Can Don Quixote Tilt at William James? or: How I Learned to Stop Worrying and Love Pierre Menard

Charles Johnson
2 March 2002

This essay is © 2001-2002 by Charles Johnson, and is freed for credited reprinting and derivative works (but not, of course, for plagiarism) under the terms of the Creative Commons Attribution-ShareAlike 2.0 copyleft license. If you have any questions or comments about the essay, feel free to contact the author.

I. The Quixotic Career of Pierre Menard

In Pierre Menard, Author of Don Quixote (1962), Jorge Luís Borges recounts how a friend of his, a French poet and sometime philosopher, decided in the mid-1930s to write Don Quixote. He did not, according to Borges, want to compose another Don Quixote—which would be easy—but the Don Quixote (48). Menard intended to produce pages which would coincide—word for word and line for line—with those of Miguel de Cervantes (49)—not by a mechanical transcription of the original (49), but by his own literary efforts. With only fragmentary memories of his adolescent reading of Quixote—a vague image little different from the outlines that an author would have of a work she has not yet begun—Menard labored for years. He considered, but rejected, the project of becoming Miguel de Cervantes. To rewrite Don Quixote through the experiences of Cervantes would be a far less interesting task than to continue being Pierre Menard and to arrive at Don Quixote through the experiences of Pierre Menard (49). And, indeed, by his death Menard had succeeded in replicating the ninth and thirty-eighth chapters of Part One of Quixote as well as a fragment of the twenty-second. (I.1)

Borges claims that on reading Menard, one finds that The text of Cervantes and that of Menard are verbally identical, but the second is almost infinitely richer (52). As three hundred years have passed, charged with the most complex happenings—among them, to mention only one, that same Don Quixote (51), and authored by a French devotee of Baudelaire rather than an old soldier of seventeenth century Spain, Menard’s work probes history and philosophy with intellectual sophistication in ways Cervantes’s cannot. Whereas (I.2)

The latter indulges in a rather coarse opposition between tales of knighthood and the meager, provincial reality of his country; Menard chooses as reality the land of Carmen during the century of Lepanto and Lope. What Hispanophile would not have advised Maurice Barrès or Dr. Rodríguez Larreta to make such a choice! Menard, as if it were the most natural thing in the world, eludes them. In his work there are neither bands of gypsies, conquistadors, mystics, Philip the Seconds, nor autos-da-fé. He disregards or proscribes local color. This disdain indicates a new approach to the historical novel. This disdain condemns Salammbô without appeal. (51-52). (I.3)

On the other hand, Menard’s text has deficiencies of style compared to the identical text of Cervantes; the intentional use of seventeenth century Spanish by the twentieth century French intellectual suffers from a certain affectation whereas Cervantes handles easily the ordinary Spanish of his time (53). (I.4)

Menard’s Quixote raises an important question. Is it possible that through repetition of a pre-existing book in a foreign tongue (54), Menard has authored a [new] work, possibly the most significant of our time (48)? To generalize the problem, is it possible that two pieces of art might be perfect physical replicas of one another, and yet have radically different artistic properties? Is it possible for Cervantes’s Quixote to tilt only at windmills, whereas Menard’s tilts at the philosophy of William James? (I.5)

If Menard could not have written a better book than Cervantes because he produced nothing which was not already there, then we must exclude from critical consideration anything and everything other than the bare physical details of a work and whatever moral, emotional, or other properties that arise directly from those details. We must, as Kendall Walton (1970/1995) summarizes a position not his own, ignore how and when a work was created, the artist’s intentions in creating it, his philosophical views, psychological state, personal life, the artistic traditions and intellectual atmosphere of his society, and so forth (332). This view has been propounded by, among others, Formalists such as Clive Bell, and since it argues that we must only consider the perceptible elements of a work (and perhaps the properties entailed by those elements), we may dub this family of positions Aestheticism. (I.6)

If, on the other hand, Menard has succeeded, then we must admit at least some of these historical properties that the Aestheticist excludes, and reject any aesthetic theory which bases a work’s artistic properties only on the perceptible forms or elements of the work. Since such a view considers contextual properties external to the perceptible properties of an artwork, these theories may be called Contextualist1. The groundwork for a sophisticated Contextualist position is laid by Kendall Walton in Categories of Art, and in light of Walton’s theory—supplemented by concepts from Arthur Danto’s The Artworld—the Aestheticist claim should be rejected. Menard could, indeed, create a new artwork—although perhaps not in the way that Walton or Danto thinks he could. (I.7)

