SciELO - Scientific Electronic Library Online

vol.4 special edition author indexsubject indexarticles search
Home Pagealphabetic serial listing  

Horizontes Antropológicos

Print version ISSN 0104-7183

Horiz.antropol. vol.4 Porto Alegre  2008



The agency of Gell in the anthropology of art


Caleb Faria Alves

Federal University of Rio Grande do Sul – Brazil

Translated by Bruno Reinhardt
Translated from Horizontes Antropológicos, Porto Alegre, v.14, n.29, p. 315-338, Jan./June 2008.




While delineating the parameters for an anthropology of art, Alfred Gell's famous book, Art and Agency, overlooked, for the most part, anthropological tradition. This raises questions as embarrassing as they are ignored: is it possible to produce good theory with no references to achieved knowledge in this particular field? Are the subjects within anthropology so differently pursued that it is not possible to refer to a common way of approaching them? What exactly do we lose with such a self-centered narrative? My point of view is that theory cannot be treated like a list of sentences that can be added to one another, according to their isolated importance. This article proposes an analysis of Alfred Gell's narrative, of how he connects its propositions. I will examine, overall, his readings, the authors he quotes like Peirce, Sally Price and others, and how he fits them along his argumentation. The objective of this exercise is to put in evidence, beyond Gell's definitions, some conceptions about art contained in his formulations, and to enlarge the range of potential references available for an anthropology of art.

Keywords: Alfred Gell, anthropological theory, anthropology of art, symbolic conventions.


Ao delinear os parâmetros para uma antropologia da arte, o famigerado livro de Alfred Gell, Art and Agency, deixou de lado boa parte da antropologia, o que coloca algumas questões tão embaraçosas quanto pouco tratadas: pode-se fazer boa teoria sem contar com o acúmulo de conhecimento nessa área? Ou os temas recebem tratamento tão díspar que realmente não faz sentido falar mais em princípios teóricos comuns que podem e devem ser aplicados a qualquer objeto? O que perdemos com uma narrativa tão autocentrada? Partindo do pressuposto de que não é possível tratar teoria enquanto um conjunto de máximas que se somam ou se substituem, ou como um tabuleiro de peças que se acomodam umas às outras segundo sua validade isolada, este artigo procede a um exame da narrativa contida no livro de Gell, das articulações que realiza entre suas proposições. Examinarei, sobretudo, a leitura que faz dos autores que cita, como Peirce, Sally Price, e outros, e como os encaixa na sua argumentação. O objetivo desse exercício é evidenciar, para além de suas próprias definições, certas concepções sobre arte contidas na abordagem que Gell sugere e ampliar as referências para uma antropologia da arte.

Palavras-chave: Alfred Gell, antropologia da arte, convenções simbólicas, teoria antropológica.



Artworks are equivalent to persons, and art is a system of action. These are the axioms that underpin Alfred Gell's (1998) reframing of the anthropology of art. These ideas, central to his thought, are not necessarily counter-intuitive to anthropologists. He openly evokes other systems of thought in order to circumvent the common attribution of absolute passivity to human artifacts, otherwise assuming that these objects are real beings (and therefore, no longer objects). Gell therefore does not intend to limit himself to a theory of indigenous art; he boldly extends his thesis to Western artistic production. Crudely stating, it is tempting to accept and celebrate Gell's point of view, as it corroborates modernist theories that preach respect for different logics of thought (whose comprehension is, according to him, one of anthropology's task) and also richly contemplates agency, one of the most cherished concepts of contemporary anthropology. Recognizing the impact of his book, I would like to critically rescue his arguments as part of a simple exercise of methodological reconstruction1 whose aim is to ponder the parameters presently available for the anthropology of art.

Gell practically never quotes the great authors of the anthropological tradition. Evidently, a work that simply ignores the disciplinary history to which it belongs fatally incurs in inconsistencies. Otherwise, one could simply assume that our great pioneers have nothing to teach us, neither in terms of anthropology of art, nor in terms of the application to art of general principles of anthropological knowledge. These gaps cannot just be added to what Gell wrote. Indeed, thought itself is rarely cumulative. These absences, as I hope to demonstrate, in part sacrifice the very consistency of what Gell himself proposes; many ideas present in his work have in fact been already elaborated upon and in due depth by a variety of authors, including the very notion of agency. In this article I aim not only at reviewing this author's work, but at rendering evident the implicit meanings of some propositions and associations of ideas he relies upon. I intend to apprehend from them a more precise picture of how Gell understands art and what it entails, especially as it appears in modern society.

In this sense, my intention is neither to solve major questions of the anthropology of art nor restrict myself to the virtues of this book, already stressed by a large number of commentators, but to expand the horizons within which they should be considered. Generally, the impression I was left with after reading Gell's book does not differ from the one expressed in Robert Layton's (2003) critical article: the work is brilliant, at the same time that it shows serious structural problems, especially its neglecting of possible contributions from semiotics (as long as the due precautions with the linguistic model are taken). Hence, my point of departure is going to be Gell's proposition of an anthropology focused on how art operates in its context of production (and I would add as a first supplement: and use2), and whose conclusions are applicable to any society. The term "esthetics" will have a great centrality in my argument, as it sheds interesting light on Gell's understanding of art and of different areas of knowledge, including the specificity of anthropology.3

As I mentioned above, the book does not quote extensively from the anthropological literature on art. However, right on its second paragraph, Gell makes reference to the recent output of an important anthropologist. That is not exactly a consecrated work, such as those of Geertz or Levi-Strauss, but Sally Price's Primitive Art in Civilized Centers. The reasons behind this choice and this work's influence within the disciplinary field are not presented. Nevertheless, that is the token that works for Gell as a means to represent and criticize the state of art of the whole anthropology of art. This criticism is thus stated:

Anthropology, from my point of view, is a social science, not part of the humanities. The distinction is, I admit, elusive, but it does imply that the 'anthropology of art' focuses on the social context of art production, circulation, and reception, rather than on the evaluation of particular works of art, which, to my mind, is the function of the critic (Gell, 1998, p. 3).

