CHAPTER XIII
Split Representation in the Art of Asia and America
CONTEMPORARY ANTHROPOLOGISTS seem to be somewhat reluctant to undertake comparative studies of primitive art. We can easily understand their reasons. Until now, studies of this nature have tended almost exclusively to demonstrate cultural contacts, diffusion phenomena, and borrowings. The discovery of a decorative detail or an unusual pattern in two different parts of the world, regardless of the geographical distance between them and an often considerable historical gap, brought enthusiastic proclamations about common origin and the unquestionable existence of prehistoric relationships between cultures which could not be compared in other respects. Leaving aside some fruitful discoveries, we know to what abuses this hurried search for analogies “at any cost” has led. To save us from these errors, experts in material culture even now need to define the specific characteristics which distinguish a trait, trait complex, or style that may be subject to multiple independent recurrences from one whose nature and characteristics exclude the possibility of repetition without borrowing.
It is, therefore, with some hesitation that I propose to contribute several documents to a hotly and legitimately debated body of materials. This voluminous collection involves the Northwest Coast of America, China, Siberia, New Zealand, and perhaps even India and Persia. What is more, the documents belong to entirely different periods: the eighteenth and nineteenth centuries for Alaska; the first to second millennia b.c. for China; the prehistoric era for the Amur region; and a period stretching from the fourteenth to the eighteenth century for New Zealand. A more difficult case could hardly be conceived. I have mentioned elsewhere1 the almost insuperable obstacles generated by the hypothesis of pre-Columbian contacts between Alaska and New Zealand. The problem is perhaps simpler when one compares Siberia and China with North America: Distances are more reasonable and one need overcome only the obstacle of one or two millenia. Even in this case, however, and whatever the intuitive convictions which irresistibly sway the mind, what an immense marshalling of facts becomes necessary! For his ingenious and brilliant work, C. Hentze can be called the “scrap-collector” of Americanism, pulling his evidence together from fragments gathered from the most diverse cultures and often mounting insignificant details2 for exhibition. Instead of justifying the intuitive feeling of connectedness, his analysis dissolves it; nothing among these membra disjecta poetae appears to justify the deep sense of affinity which familiarity with both arts had so strongly elicited.
And yet, it is impossible not to be struck by the analogies presented by Northwest Coast and ancient Chinese art. These analogies derive not so much from the external aspect of the objects as from the fundamental principles which an analysis of both arts yields. This work was undertaken by Leonhard Adam, whose conclusions I shall summarize here.3 The two arts proceed by means of: (1) intense stylization; (2) schematization or symbolism, expressed by emphasizing characteristic features or adding significant attributes (thus, in Northwest Coast art, the beaver is portrayed by the small log which it holds between its paws); (3) depiction of the body by “split representation”; (4) dislocation of details, which are arbitrarily isolated from the whole; (5) representation of one individual shown in front view with two profiles; (6) highly elaborate symmetry, which often involves asymmetric details; (7) illogical transformation of details into new elements (thus, a paw becomes a beak, an eye motif is used to represent a joint, or vice-versa); (8) finally, intellectual rather than intuitive representation, where the skeleton or internal organs take precedence over the representation of the body (a technique which is equally striking in northern Australia).4 These techniques are not characteristic solely of Northwest Coast art. As Leonhard Adam writes, “The various technological and artistic principles displayed in both China and North West America are almost entirely identical.” 5
Once these similarities have been noted, it is curious to observe that, for entirely different reasons, ancient Chinese and Northwest Coast art have been independently compared with Maori art in New Zealand.6 This fact is the more remarkable when we note that Neolithic art of the Amur—some of whose themes (such as the bird, with wings unfolded, whose abdomen is formed by a solar face) are almost identical with themes of the Northwest Coast—exhibits, according to some scholars, “an unexpectedly rich, curvilinear ornamentation related to that of the Ainu and Maori on one side and to the Neolithic cultures of China (Yangshao) and Japan (Jomon) on the other; consisting particularly of that type of ribbon ornamentation characterized by complex motifs such as the weave, spiral and meander in contradistinction to the rectangular geometric decoration of the Baikalian culture.” 7 Thus art forms from very different regions and periods which exhibit obvious analogies suggest, each of them and for independent reasons, relationships which are, however, incompatible with geographical and historical requirements.
Do we rest, then, on the horns of a dilemma which condemns us either to deny history or to remain blind to similarities so often confirmed? Anthropologists of the diffusionist school did not hesitate to force the hand of historical criticism. I do not intend to defend their adventurous hypotheses, but it must be admitted that the negative attitude of their cautious opponents is no more satisfactory than the fabulous pretensions which the latter merely reject. Comparative studies of primitive art have probably been jeopardized by the zeal of investigators of cultural contacts and borrowings. But let us state in no uncertain terms that these studies have been jeopardized even more by intellectual pharisees who prefer to deny obvious relationships because science does not yet provide an adequate method for their interpretation. The rejection of facts because they appear to be unintelligible is surely more sterile from the viewpoint of scientific progress than the formulation of hypotheses. Even if these should prove to be unacceptable, they will elicit, precisely because of their inadequacy, the criticism and research that will one day enable us to progress beyond them.8
We reserve, therefore, the right to compare American Indian art with that of China or New Zealand, even if it has been proved a thousand times over that the Maori could not have brought their weapons and ornaments to the Pacific Coast. Cultural contact doubtless constitutes the one hypothesis which most easily accounts for complex similarities that chance cannot explain. But if historians maintain that contact is impossible, this does not prove that the similarities are illusory, but only that one must look elsewhere for the explanation. The fruitfulness of the diffusionist approach derives precisely from its systematic exploration of the possibilities of history. If history, when it is called upon unremittingly (and it must be called upon first), cannot yield an answer, then let us appeal to psychology, or the structural analysis of forms; let us ask ourselves if internal connections, whether of a psychological or logical nature, will allow us to understand parallel recurrences whose frequency and cohesion cannot possibly be the result of chance. It is in this spirit that I shall now present my contribution to the debate.
