Art and Representation

This is an outline of Carroll, Philosophy of Art, Chapter 1

My personal comments are in red

Analyzing Concepts (from the book's Introduction) 

In philosophy, the word 'art' indicates a concept. One concept differs from another in covering a different category of things, and we assume that there are distinct criteria for being in one category than another. Finding these criteria is known as conceptual analysis. For example, "Norwegian" and "likes lutefisk" are different categories even if both are true of Sven Svenson. Because only some Norwegians like lutefisk, and some people who aren't Norwegian like lutefisk, Sven must belong to the category of Norwegians by meeting different criteria than Sven meets in belong to the category of those who like lutefisk. Likewise, "musical work" and "work of art" might both be true of Beethoven's Fifth Symphony, but some music isn't art and some art isn't music

With most concepts, a thing will have to meet SEVERAL criteria to fall into that category. Each of the individual criteria that it must meet are said to be NECESSARY CONDITIONS for being in the category. Once we have the total SET that we need, we have fulfilled the SUFFICIENT CONDITIONS for the concept. For example, the concept of fatherhood involves two necessary conditions, being male and being a parent, and those two conditions are together or jointly sufficient. We now have a conceptual analysis of "father." 

In defining art, philosophers hope to spell out the necessary and sufficient conditions for belonging to the same category as all the other things that we call "art." 

We TEST or EVALUATE a conceptual analysis by identifying some things that we already count as members of the category, and making sure that these examples really do satisfy all the criteria we have identified as sufficient conditions. If we find a work of art that fails to meet one of the necessary conditions, we know we've made a mistake. Our list is too exclusive (it excludes too many things). But we also need to identify things that do NOT belong in that category, and if we find one that satisfies our current list of sufficient conditions, we again know that we've made a mistake. Our list is too inclusive (it includes too many things). So conceptual analysis requires thinking about things that ARE covered by our concept as well as things that are NOT covered by our concept. While being a parent is a necessary condition for fatherhood, by itself it is it is too inclusive to be sufficient (it includes moms, but moms aren't fathers). And while being a mom is sufficient for being a parent, it is too exclusive to be necessary for parenthood (it excludes dads). It can be very difficult to find the necessary and sufficient conditions for a concept, and art is one of the most difficult and subject to debate.


Representation as a necessary condition for being art 

First Candidate for a definition of art: x is an artwork only if x is an imitation. It is important to notice that this proposal only stipulates a necessary condition for being art. It's like asking for a definition of "cat" and being told that cats are animals. It helps, but it's not intended to be a full definition. 

This captures the ideas of Plato and Aristotle, and captures the core idea of the first systematic definition of art (in the "modern" sense of "art"):

Batteux (1746). To be fine art, it is necessary that something be an imitation of something else.
For Batteux, that "something else" is beautiful nature (including imitations of events and attitudes).

According to Paul O. Kristeller, Batteux and his contemporaries were mainly thinking about five arts: painting, sculpture, architecture, music, and poetry. And when it came to music, Peter Kivy argues, they were thinking of music with words.

Modern abstract paintings demonstrate that this definition is incorrect. Prior to the 20th century, we might have used instrumental music as our counterexample. Literature is another problem case. These are, at best, representations rather than imitations.

But ARE these really counterexamples? Couldn't we instead conclude that these aren't art? Why should we assume that they are art?

Second candidate: x is an artwork only if x is a representation, where something is a representation if it is intended to stand for something besides itself and others can recognize that it does so. 

Architecture remains a counterexample. Churches do not "stand for" anything, yet they are works of art.

This counterexample is weak. First, churches might represent God's power and glory, and thus count as representations in the necessary sense. Second, Batteux did not count architecture as a fine art. He put it in a different category, that of artifacts "both useful and agreeable." (Public speaking is another case of this "mixed" category.) Why assume that architecture really is a fine art, and not, as Batteux suggests, a borderline case? 