Price's reference carries a clear rhetorical function, since Gell wants to distance himself from two propositions defended by the passage he transcribes: firstly, that the eye of the connoisseur is not immune,4 and, secondly, that the primitive's outlook is also endowed with its own characteristic discernments.5 The first point made by Price warns us to the risk of ethnocentrism, to how culturally embedded the esthetical judgment is; the second to the fact that the eye is also socially cultivated and it sees what it is trained to see. Gell does not disapprove these statements, but he also avoids understanding them as axiomatic to the anthropology of art. The finality of the general project contained in both assertions is deemed to be the elucidation of non-Western esthetical systems.

Following Price's reference, Gell proceeds to comment on Michaell Baxandall's work on the ways of seeing in Italian Renaissance as another representative of this tradition. Gell assumes this position momentarily only to express a reservation: if this commonality is true, the only distinction left between anthropology and history refer to the systems each discipline tries to elucidate. Gell's criticism carries a first subtle but still important equivocation, which introduces a series of imprecisions that wind up jeopardizing the scope of his claims: the sequence used to present his arguments infers that the esthetical system and the ways of seeing belonging to a cultural system are the same thing.     We already realize that these first passages both propose an interesting renewal of the discipline and overshadow conceptual mistakes – i.e., indistinctions between levels of analysis and tendentious examples – therefore making room for conservative practices and forms of thought that ultimately compromise the very construction of an anthropological theory along the lines Gell initially proposes. This aspect becomes even more evident if one pays closer attention to the book's first pages and to the elements they mobilize in order to build a distinction between Sally Price and himself, between "elucidating modes of seeing" and an anthropology attuned to social relations. The opening phrase of Art and Agency asserts that an anthropological theory of the visual arts is usually understood, firstly, as a theory of artistic production in the colonial and post-colonial societies typically studied by anthropologists, and additionally, as a theory about the so-called "primitive art" exhibited by museum collections, now renamed "ethnographic art". An anthropological theory of art would be a theory of art turned into anthropological art. Challenging these notions, he grounds his first axiom: that an anthropological theory of art cannot distinguish art produced in our society from that produced in other societies. Subsequently, he introduces Price, goes through Baxandall and arrives at the aforementioned proposition on social relations.

The indirect criticism of Baxandall is meaningful because the author is equally cited by Geertz (2000) in his famous essay Art as a Cultural System. There, it stands for an excellent raw model for an ethnography of artworks, an opinion corroborated by many other anthropologists. Recently, the same intellectual school that gave origin to Baxandall's work inspired Carlo Severi in a series of innovative texts belonging to the anthropology of art, image and memory. However, contrarily to what Gell wrote, to this historian a "way of seeing" is not equivalent to an esthetic system.  In fact, his book argues exactly against this association. The first sentence of Painting and Experience in Fifteenth-Century Italy states: "A fifteenth-century picture is a testimony of a social relationship" (Baxandall, 1991, p. 11). His central interests are, therefore, social relations. This distinction is fundamental for a very simple reason: the comprehension of a "way of seeing" is never enough to evaluate any artwork. The unveiling of the "visual education" of an epoch does not allow us to judge or presuppose judgments vis-à-vis the beauty of a particular work (assuming that the criterion of "beauty" is important at all). It only allows us to be able to understand what forces are mobilized by the act of appreciation. The analysis of judgments would be a different research step, which may or may not unfold according to the original objectives set by authors or texts. Baxandall simply reveals to his readers what was probably seen, regardless of whether those who saw it actually liked it, whether they found it technically well-executed, if their particular opinion had any relevance or if any esthetic parameter was ultimately mobilized by their opinion.

Moreover, despite the connections Gell wishes to establish between Price and Baxandall, one could argue that her book carries no solid attempt at understanding ways of seeing, unless negatively, as she only shows how the way we are used to approach primitive art is loaded with prejudices. Neither her research among art critics follows this path, and it could even be classified as an anthropology of professions, as the book converses only on their identities and their comprehensions of the work they perform. Therefore, differently from what Gell presents us, what Price intends is not to discuss the value of artworks, be they traditional or not; she is interested in the application of esthetic judgments, something which is fairly different from Gell's emphasis on classes of objects. Her conclusions elaborate on what the preconceptions commanding the application of our esthetic judgment can show about our society. Her own point is selection, not judgment. She does not even consider questioning if, and by which means, these values resonate outside of this collective – which would be in fact a quite interesting project.

Although I disagree with Gell's reading of Price and Baxandall, I do recognize that his critique of the former is not completely groundless. Her text lacks a robust analysis of social relations and is certainly fragile in this aspect. This absence is rendered clear when she engages in a survey of opinions without first situating her readers in terms of what they refer to. For instance, what is her interview with the museum security about? British common sense? The background and previous coursework required to work in a place like that? The police? Religious convictions? Despite its groundlessness, Price's general thesis, that museums and their collections have more to say about ourselves than about the "primitive" people they portray, is rich and instigating, being in some ways closer to what Gell himself proposes.

