The American Indian/Chapter 5

by Clark Wissler

CHAPTER V

DECORATIVE DESIGNS

In the preceding discussions we have ignored the most interesting and suggestive sides of textiles and ceramics, namely, their decorations. Wherever such products occur we almost always find them richly ornamented by designs in color which constitute the greater part of the decorative art of their makers. Taking the New World decorative designs as a whole, we are impressed with their extreme geometric unrepresentative character and the rarity of realistic art. A stroll through a large museum reveals an astonishing complexity of geometrical design in contrast to similar collections from the Old World. Nowhere else do we find anything in basketry approaching the finest basketry decorations of the Pacific Coast or in pottery that of the Andean region. From the standpoint of æsthetic values, the ancient Old World products may be rated as superior, but the range and richness of geometric design in the New World cannot be denied.

Anthropologists have given the subject of decorative design a great deal of attention, and we consequently have for the tribes of the northern continent a body of special research literature not equalled by that for any other part of the world. Quite recently, the use of ceramic design as an index to chronology and relationship in extinct cultures has appeared as a special method in archæological research and promises a considerable development in the near future.^[1] Unfortunately, no such progress has been made in the art of the southern continent or even for the Antilles and Mexico. Our first task, therefore, is to consider rather fully the status of the North American design problem and then to view the southern continent from that horizon.

TEXTILE DESIGNS

If we compare the decorations upon a representative series of baskets from the Rocky Mountain region with those upon a series of pottery vessels from Arizona and New Mexico, there seems to be a definite similarity. Closer inspection suggests that this is true because certain combinations of angles and checker patterns are common to both. The chief point of difference is that curved lines and realistic figures are rare in basketry, whereas they occur with somewhat greater frequency on the pottery in question. Again, if we examine the blankets of the Navajo, we find a series of designs strikingly like those upon the basket series. Since we know that the Navajo weaving is of recent origin, we infer that many of their blanket designs were borrowed from basketry and because of the much greater distribution of the latter, that the pottery designs were also greatly influenced thereby.

An important point has been made that the technique of weaving itself places certain form limitations upon designs which tend to make them similar, irrespective of the wishes of the artist.^[2] In all weaving we have a geometrical relation between the warp and weft elements since they have a right-angle relation to each other and, in the main, can build up a design by equal rectangular units only. In basketry these units are usually so large that diagonals can only be run as steps and even in cloth it is difficult to escape this effect. These stepped designs and diagonal rows of small squares constitute one of the prevailing characteristics of textile art, so that in our discussions of design distribution we must make full allowance for similarities due to the limitations imposed by the weaving technique.

For example, we find a certain type of designs for cane baskets in Louisiana, and passing over to northern South America,^[3] we find baskets of similar materials with designs almost identical. In this case we have other facts that suggest this similarity to be but another example of culture diffusion. Yet, we can find baskets in some of the Pacific Islands which can scarcely be distinguished from cane baskets of the New World, if we consider the designs only, and in this case there is no good reason for expecting diffusion.

The limitations set by weaving are more clearly shown when realistic figures are attempted (Fig. 35). Painted pottery, on the other hand, imposes no such restrictions in the matter of design, but leaves the hand free to make curves of any form. Accordingly, when we find the aboriginal potters of Arizona

Fig. 35. Types of Textile Design. a. Checker; b. Twill; c. A Typical Pattern from Cane Basketry; d. Design from the Penn Treaty Wampum Belt; e. Bird Figure from Peruvian Cloth; f. Design on a Peruvian Basket. Holmes, 1888. I

and New Mexico using a great array of checked and angular patterns, with stepped lines, we must necessarily refer them to textiles.

Fig. 36. Types of Prehistoric Peruvian Textile Designs

Another significant point is that the extensive use of realistic figures in cloth occurs only where weaving is highly developed, as in Peru and Mexico. When we examine examples of such decorations as are preserved in our museum collections, we note that even so, these figures are greatly distorted to make their contours coincide with the fixed lines of weaving. Further, it is also in these same localities that pottery decorations become more realistic, suggesting that some allowance must be made for the degrees of complexity in the culture of the weavers. It may be that the simple designs upon New England pottery are about all that can be expected from such a crude cultural setting. Yet, we must conclude that in the earlier stages of their historical developments in the New World both textiles and pottery were decorated with geometric designs and that the use of realistic figures came later. This is somewhat at variance with a current theory of art genesis which considers geometric art to be mere conventionalizations of earlier realistic figures. We have already noted how the weaving technique itself conventionalized all figures and have recognized other factors producing conventional effects, but the cultural conditions in the New World do not seem consistent with the above theory of design origin. The total distribution of the several types of design points clearly to a development from the simplest geometrical textile designs to the realistic textile figures.

