The Sociology of Art

P R O B L E M S O F T H E S O C I O L O G Y O F A R T

T H E W O R K O F P I E R R E F R A N C A S T E L

I

LIKE others who work on the frontiers between disciplines, the latePierre Francastel (who died in Paris in January 1970) runs the risk of neglectby both sides, of being considered a sociologist by art historians and an arthistorian by sociologists. In the English-speaking world, his work is littleknown*; none of his books and virtually none of his articles have been trans-lated into English, as they have been into Italian, Polish, Spanish and otherEuropean languages (1). Yet Francastel well deserves to be known toanyone interested in the sociology of art. His writings cover a wide rangeof subjects in the history of painting, sculpture and architecture, from theXlth century to the present day (2). Although he spent his life in studyingthe arts in particular societies, he was fascinated by the possibilities ofgeneralisation about the relationships (the plural is important) of works ofart to the societies in which they were created. The power of his workhas a great deal to do with the fact that his insights are not separate butcoordinated. His main ideas form a system. This system can easily bemisunderstood; it looks more like Marxism and more like some currentforms of structuralism than it really is. So it may be useful to summariseFrancastel's central ideas as a set of connected propositions before going onto place them in an intellectual and social context and to assess their fruit-fulness.

1. Art is a system of signs. Francastel often attacks the view that art(and Renaissance art in particular) is (in Alberti's famous image) a windowopened on reality, the representation of things as they "really" are. "Apicture is not the double of reality, it is a sign. Paintings can be seen asforming "systems of signs (des systemes de signes) which can be [...] putin parallel with the various languages—verbal, mathematical, musical—which societies use to communicate" (3). Art is like language, and lookingat pictures is like reading them; to see is to decipher. The point of this

* FRANCASTEL'S best-known works include Peinture et socie'te (London, Audin,1951); Art et technique au XIXe et XXe siecles (Paris, Ed. Minuit, 1956); Lare'alite figurative (Paris, Gonthier, 1965); La figure et le lieu (Paris, Gallimard,1967). These works are henceforth abbreviated to... P&S; A&T; RF; F&L.

(1) FRANCASTEL'S article, 'Main trends (2) Little has been written about Fran-of European art', is translated in G. S. M E - castel, but an interesting article is S.TRAUX and F. CROUZET, The nineteenth- PEROTTINO, La notion de structure danscentury world (New York 1963), and a sec- l'ceuvre de Francastel, La Pensie (1967),tion from his essay on "the destruction of a n° 135, pp. 153-164..plastic space" is translated in W. SYPHER (3) P&S, p. 45; F&L, p. 28. Here and(ed.), Art History (New York 1963). elsewhere the translations are my own.

Arch, europ. social., (1971). 141-154.

PETER BURKE

art-as-language metaphor is to make it clear that the'innocent eye' is impos-sible, and that artists always operate according to a set of conventions. Theobjects which may be seen in paintings are not a random collection; ineach society there is a 'repertory' of objects on which the artist draws.In the art of XVth-century Italy, for example, this repertory includes therock (as in paintings of St Jerome and other saints in the wilderness); theroof (under which the Nativity is often set); the carroccio or chariot, usedin scenes of triumph; and the nuvola or cloud, from which God or theVirgin or saints sometimes appear. These objects are in fact propertiesfrom the religious theatre of the time, and so could easily be recognised bycontemporary spectators (4).

This repertory of objects may be considered the vocabulary of the"language of art". The method of combining objects is also subject torules or conventions—the grammar and syntax of the language of art (5).That is, Francastel is recommending that the art-historian pay attentionto le lieu as well as la figure ; to the space between the objects in a painting,to the structure or organisation of that space. The subtitle of his Peintureet sodete is "birth and destruction of a plastic space from the Renaissance tocubism". Its point of departure is the idea that linear perspective "doesnot correspond to any absolute progress on the part of humanity in thedirection of an ever more adequate representation of the outward world ona two-dimensional surface; it is one of the aspects of a conventional mode ofexpression based on a certain state of technique, knowledge and society ata given moment " (6). For example, the paintings of Piero della Francescado not represent the world as it is, but rather a set of pictorial conventionswhich include the rule that light must come from a single source, and thatpictures should ideally be seen (as in the case of Brunelleschi's famousbox) with one eye closed. Renaissance linear perspective, like Euclideangeometry (to which it is closely related) and Newtonian physics, is only onesystem among possible others. It is as absurd to accept it as the best solu-tion of man's visual needs as to believe that a particular language is thebest solution of man's semantic needs.

This idea of art as a language is, of course, not unique to Francastel;the idea that paintings contain a conventional repertoire of objects goesback to Aby Warburg at least. He showed, for example, that the flutteringdraperies in Botticelli's Birth of Venus derived from Roman reliefs. Again,Erwin Panofsky (in a paper of the 1920s which has not been translated intoEnglish) wrote of perspective as symbolic form. Levi-Strauss has writtenon primitive art as a "system of signs", and Gombrich's Art and Illusionis very much concerned with the "linguistics of the visual image" (7).

