Lisboa: Gradiva, 2006.

_____. Modos de fazer mundo. Porto: edições Asa, 1995.

_____. Languages of Art. Indianapolis: Hackett Publishing Company, 1984.

_____. Ways of Worldmaking. Indianapolis: Hackett Publishing Company, 1984.

KINSELLA, Eileen. “The Brillo-Box scandal”. Art News, 2009. Disponível em:<>.

RAMME, Noeli. “A estética na filosofia da arte de Arthur Danto”. Artefilosofia, v. 3, n. 5 (2005), p. 87-95.

SANTOS DO CARMO, Juliano. “Nelson Goodman: sobre a natureza da experiência estética”. Intuitio, v. 8, n. 1 (2015), p. 04-14.

* Walter Menon Junior é professor do Departamento de Filosofia da UFMG
1 KINSELLA, 2009.
2 DANTO, 1981, p. 35.
3 DANTO, 1981, p. 35.
4 DANTO, 1981, p. 36.
5 DANTO, 1981, p. 39.
6 DANTO, 1981, p. 40.
7 DANTO, 1981, p. 40.
8 DANTO, 1981, p. 40.
9 At this point in the text, Danto discusses Goodman’s conception that it is possible, with visual accuracy and training, to recognize the style, materials, design, and so on, characteristic of an artist’s work to differentiate an authentic piece from a forgery. The emphasis is on a perceptual capacity developed through training and knowledge. It is evident that the perceptual criterion is not suitable for identifying works that fit the problem of indiscernibles, precisely because perceptual or aesthetic properties are neither sufficient nor necessary to identify a work of art. Instead, we should consider other data – for example, historical data, the sequence of events that presides over the making of a work – and certain necessary elements, such as an author. Despite this debate with Goodman, it should be noted that even if he defends the possibility of recognizing a work of art by its perceptual properties, this can only happen with training and in specific situations. For instance, when comparing two similar paintings where no difference in time, material, etc. is recognized. In such cases, Danto agrees that one must rely on the expertise of specialists. This does not call into question Goodman’s theory of representation, which Danto seems to appropriate at certain points, as I try to show throughout this article.
10 DANTO, 1981, p. 60.
11 COSTA, 2014, p. 86.
12 DANTO, 1981, p. 18.
13 GOODMAN, 1968, p. 27.
14 See the discussion that Bogéa (2023) carries out around the metaphor issue in Danto, especially as incorporated content.. The metaphor would necessarily have an intensional structure. As Bogéa states “Thus, within the scope of Danto‘s discussion, saying that metaphors have an intensional structure is to state that the elements that make up the metaphor (whether grammatical or artistic) cannot be replaced by identical objects without the context being considered. In these cases, there is a contextual peculiarity and historical, in which the form of presentation in metaphors occurs evidently according to the meanings and associations they have in the cultural framework of an era” (2023, p. 116). Therefore, the historicity of predicates is fundamental, that is, their context of production and use, which is the same thing in indicating an object. As we propose in this text, the relevant artistic predicates are unary and form a name or description of the object; these unary predicates depend on the context of their historical formation.