DISTRIBUTION OF DESIGNS

Like many other culture traits, designs tend to fall into geographical groups. While the boundaries to such areas cannot always be drawn with great precision, their centers can be located without much difficulty. We have noted that California seemed to be the center of the highest attainments in basket-making, and it so happens that this is also the great center for basketry designs. As indicated on the map (Fig. 17) the basketry area includes the great plateau region extending from well up into British Columbia southward to the non-Pueblo tribes of Arizona and New Mexico. Here we saw that two kinds of technique were in use, coil and woven basketry, usually twined weaving, and that while tribes tend to specialize now in one of these and then in another, this variation seems to have little effect upon the designs, for the same

Fig. 37. Decorative Design Areas in North America: 1. North Pacific Coast Center; 2. California Center; 3. Southwest, or Pueblo Center; 4. Plains, or Bison Center; 5. Eastern Center—subtypes a and b; 6. Southeastern, or Gulf Center

designs occur upon both. The coil technique offers great freedom in design because of its similarity to embroidery; but this is, perhaps, compensated by the trick of overlaying

Types of North American Basketry

Fig. 38. A Pomo Basket. California

twined strands with thin strips of colored materials to produce the designs. In the inland salmon area, coil baskets are decorated by imbrication, which is also an ingenious overlay, and for that reason was most likely derived from twine overlay. The basketry of the Tlingit gives a different type of decoration, chiefly in the use of bands of overlay, but these are a secondary part of the art of the North Pacific area to which we shall return later.

Another important art center is to be found among the Pueblo peoples of the Southwest where we see an exuberant development of pottery designs and blanket decorations.^[4] Archæological collections^[5] show that, if anything, there has been a deterioration in pottery decoration during the historic period but, on the other hand, there seems to have been a marked development in blanket designs. We are clear that Navajo textiles have passed through a development of this kind, for the old specimens are almost entirely striped.^[6] While Navajo weaving is supposed to be of recent origin, it is obvious that the designs were not copied from European techniques, but from aboriginal American models. Further, we have some textile remains from cliff ruins in which striped decorations are the rule and the same tendency is shown in Hopi and Zuñi weaving. A few exceptional specimens have come to notice that bear designs of another character, particularly those from the Gila River;^[7] but these are toward the south and may therefore be intrusive.

However, our most important problem in this area is to be found in pottery decoration. If we consider modern Pueblo pottery only, we find that its designs are largely geometric in appearance, although a strong realistic tendency is also plainly evident. Even many of the highly conventionalized geometric forms prove to be symbols of mountains, clouds, thunder, rain, etc., while among them appear unmistakable drawings of plants and animals. Yet, taking modern pottery as a whole, the geometrical character of the designs seems to predominate. In the discussions of southwestern chronology, we shall see that the more widely diffused and, older type of pottery is decidedly geometric in character. Thus, two of

Fig. 39. Beaded and Painted Designs of the Plains Indians. Kroeber, 1902. I

the favorite design concepts are the simple checker textile pattern and the step or "terrace." Again, if we look almost anywhere in the Pueblo area we shall find these patterns occurring. They have so sure a place in textile art and lend themselves so much less readily to freehand work that a non-pottery origin is suggested. Quite recently a localized tendency toward realistic pottery painting has come to notice in southwestern New Mexico, but even here we also meet with the familiar geometric designs.^[8]

Adjoining the southwest and east of the great basketry area is the bison area, which is weak in basketry and cloth, but still has a highly developed embroidery of beads and quills in which the designs are geometric and manifest many of the characteristics of textile designs.^[9] In fact, the way in which beads and quills are handled in this area requires that designs be built up by accretions of small rectangular surface contours, which is just what we have in weaving. If our general principle of technique limitations holds, we should expect to find geometric forms prevailing. This is exactly what we do find (Fig. 39).