Split representation in Northwest Coast art has been described by Franz Boas as follows: “The animal is imagined cut in two from head to tail . . . there is a deep depression between the eyes, extending down the nose. This shows that the head itself must not have been considered a front view, but as consisting of two profiles which adjoin at mouth and nose, while they are not in contact with each other on a level with the eyes and forehead . . . either the animals are represented as split in two so that the profiles are ioined in the middle, or a front view of the head is shown with two adjoining profiles of the body.” 9 Boas analyzes the two paintings reproduced here (Figures 17 and 18, which correspond respectively to Figures 222 and 223 in the text cited) in the following terms:
Figure 222 (a Haida painting) shows a design which has been obtained in this manner. It represents a bear. The enormous breadth of mouth observed in these cases is brought about by the junction of the two profiles of which the head consists. This cutting of the head is brought out most clearly in the painting Figure 223, which also represents the bear. It is the painting on the front of a Tsimshian house, the circular hole in the middle of the design being the door of the house. The animal is cut from back to front, so that only the front part of the head coheres. The two halves of the lower jaw do not touch each other. The back is represented by the black outlines on which the hair is indicated by fine lines. The Tsimshian call such a design “bears meeting,” as though two bears had been represented.10
FIGURE 17. Haida. Painting representing a bear. (After F. Boas, Primitive Art, Fig. 222.)
FIGURE 18. Tsimshian. Painting representing a bear; from the front of a house. (After Boas, op. cit., Fig. 223.)
Let us now compare this analysis with that given by H. G. Creel of a similar technique in the art of ancient China: “One of the most distinctive characteristics of Shang decorative art is a peculiar method by which animals were represented in flat or in rounded surfaces. It is as if one took the animal and split it lengthwise, starting at the tip of the tail and carrying the operation almost, not quite, to the tip of the nose, then the two halves are pulled apart and the bisected animal is laid out flat on the surface, the two halves joined only at the tip of the nose.” 11 The same author, who apparently does not know Boas’ work, after having employed almost exactly the same terminology as the latter, adds: “In studying Shang design I have constantly been aware of the feeling that this art has great resemblance, certainly in spirit and possibly in detail, to that of . . . the Northwest Coast Indians.” 12
FIGURE 19. Bronze discovered near An-Yang (China). In the middle panel, a split t’ao t’ieh mask without a lower jaw. The ears make up a second mask above the first. The eyes in the second mask may also be seen as belonging to two small dragons represented by the ears of the first mask. The two small dragons are shown in profile and face to face, like those in the upper panel. The latter may in turn be seen as a ram mask shown in front view, the horns being represented by the bodies of the dragons. The design on the lid can be similarly interpreted. (After W. P. Yetts, An-Yang. A Retrospect.)
PLATE I. Shang bronze, China, 1766–1122 B.C. ( After W. P. Yetts, The George Eumorphopoulos Collection Catalogue.)
PLATE II. Storage box representing a frog, Northwest Coast, nineteenth century. (Collection of Dr. J. Lacan.)
PLATE III. Haida. Painting representing a shark. The head is shown in front view to bring out the features characteristic of the shark, but the body is split lengthwise, with the two halves laid out flat on the surface to the right and left of the head, (After Bureau of American Ethnology, Tenth Annual Report, Plate XXV.)
PLATE VI. Caduveo woman with painted face. Drawing by Boggiani, an Italian painter who visited the Caduveo in 1892. (After G. Boggiani, Viaggi d’un artista nell’ America Meridionale)
PLATE VII. Maori chief’s drawing representing his own tattoed face.(After H. G. Robley, Moko, or Maori Tattoing.)
PLATE VIII. Caduveo woman’s drawing representing a figure with painted face. (Author’s collection)
PLATE IX. Jade figure (tiki), New Zealand, characterized by the three-lobed division of the face. (Courtesy of the American Museum of Natural History.)
PLATE X. Maori wood carving, New Zealand, eighteenth century (?). (After A. Hamilton, The Art Workmanship of the Maori in New Zealand.)
PLATE XI. Head ornament, carved wood, Northwest Coast, nineteenth century. Note the two small human heads which adorn the solar plexus and the abdomen as well as the cleft end of the sternum. (Formerly in the author’s collection.)
PLATE XII. Three Maori tattooing designs carved in wood, late nineteenth century. (Top row:) Men’s faces. (Bottom row:) Woman’s face. (After Hamilton, op. cit.)
PLATE XIII. Three Maori wood carvings, eighteenth or nineteenth century. (After Hamilton, op. cit.)