Relevant to the point I am attempting to make are the consequences of Gell's particular reading of these authors for the totality of his arguments: the way he expresses himself conveys the impression that to investigate the functioning of a social group is tantamount to assuming its values. Esthetics can perfectly be the object of anthropological analysis without the latter ultimately belonging to the field of esthetics. The issue elicited by this problem is to what extend would Gell not incur in the same mistakes he criticizes, that is, a discourse absent of social relations, if we understand them along the lines proposed by Marcel Mauss. Contrary to Mauss, his declared major inspiration, Gell does not incorporate the definition of art present in Ethnology and Anthropology, which is one of the main guidelines of our discipline. In this work, Mauss (1993, p. 9) proposes that art is that which is socially recognized as such6. There is nothing tautological about this definition. Mauss is concerned with how the institutionalization of practices ultimately defines what they are, being inscribed on their categorization. In other words, he is interested in the object's absorption into a classificatory system, which may or may not be shared by the researcher. With this orientation, Mauss wants to avoid the researcher's preconceptions about his object of inquire. We should not depart from any a priori identification of the object; we must rescue how it is conceptualized by those among whom it "lives". How do we do this? A main research tool is the understanding of nomenclatures, their meanings, the general system in which they are inserted and the social role of their contextual enunciations. From this perspective, within our society it would be absurd to avoid, as part of a debate on art, the reference to esthetics, artworks and art critique. In sum, instead of approaching in anthropological terms art criticism and the objects mobilized by the esthetic discourse, Gell opts for a general refusal. This however is inaccurate, hasty, and perhaps inconsistent, since it imposes undesirable limits to anthropological thought that convert it into a boundary of reality, and not a mere analytical perspective.

From this imprecision stem consequences which only reiterate these original slippages: according to Gell, the statute of art is irrelevant because the anthropology of art, aiming to be distinct from sociology, cannot restrict itself to the scope of what is officially recognized as art. It cannot, in fact, talk about the statute of the artwork, because these terms carry undesirable partial connotations: "An object which has been 'enfranchised' as an art object, becomes an art object exclusively, from the standpoint of theory, and can only be discussed in terms of the parameters of art-theory, which what being 'enfranchised' in this way is all about" (Gell, 1998, p. 12). The fact that something is noted as artistic does not make it an object of art only from the perspective of art theory, but also from those who deem critics to be authorities defining what art is or is not. This does not mean that these opinions are based on an obvious consensus, but that there is a set of experts, widely recognized as such, who project to the rest of society a particular comprehension of what is and what is not art, by means of their performance as critics, professors, curators, etc. This is also a meaningful social phenomenon, which deserves to be studied as much as any other. Furthermore, that is how any society or any sphere within it works. We cannot take Catholicism only as that which is defined by priests and bishops, but we cannot ignore the Church either. It ultimately depends on what our object is and how we circumscribe it. Elaborating on the previous example: to study priests, their discourse, upbringing, etc. is neither synonymous to a Catholic viewpoint nor a way of ignoring the fact that some people may recognize themselves as Catholics regardless of the Catholic Church and even against her.

The care that Gell embodies in his rejection of esthetic recognition, however, is extremely important and must be observed: we cannot transplant our discourse about the esthetical to other cultures and hold in their branches the fruits they are supposed to carry. Here we realize again the rare balance between brightness and naiveté. The premise is more than correct, and it is reproduced over and over through the various conceptual forms Gell introduces. However, the application of this principle, especially in regard to Western societies, is biased, as it supposes a mechanistic relation between the spheres of society and an excessive degree of homogeneity in Western culture. There is no equivalence between the attribution of the status of artwork to an object and its official recognition as such. The artistic attribution can take place even against official recognition. Here I agree with Layton when he highlights the interested and selective way whereby Gell mobilizes examples to corroborate his theory. Moreover, official recognition is never equivalent to institutional recognition. Art produced in the academy, for instance, can be opposed to what the State officially recognizes as being representative of the nation. But the situations met by the anthropologist during fieldwork are still more complex. An example are outsider artists who struggle for the same official recognition dedicated to canonic works, for their incorporation in museum collections and expositions, for art contests based on more inclusive criteria, and who engage in these disputes with no expert knowledge about esthetics or statecraft. With this attitude, Gell compromises mechanisms of circumscription of the object that are more suitable to the phenomena one finds on the ground.

In these terms, the status of artwork tends to become central if one studies societies that take this status as a reference, and this may be an interesting point of comparison between societies. I fully comprehend Gell's claim that the anthropological definition cannot be confused with the esthetic definition. What I have been striving to alert, though, is that he should make more evident the level of discourse and inquiry at which this refusal is necessary, and at which level its presence is indispensable. The truth of the matter is that the existence of the art object is impossible without a minimal recognition of this object as such. In this sense, the proximity between the official and the sociological discourses does not respond to the methodological principles of sociology or anthropology. This confusion reduces the complexity of social life to simplistic dichotomies. The central question is: we cannot ignore the social mechanisms that make something what it is; otherwise, we would be naturalizing the social life of objects as if their natures were ultimately defined by something intrinsic to them, entirely beyond the agency of men.

None of my arguments so far contradicts the incorporation of the art object's agency in research. This point, by the way, is not new. Merleau-Ponty (2004, p. 23) had already postulated it at least 50 years ago:

Things are not, therefore, simple neutral objects that we contemplate ahead of us; each of them symbolizes and evokes to us a certain conduct, they provoke on us favorable and unfavorable reactions, and that is why men's tastes, their characters, the attitude they assume toward the world and their exteriors can be read on the objects, the colors they prefer and the places they prefer to perambulate.

It is not necessary to lecture here on the centrality of this philosopher or the whole hermeneutic tradition to anthropology, which not only proposes that objects should be treated as persons, but goes even further to assert that language expresses our comprehension of objects in a spontaneous manner, a fact that must be accounted for by our current analysis:

Our relation with things is not a distant one; each of them speaks to our body and to our lives, as they are marked with human characteristics (docility, tenderness, hostility, resistance) and, inversely, they live in us as a multitude of emblems of the conducts we love or despise. Man is invested on things, and things are invested on him. To use the psychoanalyst jargon, things are complexes. That is what Cézanne wanted to say when he called to mind the "halo" things transmit on painting (Merleau-Ponty, 2004, p. 24).

With this attitude, cultural transposition gains much more interesting and responsible dimensions. The problem is no longer simply to understand the limits and possibilities of the agency of objects or esthetic concepts, but the whole set of terms and situations which concern objects, and that requires to be approached both comparatively and in its internal relations.7

Regarding the problem of the circumscription of the esthetic object, Shaeffer (2004, p. 25) remarks: "A peculiar aspect of this perspective is that it defines esthetic facts as a class of objects that is opposed to all other classes of objects". That would be the temptation behind the tendency to delimit a proper ontological class to these objects, instead of locating them into a subclass of objects in general. Esthetic objects share an esthetic property. It is exactly against this kind of supposition that anthropology evidences the agency that institutes art, which is later broken down through art critics' actions, journals, courses, etc.  The assumption that there is a truth about the object nested in itself is the basic principle underpinning ethnocentric prejudices, including the esthetic. From this supposedly embedded truth unfolds the idea that the recognition of an object as artistic is a matter of differential capability, manifested through taste, what winds up instituting a determined taste as the parameter for all others. Anthropology is dedicated to show precisely that there is no universal taste inhabiting any social configuration, but only patterns of culture (to use Mary Douglas' terms). From this perspective, one can only react with surprise to Gell's statement, also stressed by Layton (2003, p. 448), that he is working with an intuitive identification of artistic objects: "most of the art objects I shall actually discuss are well-known ones that we have no difficulty in identifying as 'art'; for instance, the Mona Lisa". If we take as a reference the consecrated, he is indeed correct; otherwise, the affirmation is absolutely false, as only a few have the skills needed to explain the reasons behind Mona Lisa's fame.

The central issue is to know how to approach a specific class of objects – artistic objects – without ethnocentrically ontologizing the world. We ought to be able to incorporate the agency of these objects while simultaneously denouncing them as arbitrary social constructs. In other words, to respect the native terms, that is, the life of the objects, means to treat them as entities whose social existence relies upon human action. I do not even believe in a final answer to this dilemma, as sought by Gell. It cannot be unraveled by a logical arrangement. Both principles are essential to anthropology, to the respect of cultural differences. We could not solve this question without placing an undue emphasis on the object itself, which is ultimately Gell's solution. It is telling that many of the analyses he undertakes in his book in order to illustrate his points virtually omit all the objects' context of existence. They are reduced to a very limited set of agency relations, strictly typified according to a certain moment and environment. Before we move on, it is necessary to examine Gell's answer in more detail.

The definition he suggests is not institutional, esthetic or semiotic; it is theoretical.

The art object is whatever is inserted into the 'slot' provided for art objects in the system of terms and relations envisaged in the theory (to be outlined later). Nothing is decidable in advance about the nature of this object, because the theory is premised on the idea that the nature of the art object is a function of the social-relational matrix in which it is embedded (Gell, 1998, p. 7).

Few were as straightforward as Gell in inserting their own names into the canons of an investigative current, in exercising an agency as important as his in a certain field of knowledge. Anthropology is a language, and the opening of new possibilities requires a lexicon. The theory he presents embraces the art object as follows:

I propose that 'art-like situations' can be discriminated as those in which the material 'index' (the visible, physical, 'thing') permits a particular cognitive operation which I identify as the abduction of agency (Gell, 1998, p. 13).

His definition of index is of an entity from which one can make a causal inference, or an inference about the intentions or capacities of another person (Gell, 1998, p. 13). For abduction, he understands an empirical rule created in order to render predictable that which otherwise would be mysterious (Gell, 1998, p. 14). A problem seems to be whether this formulation is enough to stakeout the specific contexts of research.8 In the pages that follow, Gell deepens his extrication of these terms. According to him, the index, being visible, can be an instrument or a result of social action. Agency exists in any situation where an intention is attributed to a person or a thing which triggers a causal series.

Gell claims to be drawing on Charles Sanders Peirce9, for whom the index is a causal inference of any sort or an inference about the intentions or capacities of someone else. The example he presents on page 15, in order to clarify his ideas and notions, is (and he acknowledges that) very recurrent: smoke is an index of fire. If there is smoke, one is allowed to assume that there is fire. If the same is proven for human action, then we have agency. Peirce also resorts to the smoke example, but before that he explains that an index is a sign which differs from an icon or a symbol. The index would be thus characterized:

An index is a sign which would, at once, lose the character which makes it a sign if its object were removed, but would not lose that character if there were no interpretant. Such, for instance, is a piece of mould with a bullet-hole in it as sign of a shot; for without the shot there would have been no hole; but there is a hole there, whether anybody has the sense to attribute it to a shot or not. A symbol is a sign which would lose the character which renders it a sign if there were no interpretant. Such is any utterance of speech which signifies what it does only by virtue of its being understood to have that signification (Peirce, [s.d.], p. 131).

By introducing the smoke example, Peirce targets, in fact, not a generic (or utilitarian) explanation but a specific one. He wishes to clarify the kind of connection between the individual object and the memory that would characterize the index. This connection is neither one of similarity, nor one of analogy, but is a dynamic one. "If A says to B 'there is a fire', B will ask, 'Where?'. Consequently, A will be forced to recur to an index, even though he is still referring to an indefinite place in a real universe, past or future" (Peirce, [s.d.], p. 131).

The example above shows how Gell, in fact, has an understanding of the index which differs from that which supposedly inspired him. Peirce, who is concerned with the type of connection between memory and individual object, does not define the index in physical terms. In the next page, indeed predicting this confusion, he anticipates himself: "The considerations above might have led the reader to imagine that indexes carry an exclusive reference to objects of experience and they would not be useful in the field of pure mathematics, which deals with, as it effectively does, with ideal creations, unconcerned with their concreteness." (Peirce, [s.d.], p. 132). One of the outcomes of this biased comprehension of the index is Gell's examples of agencies grounded on real objects and on the relations we establish with them. He hypothetically asks whether a little girl would easily throw her beloved doll away from a life boat in order to make room for her annoying brother. Of course not, he answers. And what is Michelangelo's David if not a doll for adults?, he concludes. According to him, the passage from a doll to an idol is very short. Why do we avoid such an obvious comparison and refuse to see something so evident? He answers that the comparison has an awkward effect on us not only because we are unwilling to compare ourselves to children, but because we feel uncomfortable to compare dolls to idols (Gell, 1998, p. 18). The object, for Peirce, is not concrete, but an indicator that might not be material nor linked to concrete experience. The adequate comparison, if we are to follow Peirce's lead, would be between the types of connection holding together the elements introduced by the little girl situation and the connection between anyone and David, not between the doll and David directly.

Anthropology has been concerned with the second kind of question. If we were to readdress Gell's problem, we would ask: would the attitude of this little girl be the same if she were a Catholic or a Protestant? Indian or European? How did she come to imagine that the doll was a friend? Who participates on this fantasy? Is there any sort of connection between how someone conceives a doll and the general relation with the objects produced in a certain culture? What kind of connection does she think she has with the doll? This line of inquiry concerned with the investigation of social events or facts was built in opposition to positivism, with or without biological background. To compare the attitude of an adult with that of a child in fact recalls an old-fashioned supposition that the relation between the first and the latter's universes is that of a higher or lower development of innate faculties. What it evokes is human nature, the actualization of a spontaneous, natural and universal tendency. The existence of the artifact "doll", its similitude with the human being and the relation between the child and the object need to be explained in the first place, before the actual analysis takes place. What this example shows, more deeply, is the common difficulty of transposing to art, as an anthropological object, some of the procedures we apply to any object.

Gell's attachment to the object has a motivation: to distance himself from the idea that to make anthropology of art is to unveil a language. He comments: "I believe that iconic representation is based on the actual resemblance in form between depictions and the entities they depict or are believed to depict" (Gell, 1998, p. 25). Once I read a book about gnomes that argued for their reality on the basis that many cultures, in diverse parts of the world, had described them as "elementals", as much in appearance as in habits, in spite of utilizing different names to address them and understanding them each in its own way.10 Apparently, there are two ways of explaining this truth about gnomes: either they really exist and those who do not believe in them are blind and in need of releasing their sensorial energies, or we are faced with a phenomenon of symbolic convention. Strictly speaking, anthropology has established itself by rejecting both alternatives. Since the debate on rationality in Evans-Pritchard and symbolic efficacy in Lévi-Strauss, anthropologists have focused on systems, aiming to prove that a culture's truth is the culture itself. Its isolated elements, whatever they are, need a foundation, as they always make reference to ways of living.

Refusal of both the real and the conventional is clearly incompatible, but only if they are placed at the same level of reflection. As principles of reality, we assume the truths of the groups we study, but as a principle of analysis, we cannot do this. We need to treat them as conventions, otherwise we would completely lose the anthropological reason, and our objects of research would disappear. Christ's image, for instance, has been the object of heated theological discussions. Many believe that when they see the picture, they see how Jesus actually looked like. Artists even understood their works as a revelation. The question thus seems to be: Can we do an anthropology that recognizes Christ's images without discussing the basis and mechanisms through which they were produced? Without first understanding certain conventions? Many are the downside dangers of assuming that this is really the appearance of the son of God. The first of them is to eventually incur in racism. Many segregationist Christian groups have as one of the grounds of their supposed racial superiority the axiom that the chosen people is white, and they prove this argument through some of these images. Thus, an important part of the struggle against prejudice is precisely to understand that representations are what they are, representations, a fact that does not incur on judgment in terms of falsity or lie. Although there are peoples subjected to less intense cultural contacts with distinct cultural groups, this does not mean that there are no distinctions and conflicts among them about different versions of these cultural phenomena.

Bruno Latour (1996), in Petite Réflexion sur le Culte Moderne des Dieux Faitiches, lectures on a very peculiar trace of Western modernity: we denounce the material and manufactured aspect of other cultures as if this was the proof of religious inefficacy and magical stratagem and, simultaneously, we resort to objects with religious functions that are similar to those we deem to be the product of primitive minds. By questioning others about the fallacy of similitude or the divine origin of an object, we do not resort to reasoning akin to the one we dedicate to ourselves. That is one example in which both possibilities are being mobilized at the same time: a group that argues for the similitude between an object and the entity, which downgrades another group for not being able to see that the similitude that they see binding the object to the supernatural entity is the product of a symbolic convention. What the researcher believes, in this case, is highly irrelevant. What matters is how the situation was concretely lived. No judgment is expected from the anthropologist, but only the analysis of social interactions (and here I am fully in agreement with Gell, although I think he is equivocated regarding what constitutes such relation).

Gell's dismissive attitude vis-à-vis symbolic convention follows the same path as his statements: he never places himself clearly in partnership or opposition vis-à-vis the great anthropologists; as a result, we do not know exactly to whom he addresses his critiques, and we know even less about the content of this criticism. It is much simpler to build a general critique to the idea of art as language than to utter a specific formulation. His main refusal, expressed in multiple moments in his work, has to do with deciphering a visual code of communication. Art cannot be defined in terms of what is and what is not integrated into this code, neither can anthropology can put to itself the task of deciphering or translating it. One of the most important authors employing this analogy is Geertz, although he does not properly follow any of the ways Gell believes this option incur. For Geertz (2000, p. 120), "To be of effective use in the study of art, semiotics must move beyond the consideration of signs as means of communication, code to be deciphered, to a consideration of them as modes of thought, idiom to be interpreted"; not a new "cryptography", i.e., a substitution of signals, but "a science that can determine the meaning of things for the life that surrounds them".