Even among the basket-making Apache of the Southwest, we find objects of skin decorated with designs upon covered surfaces of beads. This is clearly an intrusion from the bison area because it is only now and then that we find identity between the designs on Apache baskets and objects of skin, each having a style of its own. On the other hand, these beaded designs are quite like those found far out into the buffalo country. These buffalo hunters did not decorate pottery, in fact, some did not even make it, but they did paint rawhide objects and, strange to say, even this freehand work was in geometric designs not at all unlike those in beads and quills. While the reason for this is not entirely clear, we note that all the beadwork is by women, who also paint the geometric designs, whereas the men who paint upon robes, tents, etc., use realistic figures. This suggests that the difference may be merely a matter of social convention.

An important problem is the origin of Plains art as a whole. Though we have shown that bead technique imposes textile limits upon the decorations, the fact that the Plains area is in direct contact with basketry-making peoples and the weavers of the Southwest reveals the possibility of diffusion. While this, like most other problems, is one for the future, there are several good reasons for believing that the art of the bison, or Plains area, is in the main an independent development. In the first place, its center is in the very heart of the area, while it is weakest on the margins. In its great western Shoshoni fringe we find a condition not unlike that of the Apache in that beadwork and basketry exist side by side but with different design systems.

There is, however, a more direct approach to the problem by the analytic comparison of designs. Kroeber^[10] has carefully analyzed the designs of California baskets and Plains beadwork in search of the prevalent design units. When these are found, they prove to be, in the main, very simple geometric forms and though many can be very closely matched for the two areas, their very simplicity, taken with the principle of textile limitation, lessens the probability of their common origin. On the other hand, if we take more complex design wholes we find very little correspondence between the two areas, for each has a number of highly unique designs not found in any other part of the world. Hence, even this method tends to assert independence of origin.

To the north and east of the Plains area we have another art area in which neither ceramics nor true textiles play an important part. This region comprises the greater part of the caribou area and the northern half of the eastern maize area, a region in which, although the decorations are again by beads and quills, there is yet a distinct type of design. Here we have exactly the opposite of the preceding, for instead of textile-like designs we find curved figures and more or less realistic, plant-like forms. The cause for this very extraordinary contrast is an important problem.

When we try to locate the geographic center of this art, it proves somewhat elusive, but closer inspection reveals two sub-centers, one in eastern Canada, the other near Lake Superior. The eastern sub-type has been brought to notice

Photograph of Blackfoot Women by Fred. R. Meyer

Fig. 40. The Decorative Art of the Plains Indians

by Speck ^[11] under the designation double-curve art (Fig. 42). These curved designs, while obviously resembling vines and leaves, are still somewhat less realistic than beadwork designs of the western sub-type, perhaps because of their greater conventionalization. Though these curved designs do occur in beads and quill, they are more frequently found incised in birchbark or painted on skins. The best-known examples of the latter, are the Naskapi coats in museum collections. These observations suggest that this eastern sub-center was originally dominated by freehand work upon skins and bark

Fig. 41. Design Elements Used in Plains Indian Beadwork. Kroeber, 1908. II

from which the somewhat similar beaded designs were most likely copied. No cloth is made here and practically no woven decorated basketry, but we find some woven wampum belts and some bands of quill and moose-hair interwoven with bast fiber, in which the designs tend to be geometric. So far, we do not recall a single example of the double-curve art in these truly woven objects. All this suggests that we have here at the eastern sub-center a type of design which developed from freehand drawing upon skins and birchbark.

When we turn to the western sub-type we find practically no painting upon skins or decorative drawing upon bark and the beaded decorations correspondingly more numerous and decidedly floral. From this sub-center come those admirably beaded flowers seen in our museum collections. Speck^[12] has included the more conventionalized examples of these under his double-curve motive, but all they have in common seems to arise from their universal floral foundations. While there is little in the way of skin painting accompanying the western sub-type of beading, there is a far greater textile