(4) See especially the essay "Imagination 'disanalogies' is M. MOTHERSILL, IS art aplastique, vision theatrale et signification language ? The Journal of Philosophy, LXIIhumaine", repr. in RF, pp. 211 sq. (1965)1 559-572.

(5) On the grammar of art, F&L, p. 144. (6) P&S, p. 7.The analogy with Barthes and Levi-Strauss (7) A. WARBURG, Sandro Botticellis Ge-on the languages of clothes, food, and hurt der Venus und Friihling (1893), Italiankinship is obvious. A discussion of the trans, in his La Rinascita del Paganesimo

142

THE SOCIOLOGY OF ART

Where Francastel differs from the Warburg school is not so much in hispoint of departure, but in the direction in which he chooses to travel fromthat point. For Francastel, that direction is a sociological one; it is therelation of works of art to the "state of technique, knowledge and society ata given moment". His choice of direction aligns him with such writerson the social history of art as Frederick Antal and Arnold Hauser; but hediffers from them in refusing to take 'realism' as his point of departure.

2. Art and technics are not opposite. Another common view whichFrancastel attacks again and again is the idea that art is (or ought to be) some-thing 'fine' or gratuitous or superfluous, something which does not fill a socialfunction. He argues that the opposition between art and 'technique'(it is difficult to know whether to translate this term as 'technique', 'technics'or 'technology'), the opposition between art and the machine, art andindustry is in fact a 'pseudo-opposition', because all art contains a technicalelement, and all technology contains an aesthetic element. He attacks,among others, Ruskin, Mumford and Giedion for seeing the machine asnecessarily a threat to man and the enemy of art; for denying the possibilityof adapting art to the forms of modern life in industrial society; for thinkingof man as an absolute and refusing to admit "even the possibility of asubstantial transformation of man's functions". He points out that man'stransformation of nature, his creation of his own environment did notbegin with the Industrial Revolution. In short he is anti-romantic, hostileto the idealisation of art for art's sake and of the "organic community".He argues that it was no accident that the nineteenth century was a turningpoint in the history of both art and technology, and, more specifically, thatthe use of new materials such as steel and concrete transformed modernarchitecture. But Francastel is not a technological determinist; he suggeststhat the technical element is a necessary but not a sufficient condition ofthe form of works of art, and that paintings, buildings and tools also expressvalues, ideas and world-views (8).

3. Works of art express world-views. Art is not only related to thetechnical but also to ideas. Linear perspective, for example, is related toa particular conception of space, and space is one of the fundamental cate-gories of human thought. "Space is not a reality in itself which is simplyrepresented in a different way in different periods. Space is man's expe-rience itself" (9). Like Kant, Francastel suggests that experiences neverreach us neat, but only mediated through categories of thought. Like

antico (Florence 1966), pp. 65 sq. E. PA- turies throughout A&T. Classic expressionsNOFSKY, Die Perspektive als symbolische of the attitudes he attacks are L. MUMFORD,Form, in Vortrdge der Bibl. Warburg, 1924- Technics and Civilisation (London 1934), and1925, pp. 258-330 G. CHARBONNIER, Entre- S. GIEDION, Mechanization takes commandtiens avec Claude Li'vi-Strauss (Paris 1961), (New York 1948). For positions similar toesp. the 5th and 8th interviews. Francastel's, see G. FRIEDMANN, Problemes

(8) Francastel makes his general position humains du machinisme industriel (Paris 1946),clear in the essay "Technique et esthe- and A. KOYRE, Les philosophes et la raa-tique", repr. in RF, pp. 58 sq. Detailed chine, Critique, IV (1948),334-333, 610-629.discussion of the XlXth and XXTh cen- (9) P&S, p. 29.

H3

PETER BURKE

Hegel, he suggests that these categories change over time. In medievalart, "there is no space independent of things", no fixed point of view orreference system. There are as many spaces as there are objects. Thisconception of space is sometimes called 'topological space' (10). Thenew space-conception of the Renaissance was euclidean space, space whichcould be measured or calculated, cubic space. This new space was theresult of the adaptation of mathematical researches to solve artistic problems,but it was also part of something wider, "a new total view of the world",which was now seen as subject to the laws of nature which could be discov-ered by human reason. There was a shift of interest from quality toquantity, from myth to reason, from the theological to the naturalistic (n ) .

Again, a new conception of space appeared in late XlXth-centuryFrench art, in which "what is most mysterious is what is most near".This new space was also linked to changes in world-view, to a shift of inter-est and belief from the objective to the subjective, from absolute valuesto relative ones. The "great adventure of the modern world" is phenomen-ology, the consciousness of the role of the mind in structuring the world,a consciousness apparent in the paintings of, for example, Degas, in whichfigures are no longer placed symmetrically within the picture frame (andapparent also, one might add, in Francastel's ideas themselves) (12).