Third Candidate: The neo-representational model: x is an artwork only if x has a subject about which it makes some comment. Notice that this lays down a requirement for being art, but it does not claim that satisfying this condition is sufficient for being an artwork. After all, many sentences on this page satisfy this condition, but these sentences are not works of art. 

The core idea of neo-representationalism is that art always has some semantic content (including expressive content: a work of art may denote something and express an emotion toward it, as when Picasso's Guernica expresses horror at the bombing of civilians in war).

Strong point in favor of the theory is that it shows why conceptual art is art (e.g., Duchamp's readymades -- they warrant interpretation).

BUT expressive music remains a problem. A piece of music might express sadness, but what does it express sadness ABOUT? There does not seem to be anything denoted here -- there's no SUBJECT toward which the music expresses itself. A similar problem arises for a lot of architecture.

DECORATIVE arts also pose a problem. A beautiful pattern often lacks both a subject and a comment on that subject. But why should we grant that the decorative arts are art?


The special case of pictorial representation

Even if being a representation is not a necessary condition for being an artwork, many artworks are representations. So what makes something a representation?

Four basic proposals:

  • Resemblance Theory
    x represents y if only if x significantly resembles the look of y.

    No good: two manufactured products are visually very similar, but neither represents the other. Also, the etching of George Washington on the dollar bill resembles George Washington, and vice versa, but we don't want to say that George represents his etching. So visual similarity is not sufficient.

    Second try: Visual design x pictorially represents y (which is not a visual design) if only if x significantly resembles the look of y.

    No good: a photo in an art history book is a visual design that both resembles and represents the painting that it pictures. "Tightening up" the definition excludes too much.

    FURTHERMORE, resemblance is not necessary for representation. Denotation represents without resembling, so resemblance is not necessary. For example, the word "cat" denotes cats, but the word "cats" in no way resembles cats.

  • Illusion Theory 
     x represents y if only if x causes the illusion that y is present.

    No good: who really thinks that x is y? And if we did, why would we appreciate x? If I thought I was looking at the real thing, what would I be appreciating?

  • Conventionalist or Semiotic Account
    Visual design x pictorially represents y if only if x denotes y in accordance with some established system of conventions. In other words, representation requires a conventional visual "language." Resemblance is a matter of being familiar with the governing conventions.

    No good: If it really were just a matter of convention, then new visual techniques would seem less realistic (because unfamiliar), not more. Furthermore, even very familiar styles (e.g., cubism) never seem realistic.

  • Neo-Naturalist Account (emphasizes recognition but not the deception of illusion)
    Visual design x pictorially represents y if only if (1) x is intended to be recognized as featuring y in x by looking and (2) it is successfully recognized in this way and (3) x is intended to denote y, and (4) relevant viewers recognize that x denotes y.
    This theory combines requirements for both visual recognition and denotation.

    HOW do we recognize x as y? Perhaps by resemblance! But that's for psychologists to determine.

    Go back to the argument given against Resemblance Theory: the objection about denoting without representing depends on a false assumption if we remind ourselves that we are talking about PICTORIAL representation. "Cat" denotes cats, but it doesn't pictorially do so.

    This theory captures the fact that some people recognize what a picture denotes the very first time they see the picture, even if they've seen no other pictures in that style.


Non-pictorial representation

How do we have representation in non-pictorial cases? For example, music and literature?

No single thing seems common to all cases. There appear to be four types of representation:

  1. Unconditional (requires no prior agreements or conventions)
    What we've already endorsed with pictorial representation: some natural capacity is exploited to allow us to recognize x in y.

  2. Lexical (requires a pre-established conventional code)
    What the conventionalist or semiotic approach claimed was generally the case is sometimes the case: certain movements in ballet, or the halos above the heads of saints in pictures.

  3. Conditional Specific
    Only succeeds if the audience already knows, by other means, what is being communicated. (Basically, you can "see" x in y after you're told to look for x in y.)

  4. Conditional generic
    Only succeeds if the audience is looking for a denotation via representation. (For example, in a game of charades, you're looking for x in y without knowing what x is.)

                        Last updated Jan. 12, 2008            Who made this page?