Fig. 42. Decorations on Birchbark from Eastern North America. The upper figures represent the side and bottom of a bark vessel from the Penobscot Indians. The lower sketch is a typical "double curve" design. (Center 5b, Fig. 37.) Speck, 1914. I

development, especially at the sub-center. The weaving of bands containing quills or moose-hair is a feature in the Déné region. From the Ojibway of Wisconsin southward, we find a rather high development of bag weaving and many forms of mats. The designs upon quill bands and mats are almost without exception geometric, while bags show two forms—purely textile geometric and realistic animal figures. Drawing and sketching upon birchbark was developed almost to the point of writing and in that sense was not decorative. It was, however, entirely pictographic. Wooden objects were not infrequently adorned with incised curved designs somewhat like the floral effects in beads. The chief differences, therefore, between the two sub-centers are the disturbing textile developments in the west, with a decided realistic tendency in beadwork, while in the east freehand double-curve floral figures prevailed. The similarities are in the more fundamental character—a predominatingly freehand floral decoration. In the preceding areas we found textiles or embroidery techniques in the majority, and at the same time the decorative art was geometric. But here in the north and east we find textiles extremely weak; yet, when they do appear they tend to geometric forms. Beadwork, however, more often followed the freehand motives than not. So we see here the suggestion of a chronological relation in that this particular beaded art was derived from bark and skin decorations.

The extreme floral character of some of this beadwork has led many to regard the whole as a post-Columbian development. The very wide distribution of the Cree and Montagnais, together with their very early intimate association with French colonists, presents a favorable condition to rapid diffusion. Yet, the very characteristic double-curve art on bark and painted skins cannot be attributed to Europeans. All that can reasonably be conceded is that their trade stimulated the use of beads, and their decorative preferences tended to emphasize the old floral character. On the other hand, there seems not the least reason to doubt that the very striking beaded flowers of the west are due to European influence.

Strange to say, all the regions we have so far considered are almost completely innocent of carving or modeling in the round, everything being flat. But we now turn to the North Pacific Coast and Eskimo areas where carving is the leading art. Faint traces of carving appear at the northern border of California and grow stronger as we ascend the coast until we reach the Haida of the Queen Charlotte Islands which seems to be its geographical center. In the central part of the area we find the great totem poles and colossal grave figures, besides an endless array of smaller objects, all in wood. As we proceed still farther northward, wood tends to disappear and ivory to take its place. Like most other traits, ivory carving begins to be frequent with the Tlingit and gradually grows in intensity as wood disappears, the difference appearing merely as a matter of environment. Then as we go around the north coast of Alaska and eastward along the extended habitat of the Eskimo, carving almost dies out. This peculiar distribution among the Eskimo suggests an indirect historical relation with the carving center.

The intense development of carving at this center has a noticeable effect upon decoration. Boas^[13] has shown how the very curious relief carving upon the outsides of wooden vessels results from an attempt to carry around the contour of animals or men in such a way that the whole may stand for a realistic model. Naturally, when flat surfaces are treated the whole figure is spread out upon it. Sometimes these designs are merely laid out in color and though no doubt more conventionalized thereby, they are still the undeniable offspring of carving. All this is a feature of the central group of tribes where the art is most intense and where it is, in part, at least, the expression of a very complex system of beliefs concerning family ancestors. North from the Tlingit and south from the Nootka of Vancouver Island we have many vessels carved in the life-like forms of animals, but practically no trace of the relief ornamentation just mentioned, a fact which strongly suggests that this feature is purely a development of the more intense art at the center and that it is, therefore, relatively recent.

We have noted that there is also a textile center in this region, but we now see that it does not coincide with that for carving, its location being inland among the Salish peoples. Emmons^[14] has made a good case for the relatively recent introduction of the Chilkat blanket to that tribe. In this famous textile we usually find the queer spread-out animal

Fig. 43. Art of the North Pacific Coast Indians as Represented in the American Museum of Natural History

forms noted above and not the usual type of textile designs. It has been clearly shown that this decoration was directly copied from house fronts.^[15] Now, if the blanket came from the Southeast it must have arisen in a place and at a time too remote to have incorporated this decoration at the start. In fact, there is evidence of several sorts to show that these textiles were originally decorated with bands of small geometric figures. The basketry of the Haida and Tlingit^[16] shows a similar banded style, and this, in turn, has a curious resemblance to the quill woven bands of the Déné people of the adjacent caribou area. The significance of the latter is not clear.

Thus, we find in this art area a good example of conflict between a carving center and a textile one, the Chilkat blanket being about the only compromise. The Eskimo of Alaska took up basketry but not its design decoration.