Art expresses not only a view of the physical world but a set of values,which are communicated by means of symbols and myths. The conceptof 'myth' is a central one in Francastel's work. He treats myth as a formof thought in which projection and identification are important; spaceand time, for example, being the projection of the needs of society. Heonce wrote an essay on Botticelli's Primavera entitled "A poetic and socialmyth of the Quattrocento", suggesting that the picture symbolised pros-perity and stability, liberality and good government; in short, that it idea-lised Medicean Florence. One of the social functions of art is propa-ganda. Again, he suggested that the tree in Piero della Francesca's Baptismof Christ symbolised the tree of life. Myths and symbols are the material-isations of values; the values of a particular society (13).

4. Art-criticism is a form of ideology. So Francastel sees art as 'ideology'in the sense that it expresses a world-view which is not so much personalto the artist as shared by a society, or at least a group within it. Art criti-cism is an ideology in the same sense, and also in the sense of being apossible tool for the manipulation of an audience. In 1939-40 Francastelgave a course of lectures at Clermont-Ferrand on art history as a means ofGerman propaganda. His aim was to bring to light latent pan-Germanor racialist tendencies in Frankl, Pinder, Worringer and other German art-historians. Thus Pinder wrote of Germany as the mother of European

(10) Compare J. PIAGET, La representation (13) P&S, p. 56, and Un mythe poetiquede Vespace chez Venfant (Paris 1948), trans, as et social du Quattrocento, repr. in RF,The child's conception of space (London 1956). p. 272 sq. Compare M. MAUSS, Essai sur

(11) P&S, pp. 65 sq. le don (Paris 1925), trans, as The gift (Lon-(12) P&S, pp. 129 sq. don 1954).

I44

THE SOCIOLOGY OF ART

civilisation; Frankl described the regional styles of medieval architectureas the result of the racial characteristics of the Celts in Auvergne, forexample, or the 'Germans' in Normandy. Francastel's conclusion wasthat the German art-historians were trying "to annex [...] whole provinces ofthe past; first Gothic, then Romanesque, then Baroque". Of course, theselectures were themselves a piece of counter-propaganda. Their long-termimportance was in making their author aware of the attitudes latent in worksof art-criticism and art-history. He suggested, for example, that art-history began as propaganda, with Vasari's glorification of Florence in hisLives, and that Ruskin's view of art and beauty as eternal values was relatedto his general conservatism. He characterised Le Corbusier as an authori-tarian, and Mumford as an anti-rational mystic. A sympathetic analystof the myths expressed in works of art, Francastel was a merciless criticof the myths expressed in the work of art-historians—the "myth of themachine", "the myth of the organic", or Croce's "myth of the individual",the idea that the history of art is essentially the history of individual artists.For Francastel, art is part of society (14).

5. Art is part of society. The "system of signs" and the world-vieware not created by individuals, but by groups. In each culture men aretaught to see paintings or the world in a certain way. Behind changesin figurative systems or in world-views lie social changes. So Francastelobjected to the fragmentation of the Romanesque into regional schools bythe art-historians of his day, because this obscured the fact that the develop-ment of religious architecture c. 1100 was related to two important socialchanges—the rising population and the popularisation of the cult of thesaints and their relics—and, more generally, that the artistic changes of thetime can be related to what Marc Bloch described as the transition from the"first feudal age" to the second. Again, Francastel criticised the tendencyto split the main trends of modern art into such movements and counter-movements as 'impressionism", 'post-impressionism', 'fauvism', etc.,again on the grounds that this fragmentation hid the fact that social changein the later XlXth century "destroyed the possibility of a representationof the universe which conformed to tradition and which was alive" (15).

One great danger for anyone who wants to relate art to society is that'society' is such a vague term. In a wide sense it includes ideas and techn-iques ; in a narrower sense it refers to social groups and social movements.So far I have been describing Francastel's use of the wider or 'macrosocial'approach, relating paintings to world-views and world-views to wholesocieties. But one of his great strengths as an art-historian is that he isinterested in the 'microsocial' approach as well; the study of artists, patrons,

(14) These lectures were printed, after (Venice/Roma 1961).the war, as L'Histoire de Vart instrument de la (15) On the Xllth century, P. FRANCAS-propagande germctnique (Paris 1945). On TEL, L'humanisme roman (Rodez 1942),Ruskin see FRANCASTEL'S essay, La Venise esp. pp. 15, 86, 121. On the XlXth cen-de Ruskin et les archfelogues, in C. PELLE- tury, P&S, pp. 119 sq.GRINI (ed.), Venezia nelle letterature moderne