II. The Aesthetic Hypothesis

According to the Aestheticist, since Menard has successfully replicated Don Quixote word-for-word, he really has succeeded in replicating Quixote, and despite the arduous labors it required, he has no more created a new artwork than a printer does by reprinting copies. This commits the Aestheticist to a theory under which artistic features are harshly separated from every condition external to the perceptible properties of an artwork. If no artwork with the same perceptible elements on the page can have different artistic properties, then it must be the case that all that can be relevant to enjoying art is that the object possesses certain perceptual properties, and that these properties realize some aesthetic effect. Thus, in the words of Clive Bell (1914/1995), we have only to consider our emotion and its object: for the purposes of aesthetics, we have no right, neither is there any necessity, to pry behind the object into the state of mind of him who made it (101). (II.1)

Over the past century the most significant exponents of Aestheticism have been the most severe: the Formalists. They argue that the only question of art criticism is the description of the artwork’s perceptible form—in the case of painting, in terms of lines and colours combined in a particular way, certain forms and relations of forms (Bell [1914/1995] 100).2 However, not all Aestheticism is Formalism. A Formalist and a Contextualist would argue over whether or not the artistic properties in Quixote are identical with its arrangement of perceptible elements. However, other Aestheticists would argue with a Contextualist whether the artistic effects of Quixote supervene on its perceptible elements. For example, an Aestheticist could argue a mimetic theory, that the essence of art is its ability to represent a certain object to a viewer. The appearance of an artwork, and thus its mimetic qualities, cannot change without a change in the perceptible (non-aesthetic) elements of the artwork. In this case, Don Quixote is still portraying the exact same thing, and therefore Menard has not succeeded in producing anything new. (II.2)

III. Walton’s Critique

In order to develop a critique of Aestheticism, Kendall Walton (1970/1995) elucidates the concept of artistic categories. It is uncontroversial that there are different categories of art, such as painting, cubist painting, Gothic architecture, classical sonatas, painting in the style of Cézanne (335). Works are recognized in these various categories through three different kinds of properties that artworks possess: standard, variable, and contra-standard. A standard feature of artworks within a category is among those in virtue of which works in that category belong to that category—that is, just in case the absence of that feature would disqualify, or tend to disqualify, a work from that category (335-336). A variable feature of a category is one which has nothing to do with works belonging to that category; the possession or lack of the feature is irrelevant to whether a work qualifies for the category (336). And a contra-standard feature is the absence of a standard feature with respect to a category—that is, a feature whose presence tends to disqualify certain works as members of the category (336). (III.1)

Each of the three features plays its own role in art. Artworks are categorized according to their standard and contra-standard features: the artwork belongs to whichever category in which it is qualified by the greatest number of standard and counteracted by the least number of contra-standard features. The variable features provide the content of an artwork. It is standard of a Classical bust that it is made of marble, white, and severed at the chest. The features of the person represented are variable: the bust may represent Julius Caesar, George Washington, or a completely fictional woman with curly hair and a sharp nose. It would be contra-standard and shocking if a mechanical device were used to make the bust wink every five minutes: busts are not supposed to move. However, this would not necessarily disqualify it from being a classical bust: if all the other standard features are there, then the artwork is probably best categorized as a bust—just a queer bust with a disturbing tendency to wink. (III.2)

Crucially, however, the categories in which we perceive artworks help to determine what aesthetic properties a work has (352). A masterful Petrarchan sonnet, if it were perceived as free verse, would seem affected and pretentious; a free verse of fourteen lines, if it were perceived as a sonnet, would seem undisciplined and chaotic. Fair enough—normally there is no difficulty in properly categorizing works of art solely based on the weight of their standard and contra-standard features. We do not need anything special to distinguish a fourteen-line free verse from a sonnet: the sonnet has well-defined rhyme schemes and meters which tend to disqualify irregular poetry, whereas fixed rhyme and meter tend to disqualify a poem from being seen as free verse. Perceiving the artwork is good enough to properly categorize it. (III.3)