This distinctive technique, which is found in ancient Chinese art, among the Siberian primitives, and in New Zealand, also appears at the other extremity of the American continent, among the Caduveo Indians. A drawing, which we reproduce here in Figure 20, represents a face painted according to the traditional custom of the women of this small tribe of southern Brazil, one of the last remnants of the once flourishing Guaicuru nation. I have described elsewhere how these paintings are executed and what their function is in the native culture.13 For present purposes it is, therefore, sufficient to recall that these paintings have been known since the first contacts with the Guaicuru in the seventeenth century and that they do not seem to have evolved since that time. They are not tattooings, but cosmetic facial paintings, which must be renewed after a few days and which are executed with a wooden spatula dipped in the juices of wild fruit and leaves. The women, who paint one another’s faces (and who formerly also painted men), do not work from a model but improvise within the limits of a complex, traditionally defined range of themes. Among four hundred original drawings gathered in the field in 1935, I did not find two alike. The differences, however, stem more from the ever-varied arrangement of fundamental elements than from a renewal of these elements—whether simple and double spirals, hatching, volutes, frets, tendrils, or crosses and whorls. The possibility of Spanish influence should be excluded, given the remote date when this refined art was described for the first time. At present, only a few old women possess the ancient skill, and it is not difficult to foresee the time when it will have disappeared altogether.
Plate VIII (after p. 250) presents a good example of these paintings. The design is built symmetrically in relation to two linear axes, one of them vertical, following the median plane of the face, the other horizontal, dividing the face at eye level. The eyes are schematically represented on a reduced scale. They are used as starting points for two inverted spirals, one of which covers the right cheek and the other the left side of the forehead. A motif in the shape of a compound bow, which is located in the lower part of the painting, represents the upper lip and is applied on it. We find this motif, more or less elaborated and more or less transformed, in all the facial paintings, where it seems to constitute a constant element. It is not easy to analyze the design, because of its apparent asymmetry—which, nonetheless masks a real, though complex, symmetry. The two axes intersect at the root of the nose, thus dividing the face into four triangular sections: left side of the forehead, right side of the forehead, right wing of the nose and right cheek, and left wing of the nose and left cheek. Opposite triangles have a symmetrical design, but the design within each triangle itself is a double design, which is repeated in inverted form in the opposite triangle. Thus, the right side of the forehead and the left cheek are covered, first by a triangle of frets, and, after a separation in the form of an empty oblique strip, by two double spirals in alignment, which are decorated with tendrils. The left side of the forehead and the right cheek are decorated with a simple large spiral adorned with tendrils; it is topped by another motif in the shape of a bird or flame, which contains an element reminiscent of the empty oblique stripe in the opposite design. We thus have two pairs of themes, each of which is repeated twice in symmetrical fashion. But this symmetry is expressed either in relation to one of the two horizontal and vertical axes, or in relation to the triangles defined by the bisection of these axes. While far more complex, this pattern recalls that of playing cards. Plates IV, V, and VI are other examples which illustrate variations on what is fundamentally the same pattern.
FIGURE 20. Caduveo. Facial design reproduced by a native woman on a sheet of paper. (Author’s collection.)
In Plate VIII, however, it is not only the painted design which draws the attention. The artist, a woman approximately thirty years old, intended also to represent the face and even the hair. Now she obviously accomplished this by split representation: The face is not really seen in a frontal view; it consists of two joined profiles. This explains its extraordinary width and its heart-shaped outline. The depression dividing the forehead into two halves is a part of the representation of the profiles, which merge only from the root of the nose down to the chin. A comparison of Figures 17 and 18 and Plate VIII shows that this technique is identical with that used by artists of the Northwest Coast of America.
Other important traits are also characteristic of both North and South American art. We have already stressed the dislocation of the subject into elements which are recombined according to conventional rules having nothing to do with nature. Dislocation is just as striking in Caduveo art, where it takes, however, an indirect form. Boas minutely described the dislocation of bodies and faces in Northwest Coast art: The organs and limbs themselves are split and used to reconstitute an arbitrary individual. Thus, in a Haida totem pole, “the figure must be . . . explained in such a way that the animal is twisted twice, the tail being turned up over the back, and the head being first turned down under the stomach, then split and extended outward.” 14 In a Kwakiutl representation of a killer whale (Orca sp.), “the animal has been split along its whole back towards the front. The two profiles of the head have been joined. . . . The dorsal fin, which according to the methods described heretofore [split representation] would appear on both sides of the body, has been cut off from the back before the animal was split, and appears now placed over the junction of the two profiles of the head. The flippers are laid along the two sides of the body, with which they cohere only at one point each. The two halves of the tail have been twisted outward so that the lower part of the figure forms a straight line.” 15 See Figure 21. These examples could easily be multiplied.
FIGURE 21. Kwakiutl. Painting for the front of a house, the design representing a killer whale. (After F. Boas, op. cit., Fig. 247.)
Caduveo art carries the dislocation process both further than, yet not as far as, Northwest Coast art. It does not carry it as far, because the face or body on which the artist works is a flesh-and-bone face and body, which cannot be taken apart and put together again. The integrity of the real face is thus respected, but it is dislocated just the same by the systematic asymmetry by means of which its natural harmony is denied on behalf of the artificial harmony of the painting. But since this painting, instead of representing the image of a deformed face, actually deforms a real face, the dislocation goes further than in the case previously described. The dislocation here involves, besides the decorative value, a subtle element of sadism, which at least partly explains why the erotic appeal of Caduveo women (expressed in the paintings) formerly attracted outlaws and adventurers toward the shores of the Paraguay River. Several of these now aging men, who intermarried with the natives, described to me with quivering emotion the nude bodies of adolescent girls completely covered with interlacings and arabesques of a perverse subtlety. The tattooings and body paintings of the Northwest Coast, where this sexual element was probably lacking and whose symbolism, often abstract, presents a less decorative character, also disregarded symmetry in the human face.16
In addition, we observe that the arrangement of Caduveo paintings around a double axis, horizontal and vertical, divides the face according to a process of double splitting, so to speak—that is, the painting recombines the face not into two profiles but into four quarters (see Figure 20). Asymmetry serves the formal function of insuring the distinction between quarters, which would merge into two profiles if the fields were to be symmetrically repeated to the right and left instead of being joined by their tips. Dislocation and splitting are thus functionally related.