H5

PETER BURKE

and wider groups of spectators (16). He has relatively little to say aboutthe artist and his place in society; the organization of the arts, though heargues, for example, that the Baroque is an art of travelling artisans, incontrast to Classicism, which is the art of urban milieux (17). He has muchmore to say about patrons and the wider public. The early monograph onGirardon (1928) prints some of his contracts and discusses the restrictionswhich the patronage system placed on XVIIth century artists. Again,writing on Xlth century French architecture, he criticises fimile Malefor insisting too much on the role of the Church in the development ofmedieval art, and argues that in this period "religious art ceases to addressitself to the clergy alone", for there were important lay patrons like thecount of Toulouse and the duke of Aquitaine (18). However, a widerpublic than the patrons influenced many works of art. One of Francastel'smost striking examples is that of Italian art c. 1600, with its new emphasison the ecstasies of the saints; his theory is that the Catholic reformers, suchas the Jesuits, had to support the popular cults of saints and images in orderto keep the mass of the population firm in the faith, so that it was the peoplewho converted their missionaries, rather than the other way round. Again,he is sceptical of the Warburg school's exegesis of Botticelli's Primavera,but picks up the suggestion of the French folklorist Andre Varagnac thatthe central figure is not Venus but the Queen of May, and that "Mercury'and the 'Three Graces' are also characters from traditional May festivities(as they could still be observed in Lorraine in this century) (19).

In other words, art is part of history; the moral of that conclusion isthat art-historians should study general history more than they do, andthat historians should make a greater use of art as a source of knowledgeabout society as valuable as any written document. For art is not inferiorbecause artists do not work with words (20).

6. Art is autonomous. Art is not subordinate to literature or to ideas;it is autonomous. Francastel insists that art does not 'reflect' anything.It does not reflect reality—for art is a system of signs. It does not reflectsociety either; Francastel asserts that art is not a 'superstructure' and thatit is just as useful to explain society in terms of art as the other way round.Nor does art reflect ideas, even a world-view, in the sense of a world-viewbeing developed by thinkers or writers and then translated into visualterms by artists. On the contrary, art is a specific and independent wayof exploring reality and expressing awareness of it. Artist and writer worktogether at this task. Thus Francastel argues that the optics of Chevreuldid not influence the Impressionists—nor he them: that does not meanthat there was no link between them, but the link is of another type. Art

(16) On problems of method see the essay et les arts, repr. in RF, esp. pp. 384 sq.,Art et sociologie, repr. in RF, pp. 29 sq. and Un mythe poetique et social du Quattro-

(17) P. FRANCASTEL, Baroque et clas- cento, repr. in RF, pp. 272 sq.; and A. VARA-sique, Annales, XII (1957), 222-207. GNAC, Civilisation traditionnelle et genres de

(18) P. FRANCASTEL, Girardon (Paris vie (Paris 1948).1928); L'humanisme roman, op. cit. pp. 143 sq. (20) This case is argued in the essay,

(19) See the essays La Contre-Relorme Art et histoire, repr. in RF, pp. 73 sq.

146

THE SOCIOLOGY OF ART

may be a 'language' in the general sense of being a system of signs, but it isnot a translation of verbal language; it is an independent form of thought.Art is connected with ideas and with society, but it cannot be reduced toeither. "Only art can express what art expresses" (21).

7. Art is a prefiguration of the possible. If art can be said to 'express'or 'reflect' anything, it is 'thought' or imagination in a wide sense. Penseeplastique is a favourite phrase of Francastel's. So the sociology of art ispart of the sociology of the imaginary, the study of the different forms thatthe imagination takes in different societies. Works of art, like Utopias,may come nearer to reflecting the future than the present or the past. Totake an example which brings out Francastel's gift for seeing the generalin the particular, when we look at the architectural backgrounds in thepaintings of Gozzoli or Ghirlandaio, we probably think of the painters ascopying the buildings they saw around them. In fact, it was the architectswho followed the painters, for very few secular buildings in Renaissancestyle existed before 1500, whereas architectural backgrounds of this kindare relatively common in paintings. Does that mean that we should seethe art of the early twentieth century as the prefiguration of a world whichhas not yet come about (22) ?

To 'place' Francastel intellectually it is necessary not to take his dis-agreements with other scholars too seriously. He always tended to over-estimate the ideological distance between himself and others, particularlywhen the others were not Frenchmen. But the intellectual history of thetwentieth century should no more be fragmented into movements or schoolsthan Romanesque architecture or modern painting; the social history ofart is as much related to its time as the art which it studies. Francastel'sinterest in relating works of art to world-views and both to social changehas much in common with Marxism, which his dismissal of the term'superstructure' and his acid comments on Hauser, Antal and even Lukdcsshould not be allowed to obscure. Again, his approach to the Renaissanceis not so very different from that of the 'Warburg school', and his approachto modern architecture is much like that of Lewis Mumford, whom heoften denounced. Mumford is interested in the social history of space andtime, for example. Roland Barthes came in for some harsh criticism too,but his interest in structure, system and myth is not so different fromFrancastel's either. But still more important in his development is theone group with which he did identify; the 'Annales school' of Frenchhistorians (23).