However, a deeper investigation will reveal that there are many points at which this breaks down. A still photograph of a high jumper is motionless, but if we see it in a gallery it does not look like a high jumper frozen in mid-air. Indeed, in the photograph the athelete may seem in a frenzy of activity; the pictures may convey a vivid sense of movement (339). However, if perceptually identical static images exactly like those of the two pictures occur in a motion picture, and we see it as a motion picture, they probably would strike us as resembling a static athelete (Ibid.). The Aestheticist will reply that this situation or medium in which a work is presented is itself part of the form of the work. A film (which is presented over time) is different from looking at a series of static pictures (which each portray only a static instant). (III.4)

But the move falls apart when we consider that for any piece of art, we can construct infinitely many different arbitrary categories into which it can equally well be placed. A seemingly formless and chaotic musical piece may be seen as a masterful work of meticulous structure if we invent an arbitrary set of rules about the precise number of As it must have, the times at which key must shift to C-sharp, and so on.3 Given the contradictory artistic effects of the ever so many categories that we can construct, we must ask, How is it to be determined in which categories a work is correctly perceived? (347) before we can appreciate any artistic effects at all. But we cannot do this on the basis of anything about the presence or absence of perceptual properties. The categories we are working with have been constructed in such a way that the artwork has equally good perceptual credentials as either one or the other. (III.5)

We might try to use contextual factors other than the historical concerns that the Aestheticist wants to exclude. For example, a principle of artistic charity: the proper category is whichever one makes the artwork appear most skillful. Yet if we invoke this, we have basically made it impossible for bad art to exist. For if we Take any work of art we can agree is of fourth- or fifth- or tenth-rate quality, then It is very possible that if this work were perceived in some far-fetched set of categories that someone might dream up, it would appear to be first-rate, a masterpiece (349). Some of the possibilities for this were mentioned above. We might transfigure a sorry musical piece into a masterwork by imposing arbitrary mathematical rules on the composition.4 But surely from this it does not follow that the work really is a hitherto unrecognized masterpiece (349). We would be doing violence to the possibility of art criticism if we were to impose such a condition. (III.6)

The most natural criterion remaining is the historical circumstances of the artwork and its artist. The historical context, intellectual milieu, creator’s intention, and so forth are all necessary considerations in determining the proper category for an artwork. Bell (1914/1995) is wrong to say that we have no right, neither is there any necessity, to pry behind the object (101)—in fact, the critic must go beyond the work before him in order to judge it aesthetically (Walton [1970/1995] 353, emphasis added). Indeed, the tables have been turned on Bell: if the impact an artwork has is entirely due to its formal, aesthetic properties, then we must recognize that historical properties are relevant to artistic effect, because we need them in order to be able to perceive the right formal properties. (III.7)

With this in mind, we may now see how Pierre Menard has succeeded in producing a new work of art. Because the relevant historical facts of an artwork help to determine what aesthetic properties a work has (Walton [1970/1995] 352), Menard’s Quixote has shifted to a new category. Whereas Cervantes’s Quixote was a seventeenth century Spanish popular novel, and therefore, say, incapable of commenting on the philosophy of William James, Menard’s Quixote is a twentieth century French intellectual novel, putting it in an entirely different relationship to the tradition of art, literature, and philosophy. How can the works have different effects when they are identical to the eye? Because merely examining a work with the senses can by itself reveal neither how it is correct to perceive it, nor how to perceive it that way (Walton [1970/1995] 354). The two works fall into distinct artistic categories, and therefore to perceive Menard’s Don Quixote we must attend to entirely different features, and interpret those features in a different context, from the Quixote of Cervantes. (III.8)

IV. Assets and Liabilities

Walton provides a fundamentally dynamic account of what things are included under the definition of art, since the range of standard and variable artistic properties can shift and expand over time. This will not satisfy anyone seeking a fixed, Socratic sort of a definition. However, it does help explain important features of the world of art. (IV.1)

Arthur Danto (1964/1995) observes that with the development of a new style (i.e., a new category with the added standard feature) H, every other painting in existence becomes non-H, and the entire community of paintings is enriched, together with a doubling of the available style opportunities (212). The presence or absence of H, which had once been irrelevant to artistic merit, is now part of the set of standard and contra-standard properties which delimit the various denizens of the artworld—including those which were made before H became a standard feature of an artistic style. With this retroactive enrichment of the entities in the artworld, the tradition of art has a nearly infinite capacity to expand and enrich itself, allowing us to discuss Raphael and De Kooning together, or Lichtenstein and Michelangelo (Ibid.). It also helps explain why appreciation of art grows keener through the experience of a wide variety of artworks and knowledge of art history. For those unfamiliar with the matrix, it is hard, and perhaps impossible, to recognize certain positions as artworks (Ibid.), but through greater acquaintance with the standard and variable elements of the category, one gradually acquires the ability to properly categorize—and thus, properly appreciate—the given style of art. (IV.2)