If we pursue this comparison between Northwest Coast and Caduveo art, several other points are worthy of consideration. In each case, sculpture and drawing provide the two fundamental means of expression; in each case, sculpture presents a realistic character, while drawing is more symbolic and decorative. Caduveo sculpture is probably limited, at least during the historical period, to fetishes and representations of gods, which are always of small size, in contrast to the monumental art of Canada and Alaska. But the realistic character and the tendency toward both portrait and stylization are the same, as well as the essentially symbolic meaning of drawn or painted motifs. In both cases, masculine art, centered on sculpture, expresses its representational intention, while feminine art—limited to weaving and plaiting on the Northwest Coast, but also including drawing among these natives of southern Brazil and Paraguay—is a non-representational art. This is true, in both cases, for textile motifs; as regards the Guaicuru facial paintings, we know nothing about their archaic character. It is possible that the themes of these paintings, whose import has become lost today, formerly had a realistic or at any rate symbolical meaning. Northwest Coast and Caduveo art both carry out decoration by means of stencils, and create ever-new combinations through the varied arrangement of basic motifs. Finally, in both cases, art is intimately related to social organization: Motifs and themes express rank differences, nobility privileges, and degrees of prestige. The two societies were organized along similar hierarchical lines and their decorative art functioned to interpret and validate the ranks in the hierarchy.17
I should now like to make a brief comparison between Caduveo art and another art which also used split representation—that of the Maori of New Zealand. Let us first recall that Northwest Coast art has been frequently compared, for other reasons, to the art of New Zealand. Some of these reasons turned out to be specious—for instance, the apparently identical character of woven blankets used in the two areas. Others seem more valid—for example, those deriving from the similarity between Alaskan clubs and the Maori patu mere. I have mentioned this enigma elsewhere.18
The comparison of Maori with Guaicuru art is based on other convergences. In no other region of the world has facial and corporal decoration attained such high levels of development and refinement. Maori tattooings are well known. I reproduce four of them (Plates VII and XII), which may be fruitfully compared with the photographs of Caduveo faces.
The analogies between them are striking: complexity of design, involving hatching, meanders, and spirals (the spirals are often replaced in Caduveo art by frets, which suggest Andean influences); the same tendency to fill the entire surface of the face; and the same localization of the design around the lips in the simpler types. The differences between the two arts must also be considered. The difference due to the fact that Maori design is tattooed whereas Caduveo design is painted may be dismissed, since there is hardly any doubt that in South America, too, tattooing was the primitive technique. Tattooing explains why the Abipone women of Paraguay, as late as the eighteenth century, had “their face, breast, and arms covered with black figures of various shapes, so that they present the appearance of a Turkish carpet.” 19 This made them, according to their own words as recorded by the old missionary “more beautiful than beauty itself.” 20 On the other hand, one is struck by the rigorous symmetry of Maori tattooings, in contrast with the almost licentious asymmetry of some Caduveo paintings. But this asymmetry does not always exist; and I have shown that it resulted from a logical development of the splitting principle. It is thus more apparent than real. It is clear, nevertheless, that as regards typological classification, Caduveo facial designs occupy an intermediary position between those of the Maori and those of the Northwest Coast. Like the latter, they have an asymmetrical appearance, while they present the essentially decorative character of the former.
This continuity is also apparent when one considers the psychological and social implications. Among the Maori, as among the natives of the Paraguayan border, facial and corporal decoration is executed in a semi-religious atmosphere. Tattooings are not only ornaments. As we already noted with respect to the Northwest Coast (and the same thing may be said of New Zealand), they are not only emblems of nobility and symbols of rank in the social hierarchy; they are also messages fraught with spiritual and moral significance. The purpose of Maori tattooings is not only to imprint a drawing onto the flesh but also to stamp onto the mind all the traditions and philosophy of the group. Similarly, the Jesuit missionary Sanchez Labrador has described the passionate seriousness with which the natives devoted whole days to letting themselves be painted. He who is not painted, they said, is “dumb.” 21 And, like the Caduveo, the Maori use split representation. In Plates VII, IX, X, and XIII, we note the same division of the forehead into two lobes; the same representation of the mouth where the two halves meet; the same representation of the body, as though it had been split in the back from top to bottom and the two halves brought forward on the same plane. We note, in other words, all the techniques which are now familiar to us.
How shall we explain the recurrence of a far from natural method of representation among cultures so widely separated in time and space? The simplest hypothesis is that of historical contact or independent development from a common civilization. But even if this hypothesis is refuted by facts, or if, as seems more likely, it should lack adequate evidence, attempts at interpretation are not necessarily doomed to failure. I shall go further: Even if the most ambitious reconstructions of the diffusionist school were to be confirmed, we should still be faced with an essential problem which has nothing to do with history. Why should a cultural trait that has been borrowed or diffused through a long historical period remain intact? Stability is no less mysterious than change. The discovery of a unique origin for split representation would leave unanswered the question of why this means of expression was preserved by cultures which, in other respects, evolved along very different lines. External connections can explain transmission, but only internal connections can account for persistence. Two entirely different kinds of problems are involved here, and the attempt to explain one in no way prejudges the solution that must be given to the other.