(21) The Chevreul example is discussed more generous; see RF, p. 82. He attackedin A&T, pp. 137-138. Barthes (RF, pp. 75 sq) though Barthes'

(22) See the essay, Imagination et realite emphasis on structure and on myth isdans l'architecture civile du Quattrocento, much like Francastel's. See R. BARTHES,repr. in RF, pp. 290 sq. Mythologies (Paris 1957), and Elements de

(23) Francastel discovered Lukacs late, simiologie (Paris 1964), trans, as Elementsand, when he had actually read him, was of Semiology (London 1967).

147

PETER BURKE

Annales (subtitled « Economies Societes Civilisations ») was, and is, ahistorical journal founded in 1929 by Lucien Febvre and Marc Bloch,two gifted historians of whom the second is well known in Britain and thefirst scarcely known at all. The journal has always been associated with aparticular style of history, a social history open to influences from anthro-pology, economics, social psychology and sociology; a 'total history' showingthe interrelations of different human activities taking place in any individualsociety. Febvre was a sharp polemicist against the kind of literary history,history of philosophy, and art history which did not relate its subject-matter to the general historical and social background (24). He did notassume that the economy or the social structure determined the style ofart, thought or literature in any period, nor did he assume that the cultureof any age was necessarily a unified whole (he suggested that painting mightbe 'out of step' on occasion), but he worked all his life for what he called a'historical history' of culture which would at least relate it to society.

Francastel was a lifelong admirer of the work of both Bloch and Febvre.His book on Romanesque architecture, L'humanisme roman, was an attemptto bridge the gap between art history and the social history of France in theeleventh and twelfth centuries as Bloch had written it. A favourite phraseof his, outillage mental, 'mental equipment' was a slogan of Febvre's;it sums up the idea that art is related not only to the tools with which it ismade but also to the 'mental tools' of its makers. In a given society thereare certain current forms of thought as there are current forms of tool.Writing on architecture in the XlXth and XXth centuries, Francastelsuggested that faster travel has modified our perceptions of space and time,and made generous acknowledgement to Febvre for the idea that perceptionhas a history (25). He published a number of articles in Annales, onPoussin, for example, and on "baroque and classicism". When Febvrebecame head of the Vlth section of the Iscole Pratique des Hautes Iitudes in1948, he called Francastel to teach the sociology of art there. Francastelmakes constant reference to the works of a number of social scientists suchas Marcel Mauss in anthropology, Georges Friedmann in sociology, andJean Piaget in social psychology, and the Annales school do so too.

If art criticism and art history are forms of ideology, Francastel's ownwork is no exception to the rule. His own world-view is apparent in hiswork, in particular a certain anti-romanticism in personal taste which goeswith his dry, astringent prose and his rationalist view of art as primarily aform of thought rather than primarily the expression or communicationof emotion. He might also be accused of projecting the non-representationalvalues of twentieth-century art on to the art of the past, that of the Renais-sance for example. He could be a sharp critic of the work of his colleagues

(24) For Febvre's conception of art his- McLuhan, in La religion de Rabelais (Paristory see his Combats pour l'histoire% (Paris 1942), PP. 461 sq. His views are them-1965), pp. 295 sq. selves discussed in Z. BARBU, Problems of

(25) A&T, pp. 123 sq. FEBVRE discuss- historical psychology (London i960), pp. 21 sq.ed the history of perception, long before

I48

THE SOCIOLOGY OF ART

on the grounds that they went beyond the evidence, but the same criticismcan, on occasion, be made of him. At various points in his arguments,some of them crucial, evidence is lacking and one finds instead "it hasbecome clear to me that [...]" or some such phrase. Francastel objectedto the iconographical approach to works of art, mainly on the grounds thatthis was to deny the 'specificity' of art. "Only art can express what artexpresses". Yet he did not draw the moral that we should not discussart in words at all; he relied on intuitions rather than on texts, which is avery different matter. His own 'mythological' approach ran the riskof being Warburg's iconography without the rigour. For example,his assertion that the tree in Piero della Francesca's Baptism of Christ isthe tree of life, and that columns and caryatids in other paintings carrythe same meaning cannot be proved wrong, but certainly lacks therigour of his usual approach. It is perhaps no accident that the book,promised in 1951, on the social and political myths of the XVth centurywas never written (26).

Francastel stands apart from most art historians. He preferred notto write monographs, on the grounds that "art history is weighed downby monographs and exhaustive catalogues". He did not concentrate onindividual artists. He did not study artistic forms, as Henri Focillon did,as if they had a life of their own. He did not study the intellectual meaningof works of art, as fimile Male and Aby Warburg did. One might callhim a social historian of art, but here too he stands apart. He did not likethe work of Arnold Hauser and Frederick Antal because they described artas if it were "the materialisation of a kind of collective thought via theactivity of a man reduced to the role of a pen-holder". Unlike them, hedid not discuss art in terms of social classes. On the other hand he didnot, like Professor E. H. Gombrich, or Professor Francis Haskell, thinkof the social history of art as essentially the history of the material conditionsunder which art is commissioned and created. That, for him, was animportant part of the story, but only a part. In other words, he thoughtthat both the 'microsocial' and the 'macrosocial' history of art were worthtaking seriously and could be fused, whereas most social historians of arthave tended to concentrate on one or the other (27).

Another objection to describing Francastel as a social historian of artis that this is to play down one of his most striking qualities : his gift forgeneralisation, for constructing types, for coordinating his insights. Inhis book on Painting and society, for example, he shows that scholars workingin relatively narrow fields (like G. Kernodle on art and theatre in theRenaissance, or E. Lowinsky on 'perspective' in Renaissance music) haveillustrated unaware some extremely general theses about art, ideas andsociety; the thesis, for example, that artists tend to introduce a conventionalrepertory of objects into their paintings, or that a changing sense of space

(26) The tree of life appears in P&S, (27) A&T, p. 255 denounces Antal andp. 77. Hauser.

149

PETER BURKE

is part of a changing world-view. On account of this interest in generalisa-tion one might call Francastel a sociologist of art, a title he would probablyhave accepted, since he wrote an article on problems of the sociology ofart for a collective treatise on sociology, and gave his book of essays, Larealiti figurative, the subtitle, "Structural elements of a sociology of art" (28).

II

Now just what is the sociology of art ? It seems reasonable to defineit as the study of the relationships between art and society at a generallevel, thus distinguishing it from the social history of art, the study of thesame relationships at the level of the particular. So far the sociology ofart looks like other particular sociologies, like industrial sociology, politicalsociology, or the sociology of religion. However, the sociology of art isasymmetrical, unlike most of the particular sociologies in two importantrespects. In the first place, it is not primarily the sociology of a kind ofbehaviour—the creation, commissioning and collection of works of art—butis more concerned with the results or materialisations of that behaviour,the objects themselves. Second, it is a historical sociology. So is thesociology of religion (as defined, for example, by G. Le Bras), but not tothe same extent. As for the other particular sociologies, there is in principleno reason why political sociologists or industrial sociologists should notbe concerned with societies which existed before 1900, but in practice theyusually are not. In the case of art, the variety of past styles and theirdifferences from contemporary styles are so obvious and overwhelmingthat it would be very hard to limit oneself to the twentieth century. In thisfield, it is rather the contemporary which has been neglected. Relativelyfew sociologists have used their normal methods of questionnaire andinterview on artists and their publics. Interesting exceptions to this ruleare Mason Griff, who has studied art-students and commercial artists inChicago, and Bernard Rosenberg and Norris Fliegel, who studied twenty'vanguard artists' in New York between 1961 and 1962. A bibliographyof the sociology of art would include few studies of this century but rathermore of the past, studies more often made by historians (E. H.Gombrich,F. Hartt, F. Haskell, M. Meiss, G. Pelles, M. Schapiro, etc.) than by sociol-ogists (C. White and H. White, for example). These studies do notusually attempt to generalise (29).

(28) P. FRANCASTEL, Problemes de la guard Artist (Chicago 1965); H. C. WHITEsociologie de 1'art, in G. GURVITCH, Traite" and C. A. WHITE, Canvases and careersde sociologie (Paris i960). (New York 1965), F. HARTT, Art and Free-

(29) M. GRIFF, The commercial artist, dom in Quattrocento Florence, in L. F. SAN-in M. R. STEIN, A. J. VIDICH, and D. M. DLER, Essays in memory of Karl LehmannWHITE, Identity and Anxiety (Glencoe (New York 1964); F. HASKELL, Patrons andi960); M. GRIFF, Conflicts of the Artist in Painters (London 1963); E. H. GOM-Mass Society, Diogenes, XLVI (1964), 54-68; BRICH, The Early Medici as Patrons of Art,B. ROSENBERG and N. FLIEGEL, The Van- repr, in his Norm and Form (London 1966);

THE SOCIOLOGY OF ART

What has become of the general theory of the relations between art andsociety ? Marx sketched one out, but his discussion of the arts is a frag-mentary one, leaving important questions unanswered. Granted histheory of 'superstructure' and 'base', it would still make sense to askwhether the relation between the two is the same (a) for all societies, and(b) for all art-forms, and also whether their relation is direct or indirect.An interesting attempt at developing a Marxist sociology of art was recentlymade by Ernst Fischer, but unfortunately it was not worked out in detail,particularly so far as the visual arts are concerned. Fischer criticises anyattempt to make a 'mechanistic oversimplification' of Marx's theory of thebase and the superstructure. His own account of the relation of style tosociety is that it is mediated through content or 'ideology', and that styleexpresses content not directly but obliquely. A striking example, whichFischer does not pursue, is his suggestion that Romanticism was "a move-ment of protest [...] against the bourgeois capitalist world" (30).