Next we turn to a more difficult problem; namely, the art of the southern half of the eastern maize area. The data available are so much less adequate than for the preceding that one must hesitate to even enter such a discussion. For though, as stated elsewhere, we have historical records vouching for a higher textile development in the Gulf States than in the North, no specimens have come down to us. There are reasons for suspecting that the bag weaving we have noted for the upper section of the Mississippi Valley is in a large measure the fringe of this area, but without some corroborative data we are scarcely justified in formulating it as an assumption. Basketry has survived in Louisiana,^[17] where we find cane weaving in designs of black and red. As previously stated, the material and technique restrict designs to just such as we find here, and from this it may be inferred that they truly represent the former basketry art of the whole southeastern area. What may be the relation of the pottery found here to the historic tribes is also a puzzle. If this pottery was extant at the period of discovery, then one of the most distinctive design concepts was the spiral scroll.

When we turn to the art of the intense culture area, our problem becomes very largely one of archæology and the yet undetermined sequences of culture, because the thoroughness of the Spanish conquest practically obliterated the native culture. No doubt careful research would still reveal many surviving traits in the present populations of these countries but such studies have not been sufficiently numerous to assist us. As suggestions we may cite Tozzer's^[18] study of the Lacandones as a surviving Maya people.

With numerous dense groups of people, as in ancient Mexico and Peru, where a political organization gradually overflowed and submerged the successive local groups, there must have been a great variety of art types that persisted in the homely affairs of life; but the succeednig centuries of European trade

Fig. 44. Prehistoric Sketches of Textile Designs from the Maya.
Spinden, 1913. I

seem to have swept them into oblivion. For Mexico and southward we have no clear idea of the aboriginal textile development. Among the present Huichol^[19] we find considerable weaving in which the designs have a marked realistic tendency. So far as known, this is a trait of the modern textile art for the whole stretch of country from the Rio Grande to Panama. While it is certain that we have here a result due in part to contact with Spanish culture, there is no reason for assuming that a new textile art was created since the conquest. The general similarity to Peru, in the range and direction of conventionalization, is sufficient warrant for assuming an original textile art of a similar level. We may, however, get some idea of Maya textile design from the known sculptures and codices.^[20] According to early Spanish authorities, the Maya peoples were the most expert weavers in New Spain, which statement, if true, enables us to gauge the whole state of the art from the illustrations the native artists have left us. From these the specific resemblances to modern native Mexican weaving are clear^[21] and also the general resemblance to Peruvian styles previously noted (Fig. 45).

When we turn to the pottery of this region, even less survives among the living peoples so that any study of the ceramic art also becomes essentially archæological. However, the

Fig. 45. Mexican Textile Designs. The costumed figure is from a Prehistoric Maya Drawing (Spinden, 1913. I.), but the other sketches represent the work of modern Huichol Indians (Lumholtz, 1904. I.), in Northern Mexico

ceramic collections in our museums are not extensively embellished with painted designs. In contrast to the pottery of the Pueblo area they are plain, modeling in relief and realistic painting here taking the place of mere design. In the Panama area, including Costa Rica and the Chiriqui district, we have a complex of alligator and armadillo designs, recalling again Peruvian and some modern Mexican textile decorations.^[22] In a few instances the Maya were particularly successful with

Fig. 46. A Series of Peruvian Designs. Mead, 1916. I

textile-like designs for mosaic and stucco work, but these, the well-known case of Mitla in Mexico,^[23] and the single example at Chanchan, Peru, are the striking exceptions to the rule that wherever architectural embellishment is undertaken, it tends toward realistic carving.

When we come to South America, we find that certain accidents have preserved us a good series of aboriginal textiles from Inca culture. In the technique of design these ancients were remarkably proficient, even to the extent of using complex color sequences.^[24] On the whole, their designs tend to be realistic figures: men, cats, birds and fish being distinguishable in many degrees of conventionalization. In fact, we find here the best illustrations of the geometric biases in loom weaving. Associated with this art is an equally superior development of pottery decoration. One prominent feature of this pottery is the introduction of life forms, so that we have jars representing persons, birds, monkeys, fishes, etc., in which the modeling is of a high order. The decorations are in both color and incised work. In color, we have the great triumph of Nasca and Titicaca ware, so far superior to anything yet discovered in the New World. The painted designs upon this pottery are comparable to those upon cloth in that they have the same realistic tendencies. Certain fixed conventional forms appear both on pottery and cloth, suggesting the fundamental unity of design concepts for both ceramics and textiles.