On the other hand, despite Danto’s optimism about the progress of the artworld, we must also note that there are dangers to an increasing population, and that in some cases the enrichment of styles may result in the impoverishment of existing artworks. This is unwittingly illustrated by Walton (1970/1995): (IV.3)

One such dispute might well arise concerning Giacometti’s thin metal sculptures. To a critic who sees them simply as sculptures, or sculptures of people, they look frail, emaciated, wispy, or wiry. But that is not how they would strike a critic who sees them in the category of thin metal sculptures of that sort (just as stick figures do not strike us as wispy and emaciated). (350) (IV.4)

Yet Giacometti’s art hinges, in part, on seeing the sculptures precisely as emaciated and wiry, as a shock given the monolithic standards of human figure sculpture. This expansion of the artworld has led to the disreputable process of genrification. Whenever a shocking new work is introduced as a reaction to the tradition, this work’s features enrich the properties of the artworld, and a genre of works begins to coalesce around this new way of doing art, adapting the insights of the first work. Les Demoiselles d’Avignon shocked the artworld through contra-standard distortions of proportion, form and color, but soon, these contra-standard properties became standard properties of analytic cubism. However, since Shock … arises from features that are not just rare and unique, but ones that are contra-standard relative to categories in which objects possessing them are perceived (Walton [1970/1995] 345), the entrenchment of these new genres undermines their disruptive power. Indeed, since there is retroactive enrichment of the entities in the artworld, (Danto [1964/1995] 212), it even overturns the shocking contra-standardness of the original, converting it into the matronly grandmother of the genre. (IV.5)

By this process a once-radical artwork is, quite against its will, tamed into an agreeable conservatism, and by becoming commonplace loses the boundary existence that had once made it significant and inspired vigorous reactions.5 Indeed, just this has become the condition of most art of the nineteenth and twentieth centuries. Dadaist Anti-Art is sold through art catalogues and has become a section of every history of the Canon; Les Demoiselles d’Avignon shocks no-one; and Le dejeneur sur l’herbe seems positively staid and conservative. While the backward-looking enrichment of the Artworld gives us many reasons to appreciate artistic innovation, it also puts us in danger of losing the innocence needed to appreciate many of history’s most significant works of art. (IV.6)

V. A New Way of Reading

Aestheticism fails to explain artistic perception, but Walton and Danto’s version of Contextualism may be misguided as well. The issue can be framed with another question about Menard’s success as an author. (V.1)

While Menard did in fact write two chapters of Don Quixote, Borges suggests the enrichment of the artworld by a further means: subjective re-identification in the process of reading. Borges recounts that I often imagine that he finished it and that I am reading Don Quixote—the entire work—as if Menard had conceived it (50). Has Menard succeeded only in producing a new artwork in two chapters, or can Borges’s imaginary re-attribution of Cervantes’s other chapters to Menard produce an entire new Don Quixote? (V.2)

Whether this succeeds is particularly interesting because of a one crucial fact about Pierre Menard: he is a fictional character invented by Borges. While Menard has succeeded in producing a new artwork within the story, is it possible that he has also succeeded in producing one outside?6 Is it possible for us, outside of the story, to read Don Quixote as if the fictional Menard had written it? (V.3)

Walton and Danto do not provide any framework for such erroneous attributions. Their account of the categorization of artworks is based on objective, historical circumstances under which the artist produced it. Under the objective-historical understanding of categorizing artworks, there doesn’t seem to be any way in which Menard has affected our reading of Don Quixote. It’s a nice story, after all, but really, this copy of Quixote in front of me is Cervantes’s Quixote, not Menard’s. Not surprisingly—Menard does not actually exist. (V.4)

In reply to such an artist-centric account, one might raise the objection of Roland Barthes in Musica Practica (1970), that (V.5)

There are two musics (at least so I have always thought): the music one listens to, the music one plays. These two musics are two totally different arts, each with its own history, its own sociology, its own aesthetics, its own erotic. (149) (V.6)

The statement applies equally well to other forms of art.7 If there are indeed two sets of artistic properties for every artistic object — the art produced by the artist and the art experienced by the public — then it would seem that the latter ought not to be bound by the same kind of standard criteria as the former. For the ars practica of the artist, it is correct that the categorization of a work depends upon the historical circumstances and artistic intentions through which it was produced. However, for the appreciator of a work of art, the correct categorization is a matter of a more personal attribution. Of course, purely arbitrary re-categorizations will not be easy. Vide Walton (1970/1995): (V.7)

One cannot merely decide to respond appropriately to a work—to be shocked or unnerved or surprised by its (absolutely) contra-standard features, to find its standard features familiar or mundane, and to react to its variable features in other ways—once one knows the correct categories. Perceiving a work in a certain category or set of categories is a skill that must be acquired by training (353) (V.8)

But nothing in this says that, in principle, that the criteria of correctly categorizing might not be changed to reflect the new form of art that Menard has produced. We might argue that the appreciator creates a fiction in which the actual artwork is used as a prop in a game of make-believe, and the historical properties attributed to it are fictions stipulated by the appreciator’s knowledge and interests. Engaging with an artwork already involves a game of make-believe about its content, and it does not seem difficult to account for the work’s category from additional make-believe ascriptions. The imaginative resistance to re-categorizing works remains, since it is certainly easier and more natural to use a given prop for some make-believe games, than for others—a banana is more easily a make-believe telephone than a make-believe television. (V.9)

But now we only glean the artistic experience through make-believe, and isn’t that in principle a different artistic experience from that of an actual artwork with the actual historical properties? Perhaps not. It is not necessary that an object of make-believe be unreal — in a game of make-believe, part of the game may be that I have blue eyes, and in fact, I actually do have blue eyes. Perhaps the actual historical properties that ordinarily condition our appreciation of the Quixote of Cervantes are no less objects of make-believe, than the merely fictional properties that we attribute to the Don Quixote of Menard. We do in fact tend to lean towards the actual, ars practica category of the artwork (that Quixote was authored by Miguel de Cervantes), because there is nothing easier to make-believe than what we know to be the case. But if the aesthetic of the appreciator’s art is in principle that of a make-believe game, then through considerable imaginative effort, we may liberate an infinite variety of new artworks from the old, and (V.10)

Menard (perhaps without wishing to) has enriched, by means of a new technique, the hesitant and rudimentary act of reading: the technique is one of deliberate anachronism and erroneous attributions. This technique, with its infinite applications, urges us to run through the Odyssey as if it were written after the Aeneid, and to read Le jardin du Centaure by Madame Henri Bachelier as if it were by Madame Henri Bachelier. This technique would fill the dullest books with adventure. (Borges [1964] 54-55) (V.11)


  1. In order to avoid unfortunate confusions with the views in the social sciences, I have avoided the name Historicism. (FN.1)

  2. The same account is given of other artistic forms. In music, the Formalist claims to pay attention only to the perceptible arrangement of notes; in poetry, only to the deployment of linguistic plays, rhythm, meter, figurative linguistic plays, and so on; in drama, only to the arrangement of plot, characters, and spectacle; and so on. (FN.2)

  3. Indeed, it seems that anything and everything — a print of the Mona Lisa, a copy of a sonnet, a copy of Pierre Menard’s Don Quixote — are also (potentially) Dada ready-mades, giving them all kinds of different ironic and conceptual artistic effects than they would have had if perceived according to their normal category. (FN.3)

  4. Indeed, one might argue that this is precisely what has happened with twelve-tone music in the twentieth century. (FN.4)

  5. As with scholarly interpretation of historical traumas, the artworld finally incorporates shattering works into the system of signs and interpretations that they were designed to disrupt, and a noble posterity might misunderstand the whole past and in that way alone make it tolerable to look at (§38). Nietzsche, Friedrich (1886). Beyond Good and Evil. Trans. Walter Kaufmann (1989). New York: Vintage Books. (FN.5)

  6. Nevermind the metaphysical problem of how a fictional character might effect such a thing in the actual world; the concern here is not over exactly who the causal agent was, but whether the effect is real or not. (FN.6)

  7. Except perhaps for avant garde forms such as aleatory music, in which the role of the artist is more or less removed. (FN.7)