A more fully worked-out Marxist sociology of art is to be found in thework of the late Frederick Antal. At a general level he suggests, forexample, that "popular and aristocratic art, both irrational, always show acertain kinship with one another, in contrast to rational upper middleclass art". At a more specific level he distinguishes two styles in earlyfifteenth-century Florence, the simple and the ornate, the styles of Masaccioand Gentile da Fabriano respectively. He suggests that the two stylesare related to two world-views, the progressive (rational, sober) and thefeudal (irrational, uncontrolled) and to two social classes, the upper middleclass and the nobility. At the macro-level he argues that a rational andrealistic art is to be found in periods dominated by the upper middle class.At the micro-level he argues that this upper middle class included the mostimportant patrons at a time when patrons told artists what to do. Thisseems a splendidly testable hypothesis; but in practice it runs the risk ofcircularity. When he finds Masaccio and Gentile working for father-in-lawand son-in-law, Antal, suggests that one class may take up the ideologyof another. But if the same middle class can be progressive or reaction-ary, how can one tell the two wings apart—except by looking at the paintingsthey commissioned (31)?

A more open approach to the sociology of art is offered by VytautasKavolis. In a collection of essays he discusses the possible relationshipbetween style and the economy, the political system, the social structure,images of the universe and social values. Like Antal and Fischer he emphas-ises the mediating role of world-views, and his detailed discussion of the

M. MEISS, Painting in Florence and Siena (Cambridge, Mass., 1939).after the Black Death (Princeton 1951); (30) E. FISCHER, Von der NotviendigkeitG. PELLES, Art Artists and Society (Engle- der Kunst (Dresden 1959), English trans.:wood Cliffs 1963); and, among his other Harmondsworth 1963, esp. pp. 118, 147.articles, M. SCHAPIRO, The Sculpture of (31) F. ANTAL, Florentine Painting andSouillac, in W. R. W. Koehler, Medieval its Social Background (London 1947), esp.Studies in Memory of A. Kingley Porter pp. 310 sq.

PETER BURKE

artistic preferences of different social classes is rather like an attempt totest Antal's central hypotheses. His exposition is lucid and takes accountof the conclusions of experimental psychologists and social anthropol-ogists as well as those of sociologists and art-historians. For example, hispolitical chapter discusses the hypothesis that democracies tend to favourart styles of 'informal spontaneity' and an egalitarian treatment of thedifferent parts of the work, whereas autocracies favour art-styles of 'formalrigidity' and the subordination of parts to one dominant element. Theevidence Kavolis brings to this chapter includes historical examples:Greece in the age of Pericles and XVth century Flanders among democracies,Louis XIV's France and Hitler's Germany among autocracies. But healso mentions an anthropological study of twenty-nine tribal societies, andsome evidence from social psychology to the effect that "a liking for balancedregularity of design has been found to be associated with the authoritarianpersonality" (32).

Kavolis holds a good balance between dogmatism and fear of specula-tion. He is prepared to make general assertions, and also to treat themas no more than working hypotheses. His system is an open one in thathe cites quite a large number of relevant factors, and suggests that theirrelative importance may differ from one society to another. At the sametime, his book leaves one dissatisfied. It gives an impression of remote-ness from any actual work of art, like a text-book built on secondary ratherthan primary sources. This is not an objection in principle to cross-cultural comparisons of artistic style. Where Heinrich Wolfflin did it,in his Principles of Art History, it succeeded marvellously, but where Kavolistalks about a 'rigid' or a 'painterly' style he is unconvincing, perhaps becausehe does not illustrate his generalisations. A similar objection might bemade to his treatment of social groups, like the 'upper middle' or 'newmiddle' class, irrespective of whether the culture is England or Japan.Again, this is not an objection in principle to generalisations about society,but rather to the somewhat mechanical way in which the general conceptsare handled. A third objection is that to focus on something as abstractas 'the economic factor' or 'the political factor' is not the most fruitfulway of approaching the relationships between art and society.

Has the sociologist any alternative ? One might stop and ask: whatwould Marx, Weber, or Durkheim have done ? All three of them weremuch more interested in history than many sociologists today, and on theanalogy of their work on literature, music and religion it seems reasonableto suggest that they would have concentrated on relating types of art totypes of society. This approach is surely the one which should be pur-sued, even though it runs up against two great obstacles, the need for satis-factory typologies of both art and society.

Relatively few art-historians and art-critics have been interested inartistic typologies. A significant exception is Heinrich Wolfflin, with

(32) V. KAVOLIS, Artistic Expression. A sociological analysis (Ithaca 1968).

THE SOCIOLOGY OF ART

his famous distinctions between "linear and painterly", between "planeand recession", between "closed and open form", and so on. Anotheris Eugenio D'Ors, whose idea that the baroque is a recurrent type of artwhich can be found in East and West, in classical times and in the XXthcentury, lends itself to sociological interpretation (33).

