Are Artworks Constituted by Interpretation?
Sherri Irvin
University of Oklahoma
sirvin@ou.edu
Arthur Danto has argued persuasively that an artwork cannot be identical to a mere physical
object. He holds that artworks are constituted by interpretations, which transfigure objects into
works of art. The notion of interpretation central to Danto's philosophy, and hence the thesis
that artworks are constituted by interpretation, remains ill understood. The aim of the present
paper is to demonstrate that Danto's concerns do not necessitate the view that artworks are
constituted by interpretation; and, indeed, that given his other statements about interpretation, it
should be granted at most a very restricted constituting role.
The Non-Identity of Artworks and Physical Objects
Danto offers compelling examples to show that artworks cannot be identified with mere physical
objects. First, since the mass-produced objects presented by Marcel Duchamp did not undergo
any physical change when they became artworks, their physical features cannot explain their
artwork status. (The Philosophical Disenfranchisement of Art, 26) Second, for any sculpture or
painting, it is in principle possible that an object indiscernible from it could be produced through
a fluke event (like an explosion in a quarry), but such an event would not bring an artwork into
existence. (PDA 40) Simply having a set of artwork-like physical features, then, is not sufficient
to make something an artwork, or to give it the sorts of meaning properties characteristic of
artworks. Third, there might be two or more artworks that are physically indistinguishable, yet
have different themes and meanings. For example, of two identical red square paintings, both
titled Red Square, one might be "a clever bit of Moscow landscape," while the other is "a
minimalist exemplar of geometrical art." (The Transfiguration of the Commonplace, 1)
Copyright  2007, Sherri Irvin 1
To account for these phenomena, Danto suggests, we must see artworks not as mere
physical objects, but as constituted by interpretations that transport the objects into a realm of
meaning. To assess this claim, we must first examine Danto's account of interpretation.
Danto on Interpretation
One of the chief functions of interpretation, according to Danto, is to specify which features of
the physical object belong to the work and which do not.
[T]he form of the work may be that gerrymandered portion of the object the interpretation
picks out. Without the interpretation, that portion lapses invisibly back into the object, or
simply disappears, for it is given existence by the interpretation. (TC 125)1
Interpretation, then, "constitutes" the work in at least two ways: it gives existence to the work
and picks out which things in the world belong to it. In addition, interpretation plays a
transfigurative role which is "not institutional, it is ontological": "learning [something] is a work
of art means that it has qualities to attend to which its untransfigured counterpart lacks" (TC 99),
namely, meaning-related qualities. To recognize that something is a work, Danto suggests, is to
go "from the realm of mere things to a realm of meaning." (TC 124) "To interpret a work,"
Danto says, "is to offer a theory as to what the work is about, what its subject is." (TC 119)
Interpretation, then, picks out the elements that belong to the artwork and involves claims
about the artwork's meaning. The connection between these two functions becomes clearer
when we examine what sort of account of the work's meaning an interpretation is supposed to
offer. An interpretation, Danto suggests, proposes a "system of artistic identifications" for the
work.2
(TC 120) One makes an artistic identification "when one says [of] a dab of paint that it is
1
For instance, of a statue of a cat chained to an iron railing at Columbia University, Danto says that there is no way
of telling, independent of interpretation, whether we are seeing "a chained sculpture of a cat" or "a sculpture of a
chained cat." (TC 102)
2
The relationship between interpretation and artistic identification is not altogether clear in Danto's oeuvre. He
sometimes speaks of interpretation as though it simply is (or, perhaps, presupposes) a system of artistic
Copyright  2007, Sherri Irvin 2
Icarus, or of a smudge of blue paint that it is the sky." (TC 126) An artistic identification, it
seems, proposes a relationship between a feature of the object and something that feature is
supposed to signify, and thereby constitutes an aspect of the artwork's meaning. Thus it makes
sense that interpretation both selects which features of the object belong to the artwork and gives
the artwork its meaning: perhaps it selects all and only those physical features to which it assigns
meaning through artistic identification.3
Moreover, artistic identifications are interrelated, and
altering one may cause a cascade of changes throughout the entire system. "In terms of the logic
of artistic identification," Danto says, "simply to identify one element imposes a whole set of
other identifications which stand or fall with it." (TC 119)
Danto summarizes these functions by saying that "each interpretation constitutes a new
work" and that interpretation (I) "is a sort of function that transfigures o [the object] into a work:
I(o)=W. Then even if o is a perceptual constant, variations in I constitute different works." (TC
125, emphasis in original)
Constitution of Works without Interpretation
Should we accept the claim that the artist constitutes the work through interpretation by
proposing a system of artistic identifications?4
It is important that such a view not make
interpretative statements by the artist incorrigible; for, notoriously, artists can be mistaken in the
attribution of meaning to their works. Just as a writer can be ignorant of the meaning of a word,
a visual artist can fail to recognize that a resemblance condition necessary to establish a
identifications: "[s]ave with respect first to an `interpretation' and hence some artistic identification of the relevant
element, no sensible or possible answer can be given to the question of how may elements the work contains." (TC
123) He also speaks of identifications as being "justified" by interpretation (TC 122) and says of interpretation that
it "must ... be justified through identifications." (TC 119)
3
Cf. PDA 41-42: "Interpretations pivot on artistic identifications, and these in turn determine which parts and
properties of the object in question belong to the work of art into which interpretation transfigures it."
4
As Robert Stecker notes in his contribution to this conference, in TC Danto does not draw an explicit distinction
between artists' and viewers' interpretations, and his comments tend to suggest that he accords a constituting role to
both. I will here focus on the claim that artworks are constituted by artists' interpretations, which I take to be the
Copyright  2007, Sherri Irvin 3
Copyright  2007, Sherri Irvin 4
depiction relation is not satisfied. To account for the possibility of such errors, we must appeal
to the independent existence of some entity that is capable of constraining acceptable
interpretations. On Danto's view, however, it is difficult to locate a viable candidate to play this
constraining role. The artwork itself cannot play it: for the artist's interpretation both brings the
artwork into existence and constitutes its meaning-related features. The artwork and its meaning
come into existence, it seems, simultaneously. Nor can the physical object play the required
constraining role: as Danto repeatedly avers, it is not the sort of thing that belongs to the realm of
meaning at all; and in any case, until interpretation has delineated the artwork's boundaries, we
do not know which physical object is in question.
At times, Danto seems to acknowledge that the artist's interpretative activity is subject to
constraints, as when he says that an upright rectangular painting bisected by a straight horizontal
line would not sustain the interpretation suggested by the title Old Man Planting Spring
Cactuses. (TC 130) How can we account for the existence of such constraints, while preserving
a view in the spirit of Danto's? I suggest that we separate a number of functions that Danto
merges together in speaking of the constituting power of interpretation. As we saw earlier, on
Danto's view interpretation constitutes the work by (1) bringing the work into existence, (2)
picking out which physical elements belong to the form of the work and (3) establishing a set of
artistic identifications. These functions need not be merged into a single work-constituting act;
and to the extent that they are all performed simultaneously, this need not be through the artist's
interpretative activity.
Let us take, first, functions (1) and (2). Consider the case of a straightforward
representational painting, such as a traditional landscape. An artist applies paint to canvas in a
way that is obviously indicative of artistic intentionality, and perhaps engages in some act of
presenting this canvas by displaying it in an exhibition. Certainly, upon viewing this canvas we
more plausible of the two alternatives.
will have no doubt that the artist had a particular set of artistic identifications in mind. But the
interpretation that includes or consists of these artistic identifications is not what brought the
work into existence; indeed, the interpretation may well have predated the work's existence. The
work was brought into existence by an act of painting which, while no doubt guided by an
interpretative conception, is not identical to that conception.
Nor is interpretation what gives the work its form, or physical boundaries. The artist has
no need to explicitly specify these; and, indeed, it is unclear that it would be within the artist's
power to do so. The boundaries of a traditional painting are fixed by genre conventions: the
painted surface of the canvas falls within the boundaries, and the wooden stretchers and oil
stained reverse of the canvas fall outside of them. Functions (1) and (2), then, are accomplished
in such cases through an artist's acts of making and/or presenting an object.
Of course, such acts need not always be strongly convention-governed; and within the
convention-defying environments of modern and contemporary art, they are unlikely to be.
However, this does not thrust the artist's acts of work-making into the realm of interpretation.
Take Danto's own story about a work by Marcel Duchamp:
The cracks which appear in the glass panels of Duchamp's [Large Glass]
damaged the object and obscured the work until, leaving the damaged object
where it was, Duchamp made them part of the work, enlisting as unwitting
collaborator the shoddy freight handler. (PDA 42)
The glass was cracked accidentally, then, and Duchamp incorporated the cracks into the work by
decree. This decree was work-constituting: it picked out a feature of the object for inclusion in
the work's form. However, nothing about the episode suggests that Duchamp offered a new
interpretation; the stipulative decree involved no claim about the meaning of the work or the
newly incorporated feature. Even if we accept that the artist can constitute the work, or certain
aspects of it, through such stipulations, we need not accept that such stipulation amounts to or
Copyright  2007, Sherri Irvin 5
necessarily involves interpretation.5
After all, it seems perfectly conceivable that Duchamp may
have had no specific view about how this change in the work's form would affect its meaning.
What about function (3), the establishment of a set of artistic identifications? Does the
constitution of such identifications depend on the artist's interpretative activity? In one sense, of
course, it does: the artist's interpretative ideas play a central causal role in her decisions to give
the work particular features. This does not show, however, that we should see the artistic
identifications as being constituted by the artist's interpretation rather than by the activity of
making and presenting the object. When an artist presents a painting with a particular
configuration of colors, in a context where particular conventions for apprehending
representational content are effective, the artistic identifications are, in many instances,
constituted by that act of presentation. The features of the painting, given the relevant context,
determine directly that the depicted scene is, say, a pastoral landscape; the artist's interpretative
activity, which motivated the acts of creating and presenting the painting, need play no additional
role in constituting the artistic identifications.
This view allows us to account for the fact that artists can make incorrect interpretative
statements: for the work, once brought into existence by some act of making or presentation, can
serve to constrain interpretative activity. We may find that an artist had in mind an artistic
identification that is simply unsupported by the features of the painting presented.
A Limited Work-Constituting Role
I have suggested that the artwork-constituting functions outlined above may be performed
through acts of making and presentation rather than interpretation. However, I have not
discussed the examples that centrally preoccupy Danto: cases in which the features of the
5
I argue for such a proposal in "The Artist's Sanction in Contemporary Art," Journal of Aesthetics and Art
Criticism 63 (2005), 315-326.
Copyright  2007, Sherri Irvin 6
presented object, in conjunction with the relevant conventions operative, are not sufficient to
determine a particular set of artistic identifications.6
Such cases include the red square paintings
(TC 1) as well as Bruegel's Landscape with the Fall of Icarus, which, as Danto notes, would be
an altogether different work were it titled Ploughman by the Sea. (TC 117)
Must we see the artist's interpretation as constituting the artistic identifications that apply
in such cases? Often, the act of titling the work seems to play the constituting role. Assigning a
title, like applying paint to canvas, is certainly the outgrowth of the artist's interpretative activity.
However, it is not clear to me that we must see the title as an interpretation of the work, rather
than another feature of the work on a par with its visible appearance. And even if we do see
titles as interpretations, it is far from clear that they play a direct or incorrigible role in the
constitution of artistic identifications.7
When an artist titles her red square painting The
Israelites Crossing the Red Sea, adding the explanation that "The Israelites had already crossed
over, and the Egyptians were drowned," (TC 1) are we thereby compelled to see the
monochromatic expanse of red as a representational depiction of the surface of the Red Sea?
While this may be one interpretative possibility to entertain, I do not see that the facts Danto
invokes serve to close the case; nor, I think, would a very insistent statement to this effect by the
artist. The work might be as well or better understood as a satire on the notion that
monochromatic paintings can have as much depth or artistic significance as the epic
representational paintings of old. Whether or not it is such a satire will depend, at the very least,
on some facts in addition to the artist's interpretation; for artists can fail at creating satires just as
they can fail at depicting elm trees. And, of course, those with anti-intentionalist inclinations
will resist the suggestion that the artist's interpretation plays any meaning-constituting role here
at all; on their view, the work might be a satire even if the artist had had no satirical
6
I will here ignore the question whether there are relatively few or relatively many such cases, since I take it that an
adequate account will handle them regardless of quantity.
Copyright  2007, Sherri Irvin 7
interpretation in mind at all.
It is certainly true that, in titling the work and perhaps issuing brief interpretative
statements about it, the artist affects the work's nature and the range of interpretations
appropriate to it. We might, then, wish to say that the artist's interpretation has a limited role in
constituting certain meaning-related aspects of the work. But the interpretative activity by which
the artist does such constituting is not the same as the activities of making and presentation by
which the artist brings the work into existence and establishes the physical boundaries of its
form. Just as paint is applied to canvas in layers and over time, the constitution of an artwork
may occur in stages, some interpretative and others not.
Can we, in the end, preserve the general claim that artworks are constituted by
interpretation? Only, I think, by understanding `interpretation' very broadly, as including the
artist's actions of making an object, titling it, offering it for display, and perhaps issuing a
statement about its meaning. There is nothing objectionable about the claim that artworks are
constituted by interpretation, if this is what is meant; except, of course, that it is no longer clear
that the term `interpretation' is appropriate. If, as Danto suggests, an interpretation is a system of
artistic identifications or an account of the work's meaning, we should conclude that the direct
work-constituting role of an artist's interpretation is quite limited.
Ultimately, though, this may require simply an adjustment of terminology. The central
ideas that Danto brings to our attention in The Transfiguration of the Commonplace, that the
artwork is much more than a simple lump of physical matter and that the artist has special
authority in determining the work's nature, must be incorporated into any plausible account,
whether or not the artist's authority is taken to reside exclusively in the realm of interpretation.
7
Recall Danto's example, mentioned above, of the painting titled Old Man Planting Spring Cactuses.
Copyright  2007, Sherri Irvin 8