One observation immediately follows from the comparison between Maori and Guaicuru art. In both cases, split representation appears as a consequence of the importance that both cultures ascribe to tattooing. Let us consider Plate VIII again and ask ourselves why the outline of the face is represented by two joined profiles. It is clear that the artist intended to draw, not a face, but a facial painting; it is upon doing the latter that she concentrated all her attention. Even the eyes, which are sketchily indicated, exist only as points of reference for starting the two great inverted spirals into whose structure they merge. The artist drew the facial design in a realistic manner; she respected its true proportions as if she had painted on a face and not on a flat surface. She painted on a sheet of paper exactly as she was accustomed to paint on a face. And because the paper is for her a face, she finds it impossible to represent a face on paper, at any rate without distortion. It was necessary either to draw the face exactly and distort the design in accordance with the laws of perspective, or to respect the integrity of the design and for this reason represent the face as split in two. It cannot even be said that the artist chose the second solution, since the alternative never occurred to her. In native thought, as we saw, the design is the face, or rather it creates it. It is the design which confers upon the face its social existence, its human dignity, its spiritual significance. Split representation of the face, considered as a graphic device, thus expresses a deeper and more fundamental splitting, namely that between the “dumb” biological individual and the social person whom he must embody. We already foresee that split representation can be explained as a function of a sociological theory of the splitting of the personality.
The same relationship between split image and tattooing may be observed in Maori art. If we compare Plates VII, IX, X, and XIII, we will see that the splitting of the forehead into two lobes is only the projection, on a plastic level, of the symmetrical design tattooed on the skin.
In the light of these observations, the interpretation of split representation proposed by Boas in his study of Northwest Coast art should be elaborated and refined. For Boas, split representation in painting or drawing would consist only in the extension to flat surfaces of a technique which is naturally appropriate in the case of three-dimensional objects. When an animal is going to be represented on a square box, for instance, one must necessarily distort the shape of the animal so that it can be adapted to the angular contours of the box. According to Boas,
In the decoration of silver bracelets a similar principle is followed but the problem differs somewhat from that offered in the decoration of square boxes. While in the latter case the four edges make a natural division between the four views of the animal,—front and right profile, back and left profile,—there is no such sharp line of division in the round bracelet, and there would be great difficulty in joining the four aspects artistically, while two profiles offer no such difficulty. . . . The animal is imagined cut in two from head to tail, so that the two halves cohere only at the tip of the nose and at the tip of the tail. The hand is put through this hole and the animal now surrounds the wrist. In this position it is represented on the bracelet. . . . The transition from the bracelet to the painting or carving of animals on a flat surface is not a difficult one. The same principle is adhered to. . . .22
Thus the principle of split representation would gradually emerge in the process of transition from angular to rounded objects and from rounded objects to flat surfaces. In the first case, there is occasional dislocation and splitting; in the second case, splitting is systematically applied, but the animal still remains intact at the level of the head and the tail; finally, in the third case, dislocation goes to the extreme of splitting the caudal tie, and the two halves of the body, now free, are folded forward to the right and left on the same plane as the face.
This treatment of the problem by the great master of modern anthropology is remarkable for its elegance and simplicity. However, this elegance and simplicity are mainly theoretical. If we consider the decoration of flat and rounded surfaces as special cases of the decoration of angular surfaces, then nothing has been demonstrated with respect to the latter. And, above all, no necessary relationship exists a priori, which implies that the artist must remain faithful to the same principle in moving from angular to rounded surfaces, and from rounded to flat surfaces. Many cultures have decorated boxes with human and animal figures without splitting or dislocating them. A bracelet may be adorned with friezes or in a hundred other ways. There must, then, be some fundamental element of Northwest Coast art (and of Guaicuru art, and Maori art, and the art of ancient China) which accounts for the continuity and rigidity with which the technique of split representation is applied in them.
We are tempted to perceive this fundamental element in the very special relationship which, in the four arts considered here, links the plastic and graphic components. These two elements are not independent; they have an ambivalent relationship, which is simultaneously one of opposition and one which is functional. It is a relationship of opposition because the requirements of decoration are imposed upon the structure and change it, hence the splitting and dislocation; but it is also a functional relationship, since the object is always conceived in both its plastic and graphic aspects. A vase, a box, a wall, are not independent, pre-existing objects which are subsequently decorated. They acquire their definitive existence only through the integration of the decoration with the utilitarian function. Thus, the chests of Northwest Coast art are not merely containers embellished with a painted or carved animal. They are the animal itself, keeping an active watch over the ceremonial ornaments which have been entrusted to its care. Structure modifies decoration, but decoration is the final cause of structure, which must also adapt itself to the requirements of the former. The final product is a whole: utensil-ornament, object-animal, box-that-speaks. The “living boats” of the Northwest Coast have their exact counterparts in the New Zealand correspondences between boat and woman, woman and spoon, utensils and organs.23
We have thus pushed to its most abstract expression the study of dualism, which has been commanding our attention with increasing persistence. We saw in the course of our analysis that the dualism between representational and non-representational art became transformed into other kinds of dualism: carving and drawing, face and decoration, person and impersonation, individual existence and social function, community and hierarchy. We are thus led to acknowledge a dualism, which is also a correlation, between plastic and graphic expression, which provides us with a true “common denominator” of the diverse manifestations of the principle of split representation.