The second obstacle is the lack of a historical sociology, particularly of asociology of traditional societies. Suppose one were to try to write a bookon the sociology of art which tried to correlate types of art with types ofsociety. The section on primitive society and primitive art would not betoo difficult, because there is some interesting work by social anthropolo-gists such as Paul Bohannan and Jean Guiart which could be drawn togeth-er into a synthesis (34). Again, the section on art in industrial societiessince 1800 would not be too difficult to write; art and society have changedspectacularly and in the same parts of the world, so the only problem wouldbe to find plausible explanations of their inter-relations. Francastel'swork on impressionism, cubism, modern architecture and social changes isan illustration of what can be done here; so is Fischer on Romanticism.The real problem is how to deal with everything between primitive andindustrial, for which the blanket term 'traditional' is often used. One isconfronted with one type of society and an enormous variation in art-styles.Most of the world's great art—Greek sculpture, Gothic cathedrals, Renaiss-ance painting, Japanese woodcuts, Chinese ink-painting, Indian tem-ples, and so on—has been produced in 'traditional' societies. This maybe an argument for the autonomy of art, but it may equally well show thatsociologists have been using a concept too crude to be useful, and that whatis needed is to distinguish types within traditional society. Here liesthe value of Jean Duvignaud's recent essay which is built round typol-ogies of art and society, and performs the service of introducing intothe sociology of art the typology of traditional societies put forward bythe late Georges Gurvitch (35). His types include the following five: i)"Theocratic societies", such as ancient Egypt and the Inca Empire;ii) "Patriarchal communities", such as the societies in which the Homericpoems and the Old Testament were produced; Hi) "City-states", such asAthens and Florence; iv) "Feudal societies", such as Xlllth century Franceand Xlllth century Japan; v) "Centralised monarchies", such as Franceunder Louis XIV. There are enormous possibilities here, which Duvi-gnaud did not have the space to develop. As in the case of the Marxianmodel, one wants to ask: what are the mechanics of the connexion be-tween art and society ? Is it direct or indirect ? Does the artist have a

(33) H. WOLFFLIN, Kunstgeschichtliche whole symposium is extremely relevant toGrundbegriffe (1915), English trans.: Prin- this problem. J. GUIART, The Arts of theciples of Art History (New York 1950); South Pacific (Paris/London 1963).E. D'ORS, DU baroque (Paris 1935). (35) J. DuviGNAUD, Sociologie de Vart

(34) P. BOHANNAN, Artist and Critic in (Paris 1967). G. GURVITCH, Ditermi-an African Society, in M. W. SMITH, The nismes sociaux et liberty humaine (Paris 1955),Artist in Tribal Society (London 1961); the part iii.

153

PETER BURKE

limited autonomy, or none at all ? Duvignaud implies that society affectsstyle via aesthetic attitudes, and he produces a typology of these, includ-ing i) "absolute communion" as in tribal societies, ii) "illustration of dailylife" as in XVIIth-century Holland, Hi) "the esoteric" as in RenaissanceItaly, iv) "rebellion", and v) "conspicuous consumption". A point againstthis typology is that it seems to confuse attitudes to art with functions ofart. They may well be related, but it makes for clarity to keep them con-ceptually separate. Another interesting point of Duvignaud's is his devel-opment of the concepts 'anomie' and 'marginal man' to discuss problemsof creativity. It is the outsider who can see the world in a way which isnew for people living in a particular culture, and social change may makehis world-view a more relevant one for the future than that incorporatedin tradition. It is most easy to see how this theory applies to literature, butit may be relevant to the visual arts as well.

At times, like Kavolis, Duvignaud leaves an impression of thinness.This is partly the effect of his book's being a very short one. It is alsopartly due to the fact that every typology has its price—the exclusion ofsomething relevant and the consequent danger of circularity. But sometypologies are better than others, and in this field one needs a three-waydiscussion between art critics, sociologists and historians in order to arriveat the types which do the minimum violence to the art and to the societiesstudied. When there are more case-studies like the ones by Gombrichand Haskell, Meiss and Schapiro the task of constructing types will be thatmuch easier. But then the case-studies are likely to be sharper and morerelevant if their authors have reflected on the general hypotheses that theirwork might verify, falsify or qualify. There is a particular danger at themoment of a split between the sociological and historical approaches, adanger which has been late in arriving because of the underdevelopmentof this field, with its consequence that the sociologists have not yet stoppedmaking use of books written by historians. The dang2r is that historianswill concentrate on the microsocial history of art and leave the large-scalechanges in structure and values out of consideration, while the sociolo-gists discuss the macrosocial but have little to say about typologies of theartist's role or of the patronage system. If this split is avoided, somacredit will be due to a man whose work expressed his constant interest inboth the general and the particular, the microsocial and the macrosocial;to Pierre Francastel (36) *.

P E T E R B U R K E

* I should like to thank Professor Thomas Bottomore and the members of hisgraduate sociology seminar at the University of Sussex for helpful criticisms of anearlier draft of this paper.

(36) Evidence of the split: Duvignaud systems is put forward by J. M. B. EDWARDSand Kavolis focus on the macrosocial, in D. L. SILLS (ed.), International Encyclo-while Gombrich and Haskell discuss the paedia of the Social Sciences, vol. Illmicrosocial. But a typology of patronage p. 452.