objects whose animal form did not derive from social meaning. Although a large proportion of Northwest Coast animal images depict crests associated with family histories, Boas noted that some animal representations on hunting implements and food bowls related to certain natural attributes of the animals themselves. Granting that totemism was a significant incentive in the development of Northwest Coast art, he suggests that once the use of conventionalized animal imagery to decorate objects had been established, artists began applying similar designs to objects unrelated to totemism. Thus a halibut club assumed the shape of the sea lion or killer whale because these are successful fishers, while the grease dish represented a blubber-rich seal. Disputing restrictive and limiting unicausal theories, Boas asserted that his analysis of animal imagery in Northwest Coast art was “one of the numerous ethnological phenomena which, although apparently simple, cannot be explained psychologically from a single cause but are due to several factors” (1896:102–3).23
In 1897, Boas wrote his most elaborate and important work on art thus far, “The Decorative Art of the Indians of the North Pacific Coast,” published in the Bulletin of the American Museum of Natural History.24 This was to become his first systematic argument against an evolutionary explanation of the development of conventionalized imagery. In his introductory comments, Boas seemed to accept as true in some instances the realist/degenerationist theory of Frederic Ward Putnam:25
It has been shown that the motives of the decorative art of many peoples developed largely from representations of animals. In course of time, forms that were originally realistic became more and more sketchy, and more and more distorted. Details, even large portions, of the subject so represented, were omitted, until finally the design attained a purely geometric character (1897a: 123).
Boas then pointed out that this did not occur in one region, the Northwest Coast:
The decorative art of the Indians of the North Pacific Coast agrees with this oft-observed phenomenon in that its subjects are almost exclusively animals. It differs from other arts in that the process of conventionalizing has not led to the development of geometric designs, but that the parts of the animal body may still be recognized as such (1897a: 123).
The Northwest Coast artist adhered to an iconographic canon that determined certain identifying characteristics of animals, such as the beaver’s large incisors and crosshatched tail, the killer whale’s large dorsal fin, and the eagle’s large, downward-curving beak.26 According to Boas, the artistic requirement that any animal image had to include all identifying elements of the animal led to highly conventionalized depictions when that animal decorated certain surfaces. On a three-dimensional sculpture, the artist could represent his subject, with all its characteristic features, in a fully naturalistic fashion (and the Northwest Coast artist was capable of very naturalistic representations); this became more problematic when he was presented with a two-dimensional surface. Sometimes that artist needed to abstract and distort the animal-subject in order to make it fit on the surface being decorated. Often, when a three-dimensional animal was depicted on a two-dimensional surface, the artist resorted to what is termed “split representation,” in which the animal’s body is split down the middle, flattened out, and shown in both profiles connected at the center. Sometimes, it was not possible for the artist to portray every feature of the entire animal, so he resorted to representing its characteristic motifs, in which case the rendering became truly symbolic. Here on the Northwest Coast, therefore, the relative realism or abstraction of an image depended on the shape of the surface upon which the artist depicted different animals and their culturally dictated identifying characteristics.
Boas once again discussed this relationship between animal imagery and the shape of the surface on which it appeared in his “Facial Paintings of the Indians of Northern British Columbia,” published in 1898 as part of volume 1 of the Jesup Expedition Publications.27 Repeating the introductory comments of his 1897 monograph, Boas began his text by accepting the realist/degenerationist theory that among “most primitive people we find a tendency to the development of geometric designs.” He then asserted that this did not occur in Northwest Coast art (1898b: 13). To test his theory that in this region the form of the object being decorated with all the necessary animal symbols influenced the relative naturalism of the representation, Boas analyzed the painting applied to the most complex possible surface, the human face. After describing a wide range of Northwest Coast facial paintings, Boas concluded that geometric designs in these paintings did not necessarily evolve from naturalistic ones but were artistic responses to the problems posed by the shape of the face.
To investigate this question, Boas collected face paintings from Charles Edenshaw, “one of the most famous artists” of the Haida, and arranged them in a sequence from the most realistic to the most abstract. As it turned out, the fullest and most realistic representations appeared on the faces of the highest-ranking people, while those of lower rank had more conventionalized face paintings. In some cases, the facial features became part of the design, while in others, they were ignored, with the face serving as a flat surface. While sometimes the depicted animal was readily identified, it could be so abstract that identification was virtually impossible without an explanation by the informant. In this study, Boas noted the appearance of what he claims to be unique in the Northwest Coast: animal symbols in the form of pure geometric designs. Some of these abstract motifs, moreover, could represent different animals, requiring identification by the owner of the image. Boas would periodically return to the point that at times iconographic identification could be obscure; this supported his recurrent theme that uncomplicated answers to questions of meaning simply did not exist.
In his concluding comments, Boas (1898b:24) asserts that “the collection is of theoretical interest mainly because it shows that the difficulty of adapting the subject of decoration to the decorative field has been a most powerful element in substituting geometrical forms for less conventional designs, and in showing a series of important transitional forms.” It should be pointed out that this case is less convincing than that presented the year before in “The Decorative Art of the Indians of the North Pacific Coast.” The variety of imagery on these facial paintings suggests less that the artist was trying to solve formal problems (as he seems to have been doing in the split representation) and more that he could choose from a range of styles depending upon factors that at least in some cases had to do with the rank of the individual being painted (as he himself noted early in the essay). Although he would repeat the idea that a primitive artist at times freely chose from a range of artistic possibilities when socially dictated conventions did not completely restrict his choice, Boas did not return in later publications to this topic of facial paintings.
In addition to questioning the relationship between naturalistic and stylized images, Boas studied the relationship between art style and linguistic families. In British Columbia, the Coast Salish have a distinctly different art style from their linguistic relatives, the Salish-speaking Thompson people of the interior of the province. During the Jesup North Pacific Expedition, Boas had a British Columbia resident, James Teit, collect examples of Thompson Indian art. The pieces Teit collected resembled Plains Indian art far more than they did Northwest Coast art. In his contribution to Teit’s monograph in the Jesup series on the Thompson Indians, Boas (1900b) compared Coast and Interior Salish art, pointing to the absence of plastic art among the latter, so different from the highly three-dimensional art of the Northwest Coast. Unlike the Northwest Coast artists, Thompson painters decorated their implements with designs not prompted by the shape of the surface. The iconography of the two types of artworks created by linguistically related peoples differed as well. Unlike the universally understood imagery of most Northwest Coast representations (although, of course, not all, as he demonstrated in the facial painting article), Thompson decorative designs, often abstract and ambiguous, could be interpreted differently and sometimes apparently arbitrarily by different people. Designs on implements, which related to their use, differed from those on ceremonial pieces, which depicted owners’ dreams. These comparisons implicitly demonstrated the inadequacy of any theory positing a direct connection between language and art style. In his “Conclusion” to the Thompson monograph, Boas attempted to reconstruct the history