In the end, our problem may be formulated as follows: Under what conditions are the plastic and graphic components necessarily correlated? Under what conditions are they inevitably functionally related, so that the modes of expression of the one always transform those of the other, and vice versa? The comparison between Maori and Guaicuru art already provided us with the answer to the latter question. We saw, indeed, that the relationship had to be functional when the plastic component consisted of the face or human body and the graphic component of the facial or corporal decoration (painting or tattooing), which is applied to them. Decoration is actually created for the face; but in another sense the face is predestined to be decorated, since it is only by means of decoration that the face receives its social dignity and mystical significance. Decoration is conceived for the face, but the face itself exists only through decoration. In the final analysis, the dualism is that of the actor and his role, and the concept of mask gives us the key to its interpretation.
All the cultures considered here are, in fact, mask cultures, whether the masking is achieved predominantly by tattooing (as is the case for the Guaicuru and Maori) or whether the stress is placed literally on the mask, as the Northwest Coast has done in a fashion unsurpassed elsewhere. In archaic China, there are many references to the ancient role of masks, which is reminiscent of their role in Alaskan societies. Thus, the “Impersonation of the Bear” described in the Chou Li, with its “four eyes of yellow metal,” 24 recalls the multiple masks of the Eskimo and Kwakiutl.
Those masks with louvers, which present alternately several aspects of the totemic ancestor—sometimes peaceful, sometimes angry, at one time human, at another time animal—strikingly illustrate the relationship between split representation and masquerade. Their function is to offer a series of intermediate forms which insure the transition from symbol to meaning, from magical to normal, from supernatural to social. They hold at the same time the function of masking and unmasking. But when it comes to unmasking, it is the mask which, by a kind of reverse splitting, opens up into two halves, while the actor himself is dissociated in the split representation, which aims, as we saw, at flattening out as well as displaying the mask at the expense of the individual wearing it.
Our analysis thus converges with that of Boas, once we have explored its substructure. It is true that split representation on a flat surface is a special case of its appearance on a rounded surface, just as the latter is itself a special case on three-dimensional surfaces. But not on any three-dimensional surface; only on the three-dimensional surface par excellence, where the decoration and form cannot be dissociated either physically or socially, namely, the human face. At the same time, other curious analogies between the various art forms considered here are illuminated in a similar way.
In the four arts, we discover not one but two decorative styles. One of these styles tends toward a representational, or at least symbolic, expression, and its most common feature is the predominance of motifs. This is Karlgren’s Style A for archaic China,25 painting and low relief for the Northwest Coast and New Zealand, and facial painting for the Guaicuru. But another style exists, of a more strictly formal and decorative character, with geometric tendencies. It consists of Karlgren’s Style B, the rafter decoration of New Zealand, the woven or plaited designs of New Zealand and the Northwest Coast, and, for the Guaicuru, a style easily identifiable, ordinarily found in decorated pottery, corporal paintings (different from facial paintings), and painted leatherwork. How can we explain this dualism, and especially its recurrence? The first style is decorative only in appearance; it does not have a plastic function in any of the four arts. On the contrary, its function is social, magical, and religious. The decoration is the graphic or plastic projection of a reality of another order, in the same way that split representation results from the projection of a three-dimensional mask onto a two-dimensional surface (or onto a three-dimensional one which nevertheless does not conform to the human archetype) and in the same way that, finally, the biological individual himself is also projected onto the social scene by his dress. There is thus room for the birth and development of a true decorative art, although one would actually expect its contamination by the symbolism which permeates all social life.
Another characteristic, shared at least by New Zealand and the Northwest Coast, appears in the treatment of tree trunks, which are carved in the form of superimposed figures, each of which occupies a whole section of the trunk. The last vestiges of Caduveo carving are so sparse that we can hardly formulate hypotheses about the archaic manifestations of it; and we are still poorly informed about the treatment of wood by Shang carvers, several examples of which came to light in the excavations at An-Yang.26
I would like to draw attention, nevertheless, to a bronze of the Loo collection reproduced by Hentze.27 It looks as though it could be the reduction of a carved pole, comparable to the slate reductions of totem poles in Alaska and British Columbia. In any case, the cylindrical section of the trunk plays the same role of archetype or “absolute limit” which we ascribed to the human face and body; but it plays this role only because the trunk is interpreted as a living being, a kind of “speaking pole.” Here again, the plastic and stylistic expression serves only as a concrete embodiment of impersonations.
However, our analysis would be inadequate if it permitted us only to define split representation as a trait common to mask cultures. From a purely formal point of view there has never been any hesitation in considering the t’ao t’ieh of archaic Chinese bronzes as a mask. On his part, Boas interpreted the split representation of the shark in Northwest Coast art as a consequence of the fact that the characteristic symbols of this animal are better perceived in a front view28 (see Plate III). But we have gone further: We discovered in the splitting technique, not only the graphic representation of the mask, but the functional expression of a specific type of civilization. Not all mask cultures employ split representation. We do not find it (at least in as developed a form) in the art of the Pueblo of the American Southwest nor in that of New Guinea.29 In both these cultures, however, masks play a considerable role. Masks also represent ancestors, and by wearing the mask the actor incarnates the ancestor. What, therefore, is the difference? The difference is that, in contrast to the civilizations we have been considering here, there is no chain of privileges, emblems, and degrees of prestige which, by means of masks, validate social hierarchy through the primacy of genealogies. The supernatural does not have as its chief function the creation of castes and classes. The world of masks constitutes a pantheon rather than an ancestrality. Thus, the actor incarnates the god only on the intermittent occasions of feasts and ceremonies. He does not acquire from the god, by a continuous process of creation at each moment of social Me, his titles, his rank, his position in the status hierarchy. The parallelism which we established is thus confirmed, rather than invalidated, by these examples. The mutual independence of the plastic and graphic components corresponds to the more flexible interplay between the social and supernatural orders, in the same way that split representation expresses the strict conformity of the actor to his role and of social rank to myths, ritual, and pedigrees. This conformity is so rigorous that, in order for the individual to be dissociated from his social role, he must be torn asunder.
Even if we knew nothing about archaic Chinese society, an inspection of its art would be sufficient to enable us to recognize prestige struggles, rivalry between hierarchies, and competition between social and economic privileges—showing through the function of masks and the veneration of lineages. Fortunately, however, there are additional data at our disposal. Analyzing the psychological background of bronze art, Perceval Yetts writes: “The impulse seems almost invariably to have been self-glorification, even when show is made of solacing ancestors or of enhancing the family prestige.” 30 And elsewhere he remarks: “There is the familiar history of certain ting being treasured as emblems of sovereignty down to the end of the feudal period in the third century b.c.” 31 In the An-Yang tombs, bronzes were found which commemorate successive members of the same lineage.32 And the differences in quality between the specimens excavated can be explained, according to Creel, in terms of the fact that “the exquisite and the crude were produced side by side at Anyang, for people of various economic status or prestige.” 33 Comparative anthropological analysis, therefore, is in agreement with the conclusions of Sinologists. It also confirms the theories of Karlgren, who, unlike Leroi-Gourhan34 and others, states, on the basis of a statistical and chronological study of themes, that the representational mask existed before the mask’s dissolution into decorative elements and therefore could not have grown out of the experimentation of the artist who discovers resemblances in the fortuitous arrangement of abstract themes.35 In another work Karlgren showed how the animal decorations of archaic objects became transformed in the later bronzes into flamboyant arabesques, and he related phenomena of stylistic evolution to the collapse of feudal society.36 We are tempted to perceive in the arabesques of Guaicuru art, which are so strongly suggestive of birds and flames, the final stage of a parallel transformation. The baroque and affected quality of the style would thus represent the formal survival of a decadent or terminated social order. It constitutes, on the esthetic level, its dying echo.
The conclusions of our work do not preclude in any respect the always-possible discovery of hitherto unsuspected historical connections.37 We are still faced with the question of finding out whether these hierarchical societies based on prestige appeared independently in different parts of the world, or whether some of them do not share a common cradle. With Creel,381 think that the similarities between the art of archaic China and that of the Northwest Coast, perhaps even with the arts of other American areas, are too marked for us not to keep this possibility in mind. But even if there were ground for invoking diffusion, it would not be a diffusion of details—that is, independent traits traveling each on its own and disconnected freely from any one culture in order to be linked to another—but a diffusion of organic wholes wherein style, esthetic conventions, social organization, and religion are structurally related. Drawing a particularly striking analogy between archaic Chinese and Northwest Coast art, Creel writes: “The many isolated eyes used by the Northwest Coast designers recall most forcibly their similar use in Shang art and cause me to wonder if there was some magical reason for this which was possessed by both peoples.” 39 Perhaps; but magical connections, like optical illusions, exist only in men’s minds, and we must resort to scientific investigation to explain their causes.
NOTES
1. "The Art of the North West Coast," Gazette des Beaux-Arts (1943).
2. Carl Hentze, Objets rituels, Croyances et Dieux de la Chine antique et de l’Amérique (Antwerp: 1936).
3. Leonhard Adam, "Das Problem der asiatisch-altamerikanischen Kulturbeziehungen mit besonderer Berücksichtigung der Kunst," Wiener Beiträdge zur Kunst und Kulturgeschichte Asiens, V (1931); "Northwest American Indian Art and Its Early Chinese Parallels," Man XXXVI, No. 3 (1936).
4. See, for example, F. D. McCarthy, Australian Aboriginal Decorative Art (Sydney: 1938), Fig. 21, p. 38.
5. Review of Carl Hentze, Frühchinesische bronzen und Kultdarstellungen (Antwerp: 1937), in Man, XXXIX, No. 60 (1939).
6. For China and New Zealand, see R. Heine-Geldern in Zeitschrift für Rassenkunde, Vol. 2 (Stuttgart: 1935).
7. Henry Field and Eugene Prostov, "Results of Soviet Investigation in Siberia, 1940-1941," American Anthropologist, XLIV (1942), p. 396.
8. In his book, Medieval American Art (New York: 1943), Pal Kelemen regards the resemblances between American art and some of the arts of the highest civilizations of the Eastern hemisphere as only "optical illusions" (Vol. I, p. 377). He justifies this opinion by writing that "Pre-Columbian art was created and developed by a mentality totally alien to ours" (p. 378). I doubt that in the whole work of the diffusionist school one could find a single statement so completely unwarranted, superficial, and meaningless.
9. Franz Boas, Primitive Art, Instituttet for Sammenlignende Kulturforskning, series B, Vol. VIII (Oslo: 1927), pp. 223-24.
10. Ibid., pp. 224-25.
11. H. G. Creel, "On the Origins of the Manufacture and Decoration of Bronze in the Shang Period," Monumenta Serica, Vol. I (1935), p. 64.
