the classificatory sense of “art”

The Journal of Aesthetics and Art Criticism 61:2 Spring 2003

UNCORREC

TED P

ROO

F

LAUREN TILLINGHAST

The Classificatory Sense of “Art”

The first question with which philosophers of art must deal is “What are the limits of making?” not “What arethe limits of the use of ‘art’ and “work of art?”

George Dickie, The Art Circle

I

A student of the main debates of analytic phil-osophy of art throughout the latter half of thetwentieth century might well be struck by whatis no doubt a scandalous element in one of themost influential theories of the period, GeorgeDickie’s brand of institutionalism. Dickie’s theoryhas long faced a pair of objections: Either thetheory is circular or it is not sufficient to distin-guish artworks from what are not artworks.The circularity is obvious; few critics commenton it, preferring to head straight to the problemof insufficiency. Noël Carroll, for example,remarks, “There is a real question whetherthe . . . institutional theory is really a theory ofart. For the . . . set of definitions, though men-tioning ‘art’ at crucial points, could be filled justas easily with the names of other . . . practices,like philosophy and wisecracking.” If Dickie’sclaim is simply that art is some sort of practice,continues Carroll, “then he should give uptalking about defining art. For he is no longerplaying that game according to its originalrules.”1

That Dickie’s account appears caught betweencircularity and insufficiency is not the scandal,of course. This pair of criticisms has been lev-eled against familiar strands of functionalism,historicism, expressivism, and intentionalism.The scandalous feature is that Dickie flaunts thecircle.2 Although he has refined the theoryextensively since first introducing it some three

decades ago, and although he is highly respon-sive to criticism, he continues to argue thathis set of definitions is both necessary andsufficient to characterize “art” in the “classifica-tory” sense, that neither the circularity nor theuninformative character of the theory is adefect, and anyway, he never promised to playthe game of defining “art” according to itsoriginal rules.3

I will argue that the root of the scandalouselement of Dickie’s institutionalism is that hisown considerations cast serious doubt on anassumption so basic that it never occurs to him,or to his critics, to reject it. The assumption I amtalking about is simply that the classificatorysense of “art” applies to some kinds of artifactsbut not to others. I will urge that this assump-tion is false. Once we have a proper understand-ing of the relevant notion of an institution, andonce we clarify what work the classificatorysense of “art” is to do, the real strength ofinstitutionalism is that the classificatory senseof “art” is the ultra-abstract concept of an insti-tutional kind.

In section II, I make some preliminaryremarks about the parameters of my argument.In section III, after sketching Dickie’s theory,I distinguish the three possible noncircularversions of institutionalism. In sections IV–VIII,I argue that the most plausible version ofinstitutionalism supports the view that allartificial kinds are art, in the classificatorysense.

art003.fm 33 Wednesday, February 26, 2003 5:03 PM

134 The Journal of Aesthetics and Art Criticism

UNCORREC

TED P

ROO

F

II

Before introducing the set of definitions that Dickiecalls the institutional theory of art, I want tomake three brief remarks about how I will proceed.

The first point to mention is that I will focusalmost exclusively on Dickie’s theory as it appearsin The Art Circle, as Dickie has recently saidthis is the best version of his view.4 BecauseI intend to read some of its arguments againstone of its most basic assumptions, it is impor-tant to start with the most compelling expres-sion of the view. Since that is a rather tall order,I hope it is enough that Dickie, in company withmost of his critics, takes the arguments put forthin The Art Circle to be the most successfulsupport of institutionalism.

The other two preliminary points concern theparameters of the discussion, rather than oursource material. I want to stress that the ques-tion I am raising is about the classificatorysense of “art.” Dickie is the philosopher whohas made the most of the distinction betweena “classificatory” sense of “art” and any otherpossible sense. His own remarks about theclassificatory sense are problematic, however,and I will discuss the difficulties in depth. Asa tentative gloss, we might say that the classi-ficatory sense of “art” expresses the ordinaryconcept of an art form. If that is accurate, thenI can illustrate my central thesis by saying thatwhile I do not dispute that various forms ofpainting, sculpture, music, theater, and writingare all art forms, I take issue with the assump-tion that wisecracking and philosophy, forexample, are not art forms.5 The final prelim-inary point is that to give Dickie a fair hearing,I will not enter into the controversies surround-ing the claim that institutionalism affords atleast a necessary condition of art. Aside fromexplaining what an “institution” is for Dickie,I will limit my discussion to the question ofwhether institutionalism gives us only a neces-sary condition or yields a condition that is bothnecessary and sufficient to mark out all andonly art, in the classificatory sense.

III

As philosophical theories of art go, Dickie’sinstitutionalism is very widely known. But it is

not so widely known that it is safe to assumethat everyone is able to recite it by heart. Thus,before I attempt to reconstruct the theory, it willbe useful to have it clearly before us.

The central tenet of an institutional theory ofart is, according to Dickie, the claim “thata work of art is art because of the position itoccupies within a cultural practice.”6 Notice theambiguity in Dickie’s choice of phrase. Thecentral tenet could be that there is some culturalpractice, a metapractice perhaps, in virtue ofwhich some things are art. One is likely to thinkDickie must be articulating some version of thisclaim. Indeed, Dickie says explicitly that hethinks there is some specific practice rightlycalled “the practice of art.”7 But, given thephrasing of the central tenet, the claim could bethat “art,” in the classificatory sense, is analyz-able as: that which is what it is in virtue of theposition it occupies within a cultural practice.I do not think that the ambiguity is due tocarelessness.

Dickie’s full theory does not resolve theambiguity of his formulation of institutionalism’scentral tenet. For when it comes to describingthe practice “in virtue of which a work of art isart,” we get the famous (or infamous) circularaccount:

1. An artist is a person who participates withunderstanding in the making of a work of art.

2. A work of art is an artifact of a kind createdto be presented to an artworld public.

3. An artworld system is a framework for thepresentation of a work of art by an artist to anartworld public.

4. The artworld is the totality of all artworldsystems.

5. A public is a set of persons the members ofwhich are prepared to some degree to under-stand an object which is presented to them.8

Before taking up the circularity and my diagnosisof its source, it is important to see that the firstthree definitions bear the weight of the theory.9

The first three definitions—those defining thenotions of artist, work of art, and artworldsystem—form a sort of solid block. The fourthand fifth definitions are not part of the block.The fourth definition simply gives us a name forthat set of actual systems, described by the firstthree definitions that “exist.” Metaphorically

art003.fm Page 134 Wednesday, February 26, 2003 5:03 PM

Tillinghast The Classificatory Sense of “Art” 135

UNCORREC

TED P

ROO

F

speaking, while the first three definitions con-cern the bricks, the fourth concerns the house(or wall or pile or heap) composed of the bricks.Thus, the usefulness of the notion of the art-world depends entirely on the cogency of thefirst three definitions of the theory. If all institu-tional kinds count as artworld systems, then allinstitutional kinds belong to the artworld. If thenotion of an artworld system is empty, then sois the notion of the artworld.

In contrast to the definition of the artworld,the definition of a public is independent of thefirst three. Dickie’s use of the complex expres-sion, “artworld public,” in the second and thirddefinitions shows that, as he is using the word,there are different kinds of publics, but no suchthing as a public simpliciter.10 For example,there are mathematical proof publics and carpublics and philosophical doctrine publics, aswell as Impressionist painting publics and Gothicarchitecture publics. Following J. L. Austin, wemight put the point by saying that the notion ofa public, as it occurs in the institutional theoryof art, is substantive-hungry.11 The definition ofa public as such, therefore, might be recast as:

For any kind of object, K, a K-public is a setof persons the members of which are preparedto some extent to understand a K that ispresented to them. The virtue of reformulatingthe definition of a public in this way is that itreveals that the notion of an artworld publicdepends for its cogency on the first threedefinitions as much as the notion of the artworlddoes.

Thus, the first three definitions carry theweight of the institutional theory of art. Togetherthey comprise the theory Dickie claims to bea necessary and sufficient characterization ofthe classificatory sense of “art.” The defect ofthe theory is obvious, for the three definitionsrun in a tight circle. The word “art” is broughtin at each turn to indicate the relevant kind ofwork, kind of maker, kind of public, and kindof system. It is clear, then, that at least part ofCarroll’s objection hits the nail right on thehead. It is absolutely true that the word “art”can be replaced by the names of any number ofpractices. But Dickie does not deny that one canreplace the word “art” with the name of anypractice one likes. He agrees that if one replacesthe word “art” with the name of some (nonart)practice, the theory ceases to be a theory of art.

His point is rather that there is no noncir-cular formulation of what it is to be a practiceof art—the circular account is the only account.12

This is the scandal.Dickie is not blind to the difficulty of taking

his position seriously. He offers a number ofarguments designed to show there is really noobjection to circular definitions of practices.13

While these arguments are interesting, I will notdiscuss them here. I am not asking whether thecircle can, after all, be defended. Granting thetraditional line that it cannot, the questionremains: What does Dickie see, or think he sees,that makes it seem that a circular account is theonly account that will do?

If one rejects circularity, three options remainto the institutionalist. Since the institutionalistabout art holds that institutionalism affords atleast a necessary condition of art—that is, thatall art forms are practices—the first point ofpossible divergence concerns whether all prac-tices are practices of art or not.—Let us call thethesis that all practices are practices of art“weak” institutionalism. This version holds thatDickie’s three definitions comprise an adequateimplicit definition of an art form. Such a theoryneed not be circular, but it will be uninforma-tive in precisely the way in which a formalaccount of a practice as such is uninformative;it will of course tell us nothing about the goalsor standards of any specific art form, nothingthat would, say, allow us to appreciate or createobjects of that form. This is the weak versionbecause, according to it, painting and verse areart forms merely in virtue of being institutionalkinds.14 Unlike weak institutionalism, the othertwo versions of institutionalism cannot happilyallow the theory of art to be uninformative. Ifnot all practices are practices of art, then atheory of art needs to offer some account of thedistinction between an art practice and a nonartpractice. One way to account for this wouldbe to hold that for a practice to be an artpractice, it must be embedded in a furtherspecific practice, the practice of art. Alterna-tively, one might hold that there is no such thingas the practice of art, in which case one musthold that the distinction between art practicesand nonart practices relies on some other sort ofconsideration. Let us call the thesis that there issome specific practice in virtue of which all artforms are art forms “strong” institutionalism.



UNCORREC

TED P

ROO

F

On this view, Dickie’s three definitions refer tothis institution. And, still supposing circularityis not an option, the strong institutionalist owesus some noncircular, informative account of thisinstitution.15 This leaves the thesis that thedistinction between art practices and nonartpractices must be explained in some way otherthan by appeal to the specific practice of art. Letus call this view “partial” institutionalism,because it holds that Dickie’s three definitionsare neither a sufficient implicit account nor asufficient ostensive account. At best, talk ofinstitutions gets us only part of the story aboutart. Of course, among partial institutionalists,the issue of what consideration or element isneeded to draw the distinction between art prac-tices and nonart practices is a matter of somedisagreement.16

IV

Now that we have isolated the three possiblenoncircular versions of institutionalism, I canstate my central thesis more precisely. Mycontention is that weak institutionalism makesthe best sense as a theory of art, in the classifi-catory sense, and that some of Dickie’s ownconsiderations show this.

I aim to loosen the grip of the assumption thatweak institutionalism is obviously false. To dothis, we need to start the argument for institu-tionalism more or less from scratch: Let us askwhat recommends any version of institutional-ism. The answer is that the “institutions” and“cultural practices” at work in Dickie’s institu-tionalism are simply artificial kinds.

In The Art Circle, Dickie is careful to statethat in arguing for institutionalism, it is theconcept of an action-institution that he has inmind.17 An action-institution is any type of actthat is governed by rules, and is such that under-standing the rules is a condition of performingan act of that type.18 Promising is a clear case.The “rules” of promising are simply what definesthe practice as such; among other things, theremust be a special sign in a language that meansgivers of the sign have put a limitation on theirfuture possibilities of acting well. We can seehow promising is governed by these rules byimagining a word that is pronounced and spelledthe same, but which has a different set of rules

associated with it. Suppose saying “I promise tosuch-and-such” is associated with the rule thatthe giver of the sign wishes that such-and-suchwould come true. If this were the rule, saying “Ipromise . . .” would not be making a promise.Furthermore, someone who does not understandthe (real) rules of promising does not have theconcept of a promise; hence they cannot makea promise, although they might easily and inten-tionally utter the words, “I promise” followedby a description of a future act that is in theirpower to do. In the absence of an understandingof what “I promise to . . .” means, saying thewords is merely the appearance of making apromise, analogous to saying the words in one’ssleep or while delirious.19 Thus, the centraltenet of institutionalism is unambiguously trueof promises, that is, a particular promise is apromise because of the position it occupieswithin a cultural practice.

What has been said so far will probablymeet with general approval, if only becausepromising is one of the poster children forphilosophical work on practices. What may beharder to accept is that, if the account of anaction-institution is to accommodate cases thatdo not include vows (i.e., not simply betting,promising, marrying, etc.), then pretty much thewhole range of acts of making kinds of artifactsare action-institutions. Dickie does not actuallydraw this inference, although I do not think hewould object to it. He explicitly mentionsmarriage, tool making, and storytelling asexamples of action-institutions.20 Let us con-sider how tool making turns out to be an action-institution.

It is probably misleading to say that toolmaking is an institution. The claim is moreclearly put by saying that all kinds of tools areaction-institutional kinds. That is to say, all toolmaking actions, for example, making a roofinghammer or a socket wrench, are governed byrules, some understanding of which is a condi-tion of making a hammer or a wrench. To seethis, imagine a culture rather different fromours, one in which people make objects thatlook (feel, etc.) exactly like some of our chairsand are made of the same materials from whichchairs are made. Suppose, though, that theseobjects are understood to be targets in an elabo-rate game these people play. (The gameinvolves throwing delicate glass objects onto



UNCORREC

TED P

ROO

F

the surface from some distance away, say.)Now, these objects are not chairs; they aretargets in some game. The acts involved inmaking them are not correctly described as“carving the seat,” “forging the arms,” and soforth. The acts are correctly described as, say,“carving the target,” “forging the foul posts,”and so forth. Moreover, we can speak of “rules”governing the acts of making chairs and makingthe targets. In both cases, certain things count asmaking the device poorly or improperly.Constructing the legs so that the device wobblesis contrary to one of the points of having chairs,hence, against the “rules,” but wobbling may bequite appropriate, even required of the targets.Thus, it should at least be clear that whether thething that looks like a chair is a chair brings inconsiderations of a wider reach than the phys-ical or phenomenal properties of the object atissue; just as with particular promises, particularchairs are what they are in virtue of beingembedded in a “rule” containing practicalcontext.

But now we face a difficulty. While it holdsfor promises (and vows of all varieties) thatsome understanding of the rules of promisingpresupposes any individual agent’s act beinga promise, the same does not seem to holdfor tool making and chair making and the rest.Imagine a factory in which individual workersparticipate in making some complex device,such as a rocket ship, none of whom understandin the least what sort of device it is. Eachperson’s task is, say, to push a button at statedintervals. Cases such as these are possible;hence, it is possible for an individual person tomake a tool without understanding the “rules”that describe the sort of object being made. Thisis not to allow that the factory workers under-stand nothing of what they are doing. They arefollowing orders, and so must understand theorders they are following. The point is that theydo not have an understanding of the widerpractice in virtue of which, say, “pushing thered button” amounts to “attaching the door tothe body.”

This difference between promising and toolmaking is important, but it is not enough toshow that Dickie is wrong to begin with thethought that action-institutions are central to atheory of art. What the case just considereddoes show is that the second criterion of an

action-institution—performing an act of therelevant type presupposes some understandingof the rules of the practice—is not a thoroughlyindividualistic condition. The essential point isthat the source of the chain of orders, if there isany chain, is someone who understands thepractice at issue. Dickie’s formulation is a bitmisleading, but the claim that tool making isaction-institutional is not misguided.

To see that an agent’s understanding cannotbe omitted entirely, consider a case in which themaker has no understanding of the kind ofdevice being made and is not following some-body else’s instructions. By hypothesis, thisperson has no idea about, say, chairs, and makesan object that looks just like a chair and ischemically or molecularly identical to somechair (or as identical as any two chairs can be).Has this person made a chair? The answerwould be obvious if we build into the thought-experiment that this person means to makesome other specific sort of thing. If they meanto make one of those targets or if what lookslike the seat is really the base of a podium orreally the beginnings of a roof, then, clearly,they are not making a chair. But to build insome alternative understanding on the part ofthe maker misses the point of the objection; itgrants institutionalists their claim that someunderstanding of the wider practical contextmust be in place.

A good way to address the objection head-onis to suppose for the sake of argument thata person can be making a chair without the leastunderstanding of what a chair is. If they aremaking a chair, then what they are makingcannot depend on their intention or understand-ing in any way. But if what they are makingdoes not depend on their intention or under-standing in any way, why not just as easilysuppose that they are, unwittingly, making any-thing of which they have no concept? Whychoose chairs specifically? Why not a rocketship, or a thermonuclear reactor, or a perpetualmotion machine? The only reason seems to bethat the object is hypothesized to look just likea chair. But that cannot really matter—a fertileimagination can supply any number of devices,some instances of which could look just likea chair. Thus, the objection to holding thatsomebody is making a chair, while the contextcontains no one who grasps the concept of a



UNCORREC

TED P

ROO

F

chair, is that there is no ground to fix on “chair”as the relevant description of what they aremaking. To put the point another way, there isequal ground to fix on any one of any numberof different specific kinds of thing as the type ofthing made. This is the sense in which toolmaking presupposes an understanding of thepractices that inform tools. To say that toolmaking is an action-institution is, therefore, tosay that tools are kinds of things such that someunderstanding of what belongs to tools is logic-ally prior to the existence of any particulartools. Of course, there is nothing special abouttools in this regard; the same is true of prettymuch any kind of artifact. Therefore, the notionof institution at work in the institutional theoryof art is basically the notion of an artificial, orinvented, kind.21

Before moving on, I want to point out thatthis analysis of the sort of institutions at work ininstitutionalism throws an interesting light onone of Dickie’s central arguments for institu-tionalism, the celebrated “visually indistinguish-able objects” argument, which Dickie takesover from Arthur Danto.22 The argument, recall,supposes a pair of visually indistinguishableobjects, one of which is a minimalist paintingand the other of which is a primed canvas, orone of which is a sculpture and the other ofwhich is a package of scouring pads.23 Theargument concludes that being a minimalistpainting or piece of sculpture is not a visuallydiscernible property of an object. This is cor-rect. And the same point holds for hammers,wrenches, shirts, chairs, games, houses, menuitems, forms of entertainment, and so on. In allof these cases, we can imagine two objects thatlook the same, are chemically or molecularlythe same, but are not the same sort of thingbecause they are not bound up in and informedby the same practice. Thus, the “visuallyindistinguishable objects” argument does notshow there is anything special or peculiar aboutpaintings and sculptures or the practices inform-ing them that distinguishes them from all otherartificial forms. At most it shows that thevarious kinds of painting, sculpture, prose,poem, and so on are some among the hugerange of kinds that presuppose human under-standing, that is, the range of things that arewhat they are in virtue of being embedded in acultural practice. The visually indiscernible

object argument simply trades on a conceptualtruth about the whole class of artificial kinds.

V

Clarity about what work the notion of an institu-tion does in institutionalism goes a long way tosupport some form of institutionalism. To denyeven partial institutionalism would be to denythat only artificial kinds can be art forms.Clarity about the notion of an institution alsoshows that neither the concept of an institutionnor the visually indistinguishable object argu-ment provides grounds to hold a strong orpartial institutionalist account of art as opposedto the weak institutionalist account of art.Of course, nothing said so far obviously favorsa weak account either. To make some progresson the question of which form of institutional-ism provides the most promising theory of art, itis necessary to insist on being precise aboutsome thoroughly foundational issues: Institu-tionalism is meant to be a theory of art, in theclassificatory sense. So, before we ask whichtheory best describes what art, in this sense, is,we would do well to make sure there is noconfusion in granting that there is such a senseof “art.”

Some philosophers have thought that weought to reject any theory of the classificatorysense of “art” because the very idea of a classi-ficatory sense of “art” is problematic. There issome ground for this charge, for Dickie’s onlyexplicit argument that there is a “classificatory”sense of “art” about which there is a philosoph-ical puzzle is obscure. His argument turns onthe claim that “we” say important things with thephrases “good art” and “bad art.”24 This claimraises a number of questions, quite apart fromwhether anybody other than philosophers usesthese phrases in any serious way.25 Suppose it istrue that these phrases are commonly used tosay important things. Even so, it does not followthat “good art” shows us what they are used tosay. As Peter Geach pointed out long ago, notevery phrase with “good” in it expressesa determinate sense.26 If someone says, in thenormal course of life, “What a good thing it’sraining!” they could be making any number ofdifferent statements. Perhaps the thought is thatit is a fortunate change in the weather for a



UNCORREC

TED P

ROO

F

person who does not want to go to the collegepicnic but who also does not want to makeexcuses about not going. Alternatively, the ideamight be that one’s friend now stands a goodchance of winning some money at the racetrack,since they have bet on a mudder. Whatevermight be meant, it is not that rain is here isa good feature for things to have, in the waythat a sharp blade is a good feature for chef’sknives to have.

What I think Dickie has in mind by invokingthe phrases “good art” and “bad art” as beingused to say very important things (which it isopen to him to have in mind, anyway) is theobvious fact that the phrases “good sonnet,”“bad portrait,” and “mediocre horror movie” areused to say things that any theory of art hadbetter take seriously. But, of course, these factsdo not show there is any special classificatorysense of “work of art.”

It is very important that Dickie’s mostcommon way of setting out the classificatorysense is not by deploying the idea that we sayimportant things with the phrases “good art”and “bad art.” Mostly, he simply gives a list:The classificatory sense of “art” expresses aconcept that ranges over “all paintings, statues,vases, buildings, and such,” or, all “paintings,poems, plays, and the like,” or again, “the oldfamiliar class of paintings, poems, musicalpieces, and the like.”27 If we take remarks suchas these as definitive, it follows immediatelythat we are not working with any sense of “art”according to which a particular object has to bea good something or other in order to be a workof art. Thus, if there is a sense of “art” thatranges over only good paintings, good statues,good buildings, vases, poems, plays, “and thelike,” this sense is not the classificatory sense.

We do not have everything we need to justifytalking about the classificatory sense of “art”yet, but we do have enough to show that theclaim that there is a classificatory sense of “art”does not rest on any very bold, contentious, orespecially fishy distinction between classifica-tion and evaluation. Thus, a certain group ofobjections are simply irrelevant. For example, ithas been argued against Dickie that sometimesevaluation is part and parcel of classification.28

Perhaps it is true that a piece of pocket lintcannot be classified as a can-opener, not even avery, very, very defective can-opener, because

the pocket lint does not meet even the mostminimal standard of success at opening cans. Ifthe same point holds for the range of kinds atissue, it follows that some evaluation is requiredto classify something as a painting, vase, build-ing, and the like. If so, Dickie is wrong when hesays that being a work of art in the classificatorysense does not guarantee any value at all.29 Butas long as it is possible for a portrait to be apretty miserable, pathetic excuse for a portrait,there is no serious problem.

We are also in a position to sidestep a differ-ent worry about any purported classificatorysense of “art.” It has recently been pointed outthat it does not follow from the fact that we canmake sense of “good F” and “bad F,” that “F”must be something other than an evaluativeterm.30 For example, it does not follow from thefact that we understand “the cat is in goodhealth” and “the cat is in bad health,” that“health” is not an evaluative term, for “health”means “the sound working state of an organ-ism.” This insight about “health” would posea problem for those of us interested in the clas-sificatory sense of “art” if we depended on anexclusive distinction between classificatoryexpressions and evaluative expressions to markout the special sense of “art.” However, on theaccount we have before us, the claim is not that“art” must be something other than an evalu-ative term because we can talk about good andbad art. The claim is that there is a sense of“work of art” such that if any sonnets, portraits,or vases are works of art, then all sonnets,portraits, and vases are works of art, the bril-liant, the sublime, the mediocre, the stupid, andthe incompetent ones equally.

So far, then, there is no reason to hold thatthe classificatory sense of “art” is based inconfusion from the very beginning. It has so farbeen explained as a sense of the word “art” thatswallows kinds whole, and is restricted in itsrange to kinds that provide criteria for judg-ments about the goodness or defect of particularinstances. This leaves us free to say that inwhatever way evaluation and classification areinvolved in grasping the concepts of the kinds,that will be how evaluation and classificationare presupposed by the classificatory sense of“art.” At this point, we may as well stipulate toa third thesis about the classificatory sense of“art”—the undisputed shared doctrine of all



UNCORREC

TED P

ROO

F

three versions of institutionalism: the range ofkinds that the classificatory sense of “art” swal-lows whole is further restricted to kinds forwhich some understanding of what belongs tothe kind is conceptually prior to any instance ofthe kind.31 For example, species of animal orplant, although arguably kinds that providecriteria for judgments of goodness and defect ofparticular organisms, do not belong to theextension of the classificatory sense of “art,”because they are not action-institutional kinds.

VI

If the considerations of the previous section aresound, the idea of a classificatory sense of “art”is not based on any obvious confusion. We havea sense of the word “art” that ranges overinstitutional kinds, includes all instances of anyof the kinds within its extension, and that so farhas been indicated only by an open-ended list. Itis possible, therefore, that there is some genuineand interesting question at issue for those whowork on the classificatory sense of “art.” Andhere we come to the central difficulty. Confu-sion about the question stands firmly in the wayof seeing that two very different understandingsof what is to be explained are colliding inDickie’s institutionalism.

Let us start again from Dickie’s list of kindsthat the classificatory sense of “art” swallowswhole. These are: all forms of painting, poetry,vase, building, and the like. Now, what exactlyis the puzzle? It is no good to say that these areall art forms and we want to know what an artform is. This answer forgets that we are tryingto come up with a way to characterize thequestion of what an art form is that does notleave out the articulation of the problem. We donot get much further by saying that the puzzle isto set out some sort of principle governing whatbelongs on this list and what does not. Justfaced with the list, and the question, “What elsegoes on it?” there are any number of possibleanswers. Paintings, poems, vases, and buildingsare all things that some people like to spendtheir time looking at and thinking about,particularly people with a lot of leisure time.They are all things that some people make fora living. They are all things that some peoplemake for pleasure. They are all things that have

been made, in one form or another, for a verylong time. They are all things that have servedmany different purposes—some important, suchas protection from the weather and telling thestory of a peoples’ history, some not veryimportant, such as filling up an empty space inthe living room. They are all things that can starin a “visually indistinguishable object” argument.They are all things that have been examples ina great long argument about the nature of art.Any of these answers would do if the questionwere simply, “What do all these things have incommon?”

So again, what is the puzzle? It is difficultto answer this question on Dickie’s behalfprecisely because he runs two rather differentpuzzles together.

The familiar puzzle about the classificatorysense of “art” comes out when Dickie is expli-citly distinguishing the classificatory sense of“art” from a “derivative” sense. The contrasthe draws between the two makes it clear thathe has some sort of fairly traditional distinctionin mind between arts and crafts. The “deriva-tive” sense is supposed to apply to, say, cars andpieces of furniture.32 Elsewhere, Dickie men-tions other examples of artificial kinds that donot make it into the extension of the classifica-tory sense of “art”: religious activities and theobjects belonging to them, athletic activities,fashion shows, stage props, stage sets, andplaybills.33 What is crucial in all of this is thatDickie gives no argument at all that these kindsdo not belong to the class of art forms. Noargument is needed; these examples are alloffered as definitive of the very sense of “art” atissue. Now, suppose we accept the examples assetting out the very sense of “art” at issue. Thepuzzle about the classificatory sense of “art”becomes a puzzle about what all the variousforms of painting, verse, theatre, music, build-ing, statue, and vase necessarily have incommon that they do not share with any kind ofclothing or furniture, or with sports, stage sets,cars, altars, sermons, or hymns. There is certainlya question here. And it certainly does followimmediately, and from the very sense of “art” atissue, that only strong and partial institution-alisms are possible answers. Weak institution-alism—the view that kinds of art are kinds of artin virtue of being action-institutional kinds—is clearly ruled out by the fact that kinds of



UNCORREC

TED P

ROO

F

clothing and furniture are also action-institutionalkinds. However, before we conclude that wehave elucidated the puzzle at the heart ofinstitutionalism, notice two points. Firt, it is notat all clear why this puzzle should be especiallyengaging to a philosopher. And this is so, inpart, because of the second point; the fact thatstrong and partial institutionalism are the onlytwo possible institutional theories of this senseof “art” does not imply that one of these twotheories is the truth about the favored list ofkinds. For the truth may be that there is nothingthat the kinds on the “art”-list necessarily havein common that distinguishes them from thegroup of kinds on the “nonart” list. If the ques-tion plays out in the latter way, we would notnecessarily end up with a circular account. Theclassificatory sense of “art” would simply be aname for a stipulated list of artificial kinds—and there would be nothing left about which toargue.

I do not believe that the puzzle outlinedabove is the one that is really motivatingDickie. I will state my grounds for this claim inthe next section. For now, I want only to showthat a different puzzle does in fact appear inDickie’s work. It comes out most clearly in thecourse of a discussion about when it becomesintelligible to speak of a culture as having someart or other:

Art may have emerged (and no doubt did emerge) inan evolutionary way out of the techniques originallyassociated with religious, magical, or other activities.In the beginning these techniques would have beenno doubt minimal and their products (diagrams,chants, and the like) crude and in themselves uninter-esting. With the passage of time the techniqueswould have become more polished and specialists[would] have come to exist and their products wouldhave come to have characteristics of some interest (totheir creators and others) over and above the interestthey had as elements in the religious or whateverother kind of activity in which they were embedded.At about this point it becomes meaningful to say thatprimitive art had begun to exist, although the peoplewho had the art might not yet have had a word forits art.34

While this passage contains a number of intri-guing suggestions that we cannot pursue here, italso contains one of the most useful sketches

I have found of one of the most difficultproblems of the philosophy of art. To see it, letus put aside the issue of whether evolution isa fruitful way to think about the emergence ofart. Let us also pass by the stress on “primitive,”except to notice that one should not connect ittoo tightly to the fact that Dickie mentions onlyreligious and magical activities as examples ofpractical sources of art, as if he is claiming thatthe practical sources of art have always to dowith something mystical or unearthly. ThatDickie is not making such a claim is clear fromthe fact that he adds, “or other activities”—the practical sources of art might have to dowith such mundane matters as the various prep-arations of food and shelter, the making oftools, or the forms of social interaction. Puttingall these matters aside, the central claim in thepassage is basically that to speak meaningfullyof a culture as having some art or other, it isboth necessary and sufficient that the peoplehave some techniques for making some sorts ofthings, such that some of the products of thetechniques—for example, some individualshelters, rites, or tools—have “characteristics ofinterest over and above the interest” the prod-ucts have qua shelter, rite, or type of tool.35

While the description of a particular objectas having some characteristics of interest overand above the interest it has as an element inthe practice in which it is embedded is clearlyof critical importance, it is also somewhatobscure. To make a start on clarifying it, noticethat to say of a particular object that it hascharacteristics of interest as an element in thekind of activity in which it is embedded isanother way to say that the object meets thestandards or norms of its (action-institutional)kind. For example, to say that a particularbottle opener “has characteristics of interest asan element in the activity of opening bottles”amounts to saying that it has the characteristicsof a good, or decent, bottle opener. When weturn to the more complex description of aparticular artifact as having characteristics ofinterest “over and above” the interest it has asan object of a specific artificial kind, we woulddo well to follow the same pattern, and replacetalk of “characteristics of some interest (totheir creators or others)” with talk about good-ness. For we do not mean to include all casesin which an artifact happens to have some



UNCORREC

TED P

ROO

F

characteristic that happens to be of interest tosomeone. If that were the whole point, wewould have a very contentious account of artindeed.

That the whole range of cases in which someartifact has some characteristic of interest includestoo much becomes clear if we consider a fewexamples: On the one hand, we have a bottleopener, say, the one designed in 1997 by ErikMagnussen and manufactured by Stelton. Thisopener is an elliptical piece of heavy gaugestainless steel, folded in half long-ways; thefold is rounded so that one’s thumb rests alongthe interior of the fold while one’s fingers restalong the exterior. The mouth is cut out of thefold and positioned far enough toward the topso that one can open a bottle with a very smallamount of effort. The steel is polished to a highshine and has no ornamentation, allowing thesteel itself and the opener’s utterly simple andelegant shape to be the focus of attention. Thisis a clear example of at least one type of case inwhich it would be natural to say that this bottleopener has “characteristics of interest over andabove the interest” a bottle opener has as abottle opener; I hope it is also a clear exampleof what seems right about the thought thatworks of art are those objects that have charac-teristics of interest over and above the interestthey have as instances of some artificial kind.

In contrast to the Magnussen opener, whichillustrates what is right about invoking“characteristics of interest over and above” thecharacteristics required of a good opener, thenext two cases illustrate what is problematic.First, imagine an ordinary bottle opener ofthe sort that one finds in thousands of kitchendrawers, only this one has a blue plastic handle.Imagine also somebody who just really likes thecolor blue. Given a choice, this person willprefer a blue house, a blue car, blue clothing,blue paintings in the living room, and a bluebottle opener. It is clear that it would be verystrange to take such a person’s fondness for thecolor blue as evidence that being blue is a char-acteristic that makes a bottle opener a good ordecent bottle opener. All the same, this personis interested in blue things, and the bottleopener is blue, so the bottle opener has a char-acteristic of interest to this person. May we notsay that it is a characteristic that interests thatperson over and above the opener being a good

or decent opener? While there would seem to benothing wrong with describing this person’staste for blue things in this way, it is clear thatthis sort of case is very far from providinggrounds to say that the bottle opener with theblue handle is a work of art. The reason is thatthe example gives us no way to see this person’sinterest as having any nonaccidental connectionto the goodness of any of the objects in whichshe or he has an interest. That is, we lack eventhe beginnings of a way to understand how thecolor of the handle is somehow apt, or ground-breaking, or thoughtful, or traditional, or har-monious, or funny, etc.

The fact that people’s tastes can be idiosyn-cratic is one problem for a plain “characteristicof interest” theory. But it is not the only prob-lem. Imagine someone who collects bottleopeners, and so has an interest in some bottleopeners because they are rare, old, expensive,strange, or simply not already in the collection.The collector may also have an interest in char-acteristics of bottle openers that have nothing todo with whether or not the bottle openers are inany way good openers, may even prize someespecially awkward or dangerous openers forthe characteristics that make them awkward ordangerous. Unlike the first type of case, thecharacteristics of interest to the collector neednot be entirely a matter of taste. Indeed, theremay be rather definite standards of goodness inplace for collecting; collecting, after all, is notmerely accumulating. It is a cultural practiceformally on a par with promising and toolmaking. Here again, while it would seem unob-jectionable to say that the bottle openers thatinterest the collector have characteristics ofinterest to someone over and above the interestthe openers have as openers, it would becontentious, to say the least, to hold that a bottleopener that has characteristics of interest to thecollector is thereby a work of art. The problemthis time is not that there are no standards ofgoodness in place. Rather, it would seem to bethat the standards of goodness the collector isusing are themselves only accidentally connectedto the standards of goodness that belong tobottle openers. It is because we have again lostthe sense of a nonaccidental link between thecharacteristics that are of interest to the personand the goodness of the object in which theperson is interested, that the example fails to



UNCORREC

TED P

ROO

F

be anything but a contentious example of awork of art.

If the analysis of the examples is correct, talkof an object as having characteristics of interestover and above the characteristics that make ita good one of its kind covers too much. Itincludes merely idiosyncratic matters of taste,as well as cases in which the characteristics thatare of interest are informed by standards ofgoodness, but are accidentally related to theobject being a good one of its kind. What weneed to make Dickie’s point clearer is some-thing along the lines of a description of anobject as good over and above meeting thestandards of goodness of its kind (without thisamounting to its meeting the standards of someother kind that is only accidentally related to it).

The concept of a particular object as goodover and above the standard of goodness set byits kind is enormously problematic, of course.But Dickie is not working on the question ofhow to make sense of that concept. His questionis about the classificatory sense of “art,” whichis defined in part as swallowing kinds whole,that is, as including whole kinds, irrespective ofthe goodness or badness of the individualobjects of those kinds. So, the classificatorysense of “art” applies to a certain range of kindsthat stand in some special relation to the particu-lar extra-good objects. One’s first thought maybe that Dickie’s question is, “What kinds havetraditionally, or up to now, been widely recog-nized as kinds capable of producing instancesthat are good over and above the standards setby their kind?” This question has the virtue offitting well with the first, list-given picture ofthe classificatory sense of “art.” However, if ourquestion were really about what has happenedup until now, our inquiry would be an empiricalinvestigation. It would certainly be odd to try toanswer the question through conceptual analy-sis and thought experiments, as Dickie tries todo. We would just look and see which institu-tions have been recognized as productive ofobjects of the relevant sort, and be done with it.Suppose, on the other hand, that the question isnot about what has or has not happened so far.Suppose the question is, “What kind of kindmust be in place if it is so much as intelligiblefor instances of that kind to be good over andabove whatever the standards are for things ofthat kind?” If this is the question, then at least

the classificatory sense of “art” is a properquarry for a philosopher. But it would alsofollow that we are working with a differentpicture of the classificatory sense of “art” thanthe one determined srictly by Dickie’s list; thisclassificatory sense of “art” is not set out bygiving two lists of artificial kinds, one of whichstipulates the art forms, whereas the other stipu-lates the nonart forms. Rather, this latterconception of the classificatory sense of “art” ismotivated by the intelligibility of a certain sortof judgment, the judgment that some particularobject is good over and above the norms orstandards determined by the kind of thing it is.

VII

We have uncovered two puzzles standing behindthe question, “What is art, in the classificatorysense?” We arrived at one puzzle by taking it tobe constitutive of the classificatory sense of“art” that one list of kinds belongs to itsextension and another list of kinds does not; thequestion is what, if anything, do the art-kindshave in common that distinguishes them fromnonart-kinds? We arrived at the other puzzle byfocusing on a very special sort of judgmentconcerning the goodness of some particularobjects; the question here is about the categoryof kind that is presupposed by judgments of thissort. The first puzzle is explicit in Dickie’sintroductory remarks and referred to throughouthis work. The second puzzle appears almost asan aside in some remarks that, as a whole,address a quite different point.36 If we were toconsider only these facts, it would seem that,whatever we are to say about the relationbetween the two puzzles, it must be the first onethat interests Dickie. However, it would be adisservice to institutionalism and to Dickie’swork to consider only his official statements ofthe question. For on the hypothesis that it isreally the second puzzle that is of centralconcern, Dickie’s own claims about the import-ance of the classificatory sense of “art” snapinto focus.

Suppose it is the second picture of the classi-ficatory sense of “art” that is at issue. If so, first,Dickie’s repeated claim that he is doingmetaphysics rather than ordinary languagephilosophy is clearly true.37 On the proposed



UNCORREC

TED P

ROO

F

understanding, the question, “What is art, in theclassificatory sense?” need not be a questionabout how the word “art” is used in English,now or ever. Possibly, some people sometimesuse “art” to mean “instance of such a kind that itis intelligible that some instance of that kindcould be good over and above being a good oneof that kind,” but it is entirely consistent withthere being this question that no one ever usesthe phrase in this way.38 On the other hand, ifour question was, “What do all forms of paint-ing, music, theatre, vase, and sculpture sharethat is not also shared by furniture, sports,hymns, forms of clothing, and cars?” it is moredifficult to see how the question necessarilybelongs to metaphysics. The question thatinvokes the two lists is, after all, just an instanceof a question that can by described schemat-ically as, “What is true of A’s, B’s, and C’s thatis not also true of D’s, E’s, and F’s?” The factthat we can raise such a question about somepair of lists is no reason to suppose we are doingmetaphysics; we might be simply listing outthose kinds that tend to be called “art” in ourtime and in our culture.

Second, Dickie’s often repeated insistencethat the classificatory sense of “art” marks out“the basic domain of the philosophy of art,”by capturing or expressing something that ispresupposed by the excellence-guaranteeingsense of “art,” is also true, and necessarily so.39

Moreover, and along similar lines, it becomesclear that Dickie is in pursuit of an issue that isfundamental to any analysis of the excellence-guaranteeing sense. For in the absence ofa good account of the kinds that can stand inthis special relation to some of their instances,we are apt to mistake merely prominent orstriking elements of some of these kinds fornecessary or sufficient conditions of being ofthe appropriate category of kind. This is a mis-take Dickie explicitly warns us against, andsurely he is correct to do so.40

Third, what is so often said about Dickie’stheory, that it completely ignores the evaluativeaspects of art theory, comes out as simply false.This classificatory sense of “art” depends for itsvery existence on the connections between artand value, which is not to say that the connec-tions are simple or easy to describe; the classifi-catory sense ranges over kinds of objects thatare both valuable as useful and valued, at least

potentially, not simply as useful. Someinstances of the relevant range of kinds arevaluable over and above the value of the kinds.Thus, when Ted Cohen dismisses the classifica-tory sense of “art” as the sort of topic of interestonly to “cataloguers, book-keepers and certainother philosophers,” and invites us instead tothink about the sense of “work of art” thatimplies excellence—not excellence qua bottleopener or excellence qua portrait, but excel-lence that does justice to the thought that thisparticular object demands to be taken seriouslyregardless of one’s antecedent interest in portraitsor bottle openers, it makes sense that Dickiewaves away the objection.41 The questionCohen wants to dismiss is the question aboutthe two lists, and perhaps Cohen is correct todismiss it. On our hypothesis, however, the twolists are not the crucial issue. If the correctanalysis of what it is for an object to merit aperson’s attention regardless of that person’santecedent interests is that the object is goodover and above meeting the standards of itskind, then there is an important connectionbetween Cohen’s and Dickie’s topics. Far fromforgetting about the excellence-guaranteeingsense of “art,” Dickie is relying on an interpret-ation of this sense to define the classificatorysense.

Finally, we are in a position to see what iscorrect in the idea that we say important thingswith the phrases “good art” and “bad art,”although there is no one important thing that wesay. If the classificatory sense of “art” rangesover all and only those kinds, some instances ofwhich can be good over and above meeting thestandards of their kind, then the judgmentsconcerning the goodness and defect of theinstances, qua instances of their kind, would beamong the conditions of intelligibility of theclassificatory sense of “art.” Moreover, sincethe classificatory sense of “art” includes manydifferent kinds within its extension, there wouldbe no one set of standards to provide the contentof the phrases “good art” and “bad art.”However, since the classificatory sense of “art”is now being defined partly in terms of theexcellence guaranteeing sense of “art,” it isprobably wise to stop using the phrases “goodart” and “bad art” entirely. There are hazards onall sides, and the phrases do not do any essentialwork. To see this, accept for the moment that



UNCORREC

TED P

ROO

F

the Magnussen bottle opener really is an openerthat is good over and above meeting the standardsof a good bottle opener. If so, the Magnussenopener is a work of art, in the specialexcellence-guaranteeing sense. So, does it notfollow that ordinary, perfectly adequate but other-wise unremarkable bottle openers are good bottleopeners and bad works of art? Such a view isnot tempting; luckily, nothing in our premisesimplies it. In maintaining that the Magnussenbottle opener is a work of art, in the specialexcellence-guaranteeing sense, all we arecommitted to is that “bottle opener” names oneof the kinds belonging to the classificatorysense of “art.” That is, we are committed toholding that a bottl opener is a kind of thing,such that some particular one can be understoodas good over and above the norms for its kind.It follows that a good or bad bottle opener isa good or bad work of art, in the classificatorysense—but it is “bottle opener,” not “work ofart,” that is providing the criteria of goodnessand defect here. It does not follow that ordinaryrun-of-the-mill bottle openers, be they good orbad ones, are works of art at all, in the specialexcellence-guaranteeing sense. The excellence-guaranteeing sense just does not come into it.42

VIII

I have argued that the question at the heart ofDickie’s brand of institutionalism is: Whatcategory of kind is presupposed in the judgmentthat some particular thing is good over andabove its kind? The answer I have been pushingis that the relevant category of kinds is thecategory of artificial kinds, the category, that is,of individual things that are what they are invirtue of being embedded in a cultural practice.My hope is that this line will no longer seemthoroughly implausible. One might still suggest,however, that there is in the end only onequestion at issue, because the list-determinedquestion is just another way of raising the ques-tion about the category of kinds. The suggestionamounts to the claim that particular paintings,sculptures, pieces of music, poems, and vases,but no particular sporting events, cars, pieces offurniture, jokes, or works of philosophy, can begood over and above meeting the standards oftheir kind, and so, strong or partial institutional-

ism are still the only possible theories of art, inthe classificatory sense.

My final argument against this view is simplyan argument by elimination. Consider stronginstitutionalism, the thesis that an institutionalkind is a kind of art only if it is embedded in theinstitution of art. It is not just that we havefailed to uncover any clear arguments for thisview; the theory now looks muddled. For if the“institution of art” is an institution, then it setsthe standards of goodness and defect for theobjects informed by it. But, first, we have notmade use of any distinct set of standards com-mon to all art forms. Indeed, one of the draws ofinstitutionalism is that it accommodates notonly the fact that different standards are in placefor different art forms, but that different worksof art, in the excellence-guaranteeing sense,may be excellent in many different ways. And,second, even if there are some set of standardsthat are common to all kinds of art, and whichare relevant to them in virtue of their placewithin the institution of art, the question wouldstill be whether individual objects embedded inthat institution can be good over and abovemeeting the standards of the institution of art.Whether the answer to this question is yes or noclearly cannot involve appealing to yet anotherinstitution. Strong institutionalism, therefore, isnot a possible answer to the question about thatcategory of kinds presupposed by the excellence-guaranteeing sense of “art.”

This leaves us with partial institutionalism,which holds that not all institutional kinds arekinds of art while rejecting any appeal to aspecific institution of art. It is difficult to seewhat the partial institutionalist’s grounds are forinsisting that this category must be a propersubset of artificial kinds. The partial institution-alist might argue positively that it is in factinconceivable that instances of some artificialkinds can surpass the standards of their kind;I would like to see this argument. For the fact is,we know about the thought and effort put intopackaging, flower arranging, and food prepar-ation in Japan.43 We know how seriously theFrench have taken food, wine, perfume, andclothing. We know about Noguchi’s coffee tables,lamps, and playgrounds; about the Eames’sheat-molded plywood chairs; about IsseyMiyaki’s dresses; as well as Enever andPalmer’s success in designing the engine and



UNCORREC

TED P

ROO

F

body of the MG-A. These are all pretty clearlyplaygrounds, pieces of furniture, dresses, andsports cars, the excellences of which transcendthe standards of goodness determined by thekinds: playground, chair, lamp, and so on. Andwho knows what is lost to us through history orwhat will happen next? It seems much moreplausible to hold that there is no principled wayto rule out any artificial kind as possibly beingthe focus of the thought and work involved increating something good over and above meet-ing the standards of its kind.

One might always suggest that partial institu-tionalism and strong institutionalism are theor-ies designed to answer questions other than theone about the category of kinds presupposed bythe excellence-guaranteeing sense of “art.” Farfrom having any objection to such a suggestion,I suspect that partial institutionalism at leastprovides the basic framework for both a theoryof the excellence-guaranteeing sense of “art”itself and the ordinary sense of “art form.”Indeed, some of the considerations offered inprevious sections of this essay incline me tothink that a properly developed partial institu-tionalism is the only plausible account of thesense of “art” I have called the excellence-guaranteeing sense. As an account of theordinary sense of “art form,” partial institution-alism is also plausible. As the phrase is ordinar-ily used, an artificial kind counts as an art formonly after someone has made a work of art, inthe excellence-guaranteeing sense, of that kind.Perhaps even this is not enough, and art forms,in the ordinary sense of the phrase, are strictlyspeaking only those artificial kinds that arecommonly or generally thought of as kinds,some instances of which are works of art, in theexcellence-guaranteeing sense.44 Nothing I haveargued requires that I deny the theory; I wouldonly point out that, so understood, the ordinaryconcept of an art form presupposes both thecategory of kind treated under the heading ofthe classificatory sense of “art,” and the conceptexpressed by the excellence-guaranteeing senseof “art.”

I would press that we draw two conclusions:First, the scandal of Dickie’s doctrine regardingthe theory of art is the result of being tornbetween a bad picture and some good argu-ments. He sees that the traditional list of artforms have nothing in common other than being

institutional kinds. He sees that part of thephilosophical importance of the concept of anart form has to do with the framework necessaryto make sense of the judgment that someparticular object is good over and above fulfill-ing the standards set by its kind. He evenacknowledges that there are some coffee tablesand lamps and chairs and sports cars that aregood over and above their kinds. Were it not forthe deeply but uncritically held assumption thatcars and chairs are simply not kinds of art, therewould be no scandalous flaunting of a circularaccount. The second conclusion I hope to havesupported is that we should reject the assump-tion that some kinds of artifacts are not art,in the classificatory sense. If I have failed incompellingly demonstrating that this assump-tion is false, I hope at least to have shown thatits truth is not obvious.45

LAUREN TILLINGHASTDepartment of Philosophy Knox College Galesburg, IL 61401

INTERNET: [email protected]

1. Noël Carroll, “Identifying Art,” in Institutions of Art:Reconsiderations of George Dickie’s Philosophy, ed.Robert J. Yanal (Pennsylvania State University Press,1994), pp. 12–13. The earliest essay I have found that raisesthe insufficiency objection is Kendall Walton’s review ofDickie’s Art and the Aesthetic: An Institutional Analysis(Cornell University Press, 1974). See Kendall Walton, ThePhilosophical Review 86 (1977): 97–101.

2. The Art Circle: A Theory of Art (New York: Haven,1984). Reprinted (Evanston, IL: Chicago Spectrum Press,1997), p. 12.

3. The germ of institutionalism is introduced in “DefiningArt,” American Philosophical Quarterly 5 (1969): 252–258.For the remark about never having promised to play accord-ing to the original rules, see Dickie, “The InstitutionalTheory of Art” in Theories of Art Today, ed. Noël Carroll(University of Wisconsin Press, 2000), pp. 103–104.

4. Dickie, “The Institutional Theory of Art,” p. 96, n. 14.5. Some familiarity with Wittgenstein’s Tractatus and

Marcel Duchamp’s In Advance of a Broken Arm might atleast give one pause before accepting that philosophy andwisecracking in particular are obvious examples of institu-tional kinds that are not kinds of art.

6. Dickie, The Art Circle, p. 52.7. For example, ibid., pp. 8 and 10.8. Ibid, pp. 80–82.9. I have altered the order. Dickie’s list runs: artist,

work of art, public, artworld system, and the artworld.I am indebted to Barbara Scholz’s essay, “Rescuing the

art003.fm 6 Wednesday, February 26, 2003 5:03 PM


UNCORREC

TED P

ROO

F

Institutional Theory of Art: Implicit Definitions and FolkAesthetics,” The Journal of Aesthetics and Art Criticism 52(1994): 309–325, for the idea of rearranging the definitions.For Dickie’s explanation of why he organizes the list as hedoes, see “The Institutional Theory of Art,” pp. 100–101.

10. This point is made explicitly on pp. 65–66 of The ArtCircle.

11. See J. L. Austin, Sense and Sensibilia (Oxford:Oxford University Press, 1962), sec. 7.

12. Dickie is explicit on this point. He says, for example,“If art is the sort of thing I think it is, the only correctaccount of it would have to be a circular account.” The ArtCircle, p. 78.

13. Dickie defends circularity, uninformativeness, andarbitrariness in The Art Circle, pp. 75–79. The consider-ations he offers seem to support the plausible view thatanswers to philosophical questions will be uninformativein a certain sort of way. It is interesting that many of Dickie’sarguments in support of a circular account are compellingarguments against more descriptively rich accounts.

14. I have not been able to find a contemporary philoso-pher who holds this view. Barbara Scholz’s essay, “Rescu-ing the Institutional Theory of Art: Implicit Definitions andFolk Aesthetics,” The Journal of Aesthetics and ArtCriticism 52 (1994): 309–325, looks like an example untilone notices that on Scholz’s understanding of implicitdefinition, implicit definitions are not sufficient to mark outwhat is being defined. In different circumstances, I wouldargue that R. J. Collingwood is a weak institutionalist in ThePrinciples of Art (1938).

15. If the distinguishing feature of the relevant practice issomething on the order of making pleasure (or a specialkind of pleasure), then we would have a functionalistversion of strong institutionalism. A nonfunctionalistexample is Stephen Davies’s attempt to define a special sortof entitlement that belongs to the practice of art. See Defin-itions of Art (Cornell University Press, 1991), chap. 4.

16. Noël Carroll’s remark that I quoted in the openingparagraph, that Dickie would be clearer if he gave uptalking about defining “art,” suggests that Carroll inclines topartial institutionalism. Robert Stecker also seems to endorsepartial institutionalism in his essay, “Is It Reasonable toAttempt to Define Art?” in Theories of Art Today, ed. NoëlCarroll (University of Wisconsin Press, 2000), pp. 45–64.

17. The terminology belongs to Jeffery Weiand. See“Can There Be an Institutional Theory of Art?” The Journalof Aesthetics and Art Criticism 39 (1981): 409–417. Action-institutions contrast with person-institutions. A person-institution is an abstract agent of some sort, defined in partby its having certain specific powers and determiningcertain roles for its members. Churches, museums, founda-tions, and states are all examples of person-institutions.Waging war, declaring sainthood, and electing leaders areexamples of the powers of person-institutions. Dickie grantsthat he confused the two sorts of institutions in earlier for-mulations of institutionalism. See The Art Circle, pp. 7–14and 49–53.

18. Ibid., p. 52. A developed account of action-institutions would require treatment of a number ofquestions concerning the levels and types of understandingpresupposed by action-institutions. The difficulties involvedin developing such an account, however, do not affect thepoints at issue in this essay.

19. Promising is a terrific example of an action-institution, but is a pretty difficult concept in its own right.For a discussion of some of the puzzling aspects of a signthat means that the sign-giver’s future possibilities of actingwell are limited by giving the sign, see G. E. M. Anscombe,“On Promising and Its Justice and Whether It NeedBe Respected in Foro Interno,” The Collected Papers OfG. E. M. Anscombe, vol. 3 (University of Minnesota Press,1980), pp. 10–21.

20. The Art Circle, p. 51.21. A defense of the claim that institutionalism affords at

least a necessary condition of art would require pursuingthis conceptual point.

22. Dickie is thinking of two of Danto’s essays in particu-lar. “Artworks and Real Things,” Theoria 39 (1973): 1–17.Reprinted in Aesthetics: A Critical Anthology, ed. George.Dickie and Richard J. Sclafani (New York: St. Martins,1989), pp. 551–562. Arthur Danto, “The Transfiguration ofthe Commonplace,” The Journal of Aesthetics and ArtCriticism 33 (1974): 139–148.

23. Dickie often couches the argument in terms of onlyone of the objects being a work of art, rather than only onebeing a painting or a piece of sculpture. For the reasonsI have already laid out, I regard it as begging the questionto stipulate that packages of scouring pads are not art, inthe classificatory sense. See The Art Circle, pp. 12–13 and64–65.

24. Ibid., p. 13. But see also “Art and Value,” The BritishJournal of Aesthetics 40 (2000): 228–241, where Dickieconcedes the point that this consideration is less thanentirely compelling.

25. There is an interesting passage near the end of “Artand Value,” in which Dickie acknowledges that in regularcritical discourse about paintings, pieces of theater, andso on, nobody calls these items “art” except as extremelyhigh praise. What is interesting is that he seems tosee this as evidence for a classificatory sense of “art.”See p. 236.

26. Peter Geach, “Good and Evil,” Analysis 17 (1956):103–112. Reprinted in Theories of Ethics, ed. Philippa Foot(Oxford: Oxford University Press, 1967), pp. 64–73.

27. The first list is from Dickie’s Introduction to Aesthet-ics: An Analytic Approach (Oxford: Oxford UniversityPress, 1997), p. 57; the second and third are from The ArtCircle, pp. 31 and 35 respectively.

28. For example, by M. W. Rowe, “Why ‘Art’ Doesn’tHave Two Senses,” The British Journal of Aesthetics 31(1991): 214–221. See also, Graham Oppy, “On FunctionalDefinitions of Art: A Response to Rowe,” The British Jour-nal of Aesthetics 33 (1993): 67–71, for criticism of Rowe’stripartite division of nouns and subsequent argument byelimination.

29. The Art Circle, p. 13. But see “Art and Value,” p. 229,where Dickie claims that “even the worst art will have a tinybit of value.”

30. Berys Gaut, “‘Art’ as a Cluster Concept,” in Theoriesof Art Today, ed. Noël Carroll (University of WisconsinPress, 2000), pp. 38–39.

31. We may build in this further restriction without argu-ment because we have limited the discussion to whetherinstitutionalism provides only a necessary condition, ora necessary and sufficient characterization of art, in theclassificatory sense.



UNCORREC

TED P

ROO

F

32. See The Art Circle, pp. 13–14 and 37–43. Dickiewonders whether the derivative “sense” is a genuine senseof “art” or is in fact metaphorical. Either way, the possibil-ity that the classificatory sense straightforwardly andliterally applies to cars and pieces of furniture is neverconsidered. I am ignoring Dickie’s remarks concerning an“evaluative” sense of “art” that is some sort of genericpraise word applying as centrally to the Grand Canyon andsunsets as to The Polish Rider and Hamlet. As others havenoticed, nothing in institutionalism implies that there mustbe any such sense of “art.” See Robert Stecker, “The End ofAn Institutional Theory of Art,” The Journal of Aestheticsand Art Criticism 26 (1986): 124–132.

33. See The Art Circle, p. 75, 80, and 81 respectively.34. Ibid., p. 56.35. It is clear that Dickie means to be making an histor-

ical claim in the quoted passage. One might be skepticalwhether he is also making a conceptual claim about art. Ifthe point were simply historical, however, the most Dickiewould be in a position to claim is that it is false to hold thatpeople were making art before they were making objectsthat have characteristics of interest over and above the inter-est they have as elements in the institution in which they areembedded. But Dickie does not say that such a doctrine isfalse; he says it is meaningless.

36. Dickie is arguing that institutionality does not entailthat practices must be invented in one fell swoop. See ibid.,pp. 53–56

37. For example, ibid., pp. 13 and 37.38. Thus, I would object to Robert Stecker’s assumption

that a definition of “art” is always aimed to explain howpeople actually use the term. For Stecker’s view, see “Is ItReasonable to Attempt to Define Art?” in Theories of ArtToday, ed. Noël Carroll (University of Wisconsin Press,2000), pp. 45–64, and the very similar set of suggestionsin Richard Kamber, “A Modest Proposal for Defining aWork of Art,” The British Journal of Aesthetics 33 (1993):313–320.

39. Dickie, The Art Circle, p. 13.40. Ibid., p. 5.

41. Ted Cohen, “The Very Idea of Art,” NCECA Journal 9(1988): 7–14. The remark about bookkeepers is on p. 8.See also “Reflections on One Idea of Collingwood’sAesthetics,” Monist 72 (1982): 581–583. David Novitzmakes some similar suggestions in “Disputes about Art,”The Journal of Aesthetics and Art Criticism 54 (1996):153–164.

42. This is one of the logical features of the excellence-guaranteeing sense that makes it interesting and difficult toexplain.

43. On packaging, see Yuriko Saito, “Japanese Aestheticsof Packaging,” The Journal of Aesthetics and Art Criticism57 (1999): 257–266.

44. It might be suggested that the definition of theordinary sense of “art form” proposed in the text puts thegenerality in the wrong place. Art forms, the suggestionmight go, are the practices of making that commonlyor regularly or reliably produce works of art, in theexcellence-guaranteeing sense. Thus, painting and poetryare art forms, but bottle opener making is not, since it israre indeed that a bottle opener is good over and above thestandards of bottle openers. I can see no logical or con-ceptual difficulty with a concept that works along thesuggested lines, although it is implausible as an account ofthe ordinary sense of “art form.” Painting and poetry areparadigmatic cases of art forms, in the ordinary sense; onthe proposed definition of “art form,” however, it is prob-ably false that painting and poetry are art forms. Lots andlots of people make paintings and poems; for example, inhigh school and college classes, at community centers, atsummer camps, and on vacations. And, although I do notknow what the actual statistics are, I suspect most of theproducts of these efforts do not even meet the standards ofthe kinds of painting or poem at issue, much less surpassthose standards.

45. I would like to thank George Dickie, Lance Factor,Miles Rind, Candace Vogler, Stephen Wessley, Bill Young,and an anonymous referee for The Journal of Aesthetics andArt Criticism for their discussions, suggestions, and criti-cisms of previous drafts of this article.


Author Query Form

Journal: The Journal of Aesthetics and Art Criticism

Article : 03

Dear Author,

During the copy-editing of your paper, the following queries arose. Please respond to these by marking up yourproofs with the necessary changes/additions. Please write your answers on the query sheet if there is insufficient spaceon the page proofs. Please write clearly and follow the conventions shown on the attached corrections sheet. Ifreturning the proof by fax do not write too close to the paper’s edge. Please remember that illegible mark-ups maydelay publication.Many thanks for your assistance.

QueryRefs.

Query Remarks

1 Au: Is this spelling correct? At the JAAC website it’s “Wieand.” CE

2 Au: Is this part of the name of the journal? Please clarify what these initials stand

Sriram S

Sriram S

Marginal mark

Stet

New matter followed by

New letter or new word

under character

e.g.

over character e.g.

and/or

and/or

MARKED PROOFÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐ

Please correct and return this setÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐ

Please use the proof correction marks shown below for all alterations and corrections. If you

wish to return your proof by fax you should ensure that all amendments are written clearly in

dark ink and are made well within the page margins.

Textual mark

under matter to remain

through matter to be deleted

through matter to be deleted

through letter or through

word

under matter to be changed





Encircle matter to be changed

(As above)

through character or where

required

(As above)

(As above)

(As above)

(As above)

(As above)

(As above)

linking letters

between letters affected

between words affected

between letters affected

between words affected

Instruction to printer

Leave unchanged

Insert in text the matter

indicated in the margin

Delete

Delete and close up

Substitute character or

substitute part of one or

more word(s)

Change to italics

Change to capitals

Change to small capitals

Change to bold type

Change to bold italic

Change to lower case

Change italic to upright type

Insert `superior' character

Insert `inferior' character

Insert full stop

Insert comma

Insert single quotation marks

Insert double quotation

marks

Insert hyphen

Start new paragraph

No new paragraph

Transpose

Close up

Insert space between letters

Insert space between words

Reduce space between letters

Reduce space between words

the classificatory sense of “art”

Documents