A much criticized aspect of Geertzian anthropology frequently mentioned by post-modernists is the extension of this code, or his attachment to a conception of society as a coherent and organized whole. That is also present in his formulations about art: "We could even argue that rites, myths and the organization of family life or the division of labor are actions that reflect the concepts developed by painting similarly to how painting reflects the concepts subjacent to social life" (Geertz, 2000, p. 102). When it touches this point, Gell's critique is pertinent, and yet outdated. Art does not necessarily reflect anything; it might establish a tense relation with other codes or even oppose them. Gell's book looks more consistent when we ignore these subtle passages that lead from a critique to another of the anthropological tradition, between the association of elements that seem very close but which carry no equivalence between the will to decipher and that of reconstituting a whole.

Effectively, Geertz (2000, p. 104) rejects deciphering, and that point is repeatedly made in his essay. He goes even further: for him, the relationship between the "symbolic elements […] that compose a semiotic system that, for generic reasons, we would like here to call esthetical, have an ideational – and not mechanical – connection with the society in which it presents itself". The argument makes evident how the idea of reflex carries for Geertz a double connotation: on the one hand, it insinuates a general coherence binding the spheres of social life; on the other, it argues that any translation involving art cannot possibly be but if not in art's own terms; that the relation between art and society must be established through some isolation of art as an autonomous sphere of investigation. In other terms, art is a point of entry to anthropological research as legitimate, self-sufficient and revealing as any other. Art is not explained by religion, by politics, nor by the general mood prevailing at a certain moment in a certain group. It has a life of its own. The relation with the other spheres of social life is, therefore, one of the first steps of investigation (if one ignores the continuity that Geertz calls the general experience of life, I would completely agree with this point). That is what Geertz seems to suggest when he differentiates, in general terms, art in the West and in the rest of the world from folk terminologies and those used by common language. The supposed isolation of art in the West is exactly the form whereby it is connected to the totality of society.

In regard to this specific point, Geertz is much bolder than Gell, and his proposals establish the question of the approximation with esthetics (on the latter's terms) as a primary issue. There are two ways of dealing with the relation between esthetics and anthropology. The first is through study, as the object of anthropology, in the same way that anthropology may take itself and any other area as an object of inquiry; the second is by dialoguing with aesthetics' literary canon. None of these options actually answers the question about esthetics being or not being a transcultural category. The first option accords to it a status equivalent to that of any other social group, such as punks or skateboarders.11 The second demands us to find invisible bridges between both fields of study, in case they actually exist (and I believe they do). A safeguard must be mentioned here. Aesthetics does not include a closed or defined opinion about anything; there is no consensus on the nature of its object, even less about how to approach it. To classify a production as esthetic requires the author to be aligned with a trend of thought from which we both may be allowed to talk and find the references to do so.

Hence, Gell's statement (1998, p. 3), "I believe that the desire to see the art of other cultures aesthetically tells us more about our own ideology and its quasi religious veneration of art objects as aesthetic talismans, than it does about other cultures" is slightly superficial when it comes to employ the term "aesthetic".12 Again Gell is too imprecise and eventually confuses distinct social groups in a significant misrepresentation of the constitution of our society. On the one hand, he is obviously correct when he exhorts us not to take part on a process of evaluation of primitive art and the one produced by our society; on the other, this function cannot be confused with aesthetics as a branch of philosophy. The art market, the academic world, museums and journals dedicated to the public are not a single group. They might even work with completely disparate criteria. A successful artist may not have any recognition of his production from aesthetics. Equally, a work highly esteemed by aesthetics might not have any repercussion or understandability for most people – which is, incidentally, what happens to many artworks. The question "Is aesthetics a transcultural category?" is an elusive one. It is necessary, in the first place, to define a reference for aesthetics, and second, to inquire if it is useful for anthropology and for the dialogue that it establishes internally and externally with other fields of knowledge, groups or peoples.13 An important point about utility is certainly the comprehension it generates, which demands significant care lest we do fall on the trap of ethnocentrism. In sum, and reiterating, the question itself is senseless and the possibility must not be disregarded.

Rather, Gell defines abduction as the distinctive criteria for his anthropology of art, a concept that I have not tackled yet. According to Peirce, it indicates the initial moment of an inductive process, in which a hypothesis is selected as a possible explanation of an empirical fact.14 The dictionary Aurélio defines it as an imperfect even though plausible reasoning, also referring to apagoge and violent rapture. This second set of definitions is akin to those found in the Webster's New World dictionary. I am not convinced that any of these alternatives is adequate to the kind of agency brought forth by an artwork. On page 29, and drawing from the notion of abduction, Gell creates a table that crosses types of agents and patients. The columns and rows contain the same elements: artist, index, prototype,15 recipient. The artist, in these terms, may occupy the position of either agent or patient. If in both there are two alternatives, as the agent he is the creative source, and as patient he is the eyewitness of the creative act.

The interactive processes binding people to each other and to objects of any sort seem to me to be far more complex than what this table conveys, as the previous example taken from Bruno Latour's book sought to demonstrate. A parallel with a consecrated theoretical framework, Max Weber's types of domination, might be useful here. They are powerful especially because of their simplicity and comprehensiveness, which prevents the interpreter from confusing them with the real. In Gell one sees the opposite. We have a total of 20 possibilities of relations involving art objects that intend to embrace all possible forms of interaction.

Considering all his suspicion vis-à-vis esthetics, Gell's analysis of Duchamp's work is contradictory, to say the least. Firstly, because he uncritically reproduces classifications created by critics, such as the realist, cubist and futurist tags. If Gell had had a more considerate attitude towards those that preceded him, he would have realized, pace Elias,16 that we should not surrender to these classification that easily. Secondly, he affirms that Duchamp has become a cubist belatedly and because of his satiric spirit, more than for any esthetic pretension. Gell seems to suppose that satire carries no esthetic possibility, which is definitely a more rigid position than that taken on by much of esthetic discourse, including Duchamp himself. Lastly, the idea that this artist consciously materialized ideas about the temporal flux in Russerl, and that he made possible and visible the object's agency, is not consistent. In fact, this flux might appear clear to academic artists. Before modernism, there was no consistent effort to break with esthetic norms and the construction of the artist's role as a vanguard that challenges all that precede it. Any painting presupposed a dialogue with others, and creativity was linked to the project of achieving new solutions to problems of composition based on particular works by those who were considered the great masters.

Roy Wagner, in a seminal book, revealed how the anthropologist does not capture "culture" as part of his exercise of investigation. Contrarily, he undertakes a construction that depends on his relation to one or other informants. I do not want to engage here in a long chain of citations about the contemporary critiques to the notion of culture, or to make a balance of post-modernism or other tendencies that incorporated these critiques, but only to express my surprise with an analysis that ignores these contributions. One of the main precepts of modern anthropology is that, as anyone else, the anthropologist cannot dialogue with this maximal, supra-personal, coherent, articulated, monolithic entity called culture. Duchamp, therefore, did not lead the culture of his time or interacted with it in a privileged manner. From a methodological point of view, the first question to be asked is where he exposed his work, how his works were received, by whom, when, in which circumstances, etc. This has been the path taken by much of contemporary production.17 On the contrary, what can be evidenced is that a person lives in her time, which might be interesting if one is willing to understand some common characteristics of particular times and means, but has nothing to say about this or that person or artwork in particular. I am not interested in proving wrong the characteristics Gell attributes to the works analyzed, but in showing that the history of battles is not only that of great generals, that geniality, as a potentiality of culture, is less individual than what is normally supposed, and that is what needs to be investigated and traced along its concrete realizations.18

I definitely agree with several points in Gell's book: that anthropology of art is supposed to do more than decipher codes, that it is not part of its competence to evaluate artistic works, that we must incorporate the agency of objects in more audacious ways than the usual – in this sense, I understand the celebrity achieved by his book. The manner through which he mobilized these perspectives from modern anthropology in order to propose a renewed approach to art is important and must be taken forward. However, his conceptual imprecision also elicits suspicion, as in the case of the concepts he borrowed from Peirce; the absence of anthropological theory, a cause of many important flaws in his overall argument; his distinction between levels of analysis and the principles that orient them, between methodology and theory. Lastly, it is a little unsettling the way Gell sets his focus on the object, taking art as a given (see his abovementioned analysis of Duchamp) and almost leading us back to positivistic comparisons between decontextualized elements. 

The starting point of any theory is the artistic object, regardless of its definition. The initial question of any investigation suggested by Mauss is still the best alternative: what makes something be considered what it is? Otherwise we would be committing, as Schaeffer (2004) has demonstrated, a dangerous ontologization of the object. This question is not incompatible with the incorporation of the object's agency, as it may appear. The first work I know that incorporates agency in a direct fashion is precisely Aby Warburg's ([s.d.]) ethnography among the Hopi, over a century ago. The main purpose of Warburg is to understand how it is constituted in a symbolic way; although, as Fritz Saxl ([s.d.], p. 149) argues,19 Warburg never questioned the manner in which the fusion between the lightning and the serpent during the ritual that became widely known as "The serpent's ritual" happens to the Indians themselves. For Warburg, as for the Indians, the lightning is the serpent, which clearly demonstrates that the inquiry of forms does not necessarily entail a disagreement with the ways of thinking that host them.

It is always good to recall the premonitory warning suggested by Cardoso (1986, p. 98) for any research:

However, this vogue of new techniques of investigation and the interest for the social actor in flesh and bone were not followed by a consistent theoretical-methodological critique. They answer to a malaise, a disapproval of fast generalizations and too abstract explicative schemes. But the return to the concrete has happened through the same paths first traced by positivist science.

What Cardoso denounces is how theoretical frameworks have become a declaration of principles, more than the construction of analytical references. Gell is very good at that. He indeed declares principles with which, in a good measure, I would agree. But I wonder if that is enough to establish the basis for an anthropology of art – which cannot refrain, as Layton (2003, p. 460) has stressed, from considering that art objects rely on their correct reading in order to become effective as secondary agents, a fact that demands a semiological approximation.



ABBAGNANO, Nicola. Dicionário de filosofia. São Paulo: Martins Fontes, 2007.

ALMEIDA, Kátia Maria Pereira de. Distinção e transcendência: a estética sócio-lógica de Pierre Bourdieu. Mana, v. 3, n. 1, p. 155-168, Rio de Janeiro, abr. 1997.

APPADURAI, Arjun. The social life of things. New York: The Cambridge University Press, 1990.

BAXANDALL, Michaell. O olhar renascente: pintura e experiência social na Itália da Renascença. Rio de Janeiro: Paz e Terra, 1991.

BELTING, Hans. O fim da história da arte: uma revisão dez anos depois. São Paulo: Cosac & Naify, 2006.

CARDOSO, Ruth C. L. Aventuras de antropólogos em campo ou como escapar das armadilhas do método. In: CARDOSO, Ruth (Org.). A aventura antropológica: teoria e pesquisa. Rio de Janeiro: Paz e Terra, 1986. p. 95-105.

ELIAS, Norbert. Mozart: Sociologia de um gênio. Rio de Janeiro: Jorge Zahar, 1995.

EUFRÁSIO, Mario A. Estrutura urbana e ecologia humana: a escola sociológica de Chicago (1915-1940). São Paulo: Editora 34, 1999.

GEERTZ, Clifford. Arte como sistema cultural. In: GEERTZ, Clifford. O saber local. Petrópolis: Vozes, 2000. p. 142-181.

GELL, Alfred. Art and agency: an anthropological theory. Oxford: Clarendon, 1998.

INGOLD, Tim. Aesthetics is a cross-cultural category? In: INGOLD, Tim (Org.). Key debates in antropology. London: Routledge, 1996. p. 249-293.

LATOUR, Bruno. Reflexão sobre o culto moderno dos deuses fe(i)tiches. São Paulo: Edusc, 2002.

LAYTON, Robert. Art and agency: a reassesssment. Journal of the Royal Antropological Institute, v. 9, n. 3, p. 447-464, 2003.

MAUSS, Marcel. Manual de etnografia. Lisboa: Don Quixote, 1993.

MERLEAU-PONTY, Maurice. Conversas – 1948. São Paulo: Martins Fontes, 2004.

NARAYAN, Kirin; ROSALDO, Renato. Creativity / Antropology. London: London University Press, 1993.

PEIRCE, Charles Sanders. Semiótica e filosofia. São Paulo: Cultrix, [s.d. ]

POORVLIET, Rien; HUYGEN, Wil. O livro secreto dos gnomos. São Paulo: Siciliano, 1993.

SAXL, Fritz. Le voyage de Warburg au Nouveau Mexique. In: WARBURG, Aby. Le rituel du sepent. Paris: Macula, [s.d.]. p 149-162.

SHAEFFER, Jean-Marie. Objets esthétiquees? L'Homme, n. 170, p. 25-46, 2004.

TORN, Cristina. Making the present, revealing the past: the mutability and continuity of tradition as process. Man: The journal of the Royal Antropological Institute, v. 32, n. 4, p. 696-717, dec. 1988.

WARBURG, Aby. Le rituel du sepent. Paris: Macula, [s.d. ].



Received 31/10/2007
Approved 03/03/2008



1 My reference for 'methodological reconstruction' is Mario A. Eufrásio (1999, p. 251): "A first possible characterization of methodological analysis is the process of identifying scientific discourse's constitutive elements (terms, statements, arguments, and formulations) in terms of their structure, functions and relations at various levels and aspects. In other words, of establishing modes of occurrence and relations structurally or dynamically among these elements, in their hierarchy, articulations, functions and 'ultimate' validity, and then objectively and meta-theoretically interpreting and assessing them." Strictly speaking, I am not methodologically reconstructing Gell's book, but rather incorporating this exercise to the content implicit in the association between the constitutive elements of scientific discourse.
2 An artwork's fruition does not hinge on the fact that it was produced in the environment where it was created. There is nothing artificial or illegitimate in the incorporation of products which are external to the group. See, for instance, Cristina Torn's (1988) instigating work on the appropriation of the Last Supper's images by Fijians.
3 Attention to the way Gell constructs his arguments implies less a complete examination of his book than a more microscopic look which is still representative of the whole. My focus will be on the early pages, where he lays bare the main core of his thesis with sequels and examples taken from the other chapters.
4 "[…] the eye of even the most naturally gifted connoisseur is not naked, but views art through the lens of a Western cultural education" (Price apud Gell, 1998, p. 2).
5 "[…] Primitives (including both artists and critics) are also endowed with discriminating eye […]" (Price apud Gell, 1998, p. 2).
6 In Mauss' own words: "Aesthetic phenomena form one of the largest components in the social activity of human beings, and not merely in their individual activity. An object, an action, a line of poetry is beautiful when it is recognized as beautiful by the majority of people of taste. This is what people call the grammar of art. All aesthetic phenomena are in some degree social phenomena" (p. 67).
7In this regard, perhaps even more than Sally Price, Arjun Appadurai (1990) would be an important reference to a broader theoretical construction about objects, whether artistic or not, in our society.
8 This is Gell's purpose, defined and introduced on page 4 of the same book as a criticism to Boas who, according to him, was not successful in this enterprise.
9 Robert Layton (2003) has adopted an approach to Gell's definitions which is at once broader and less precise. Besides Peirce, he resumes Saussure, Mounin and Umberto Eco, among others, in order to assess the importance of such definitions for anthropology. My criticism is different, though; as stated early on in this paper, it refers to the construction of Gell's argument. I have therefore departed from the importance that Gell himself attributes to the author he quotes.
10 This is the book O Livro Secreto dos Gnomos (Poortvliet; Huygen, 1993).
11 Kátia Maria Pereira de Almeida (1997, p. 3) underscores a much deeper and refined avenue of approximation in the tradition of Pierre Bourdieu: "In fact, as Miceli notes, Bourdieu seems to have found a way out of the dilemma pertaining to the dichotomy between 'object of knowledge' and 'real object', by acknowledging that the sociological foundations for the distinctions and categories used would themselves derive from the division of labor operating at a particular social formation." Therefore, even while defining its field of knowledge, aesthetics unveils particular social formations. The same holds true for anthropology, bringing both disciplines together in terms of a shared threshold of the division of knowledge and its relations to social formations.
12 On aesthetics as a transcultural category, see Ingold (1996).
13 What to say, for instance, about the broad debate which includes an art critic like Hans Belting, who incorporates anthropology and challenges the alleged universalism of the history of art through respect to cultural diversity? See Belting (2006).
14 See Nicola Abbagnano's (2007) dictionary of philosophy.
15A prototype is the entity which one believes to be representing (Gell, 1998, p. 26).
16 In Sociologia de um Gênio, Norbert Elias (1995) asserts that social events that were actually lived cannot be accounted for or even framed in the categories we use to distinguish among artistic processes.
17 Even though not exclusive, such questions are quite familiar to those who work with the notion of performance. This partly accounts for its importance; instead of departing from general data, pre-defined cuts of the real or pre-conceived categories, attention to performance guides the researcher to relations that are real and immediately lived in the extension that they have according to a criterion which is at once investigative and a datum of reality.
18 On this subject, see Narayan and Rosaldo (1993).
19 This is a recent edition which contains comments to Warburg's piece, including those by Saxl.