12. Loc. cit.
13. "Indian Cosmetics," VVV, no. 1 (New York: 1942). Tristes Tropiques (Paris: 1955).
14. Franz Boas, op. cit., p. 238.
15. Ibid., p. 239 and Fig. 247.
16. See, for example, the Tlingit tattooings reproduced by J. R. Swanton in Bureau of American Ethnology, 26th Annual Report, Plates XLVIII to LVI; and Franz Boas, op, cit., pp. 250-1 (body paintings).
17. I have developed this analysis further in Tristes Tropiques, chapter XX (Paris: 1955).
18. "The Art of the North West Coast," op. cit.
19. M. Dobrizhoffer, An Account of the Abipones, trans, from the Latin, Vol. II (London: 1822), p. 20.
20. Ibid., p. 21.
21. See also H. G. Creel: "The fine Shang pieces are executed with a care, extending to the most minute detail, which is truly religious. And we know, through the study of the oracle bone inscriptions, that almost all the motifs found on Shang bronzes can be linked with the life and religion of the Shang people. They had meaning and the production of the bronzes was probably in some degree a sacred task." "Notes on Shang Bronzes in the Burlington House Exhibition," Revue des Arts asiatiques, X (1936), p. 21.
22. Boas, op. cit., pp. 222-24.
23. John R. Swanton, Tlingit Myths and Texts, Bureau of American Ethnology, Bulletin 59 (1909), pp. 254-255; E. A. Rout, Maori Symbolism (London: 1926), p .280.
24. Florance Waterbury, Early Chinese Symbols and Literature: Vestiges and Speculations (New York: 1942).
25. Bernhard Karlgren, "New Studies on Chinese Bronzes," The Museum of Far Eastern Antiquities, Bulletin 9 (Stockholm: 1937).
26. H. G. Creel, Monumenta Serica, Vol. I (1935), p. 40.
27. Carl Hentze, Frühchinesische bronzen und Kultdarstellungen (Antwerp: 1937), Table 5.
28. Boas, op. cit., p. 229. One should distinguish, however, between two forms of split representation—namely, split representation proper, where a face and sometimes a whole individual are represented by two joined profiles, and split representation as shown in Plate III, where one face is shown with two bodies. We cannot be certain that the two types derive from the same principle, and in the passage which we summarized at the beginning of this chapter, Leonhard Adam wisely distinguishes between them. The split representation so well illustrated in Plate III reminds us, indeed, of a similar technique well known in European and Oriental archaeology This is the beast with two bodies, whose history E. Pottier attempted to reconstruct ("Histoire d’une bête," in Recueil E. Pottier, Bibliothèque des Ecoles d’Athènes et de Rome, section 142). Pottier traces the beast with two bodies to the Chaldean representation of an animal whose head appears in a front view and the body in profile. A second body, also seen in profile, is assumed to have been subsequently attached to the head. If this hypothesis is correct, the representation of the shark analyzed by Boas should be considered either as an independent invention or as the easternmost evidence of the diffusion of an Asiatic theme. This last interpretation would be based on evidence which is far from negligible, namely the recurrence of another theme, the "whirl of animals" (see Anna Roes, "Tierwirbel," Ipek [1936-37]) in the art of the Eurasian Steppes and in that of certain areas of America (especially in Moundville). It is also possible that the beast with two bodies derives independently, in Asia and America, from a technique of split representation which has not survived in the archaeological sites of the Near East, but which left traces in China and may still be observed in certain areas of the Pacific and in America.
29. The art of Melanesia presents rudimentary forms of split representation and dislocation. See, for example, the wooden containers of the Admiralty Islands reproduced by Gladys A. Reichard, "Melanesian Design: A Study of Style in Wood and Tortoise Shell Carving," Columbia University Contributions to Anthropology, II, no. 18 (1933), and the following comment by the same author: "Among the Tami, joints are represented by an eye motif. In the face of the fact that tattooing is exceedingly important to the Maori and that it is represented on the carvings, it seems to me more than possible that the spiral often used on the human figures may emphasize the joints" (p. 151).
30. W. Perceval Yetts, The Cull Chinese Bronzes, London, 1939, p. 75.
31. W. Perceval Yetts, The George Eumorphopoulos Collection Catalogue, Vol. I (London: 1929), p. 43.
32. W. Perceval Yetts, "An-Yang: A Retrospect," China Society Occasional Papers, n.s., No. 2 (1942).
33. H. C. Creel, op. cit., p. 46.
34. A. Leroi-Gourhan, "L’Art animalier dans les Bronzes chinois," Revue des Arts asiatiques(Paris: 1935).
35. B. Karlgren, op. cit., pp. 76-78.
36. B. Karlgren, "Huai and Han," The Museum of Far Eastern Antiquities, Bulletin 13 (Stockholm: 1941).
37. The problem of ancient relations across the Pacific Ocean has recently come to the fore again, owing to the surprising discovery, in a provincial museum of southeastern Formosa, of a low-relief in wood which could be of local origin. It represents three persons standing. Those located at the extremities are in the purest Maori style, while the person in the middle offers a kind of transition between Maori art and that of the Northwest Coast. See Ling Shun Sheng, "Human Figures with Protruding Tongue," Bulletin of the Institute of Ethnology, Academia Sinica, No. 2 (September 1956), Nankang, Taipei, Taiwan.
38. H. C. Creel, op. cit., pp. 65-66.
39. Ibid., p. 65.