As we go out from Peru in either direction, pottery decorations become inferior; consequently, we may be sure that the center of the art was in that country. The great problem for the future is to discover the historical relations of this center to the adjacent cultures. If we follow around the north coast and down into Brazil we find greater use of painted pottery decorations than in the corresponding parts of North America. No doubt one factor in this distribution is the presence of the very strong Peruvian art center. In a similar way this Peruvian influence can be seen in Chile and the adjacent parts of Argentina, presumably again connecting with eastern Brazil.

As to the textile designs in these outlying regions, we are so ignorant that little can be said, though the explorations of German anthropologists^[25] among the wilder peoples of eastern Brazil give us a fair idea of designs in a few localities. As previously noted, we find here the designs peculiar to cane basketry in all parts of the world; however, some textile work exists in which simple striped designs occur, though on the whole the designs are similar to those upon basketry. Painted decorations upon bark and wood are also found which have a geometric character; but these are almost entirely made up of triangles.

In the northwest Amazon country there is an identity between pottery designs and those used in body painting.^[26] The colors are laid on in large masses, but in the form of true textile designs. A similar style of body painting has been reported for Panama.

This relatively brief survey of New World art reveals some interesting general characteristics. The experience of anthropologists shows that by generalizing design characteristics we can consistently differentiate a few centers of development and influence. These prove also to be centers of specialization in industrial art. For example, the tribes of California are lamentably deficient in everything but basketry. Again, we see that geometric art and realistic decoration tend to be antagonistic to each other in the sense that wherever one predominates the other adjusts itself to it. But while this is so strikingly true of the centers we find many intermediately situated peoples practising the two or more special arts of the nearest centers, but less successfully. In North America particularly, we find a tendency for women to produce the geometric art and men the realistic. That this has an important psychological basis is unlikely since the distinction is clearest among the groups where hunting is the chief work of the men. Here the textile arts fall to the women, who thus find their activities limited. Among the Pueblo peoples on the other hand, where the men wove, we still find geometric art.

Finally, we must not forget that we have been but skimming over the surface of a very complex problem. Each small territory presents its own particular characteristics. Art, too,

Fig. 47. A Peruvian Poncho
This is a prehistoric textile of fine weave and typical in design

has everywhere strong individualities which tend to obscure the common elements, thus making every thorough survey of even a small area extremely exacting. The work of Kroeber^[27] in California demonstrates that often the large areas we have designated can be resolved into many small geographical sub-areas, which can, in some cases, be further differentiated into tribal types. However, all this is too intricate for discussion here.

SYMBOLISM

No discussion of our subject, however brief, can disregard symbolism. Though an old subject, it seems to have been given new life by Von den Steinen's observations in Brazil^[28] and Haddon's^[29] vigorous exposition of the realistic origin theory. Following this, with Boas^[30] as leader, a number of American anthropologists began an intensive study of designs in the basketry and bead work areas we have discussed. It was found that all tribes have names for many of their designs and in some cases, at least, employ these names to express ideas. Since these are almost always derived from familiar objects, as bird, feather, tree, etc., we are confronted with the possibility that the names were given at a time when the design was truly pictographic. This theory must be considered notwithstanding that we found certain objections to such origins in the influence of the technique. Accordingly, we have this problem: When a design is called by a definite name, is that name a clue to its historic origin?

The study of design names shows that this nomenclature develops according to the practical needs of the workers, for among the Pomo^[31] and Dakota,^[32] who lead in their respective centers, designs have been analyzed into their structural elements and names given to the same. Further, when definite composite designs have been established, the names of the separate design elements in the complex are compounded into a single term. In other words, we have an intense systematization of design composition, with a corresponding terminology. When we turn to less specialized decorations like the Maidu^[33] and Arapaho^[34] we find that they have a much longer list of design names, which upon inspection prove to be the result of a less elaborate classification and a failure to comprehend the advantages of design analysis. This forcibly suggests that the present association between a design and its name is quite likely to be the result of other than genetic causes.

Fig. 48. A Series of Designs and Their Names, from the Dakota Indians: a, Twisted; b, Full-of-points; c, Forked tree; d. Dragon-fly; e, Filled-up; f, Tripe; g, Feathers; h, Leaf; i, Tent; j, Arrow; k, Three-row; l, Vertebræ; m, Whirlwind; n, Bag; o, Pointed; p, Trails; q, Cut-out

Another way of testing the case is to compare the designs associated with one name. For example, from the special literature we find "flying goose" designs among the Tlingit, Thompson, Pit River, Maidu, Wintun, and Yurok, but we fail to find these designs identical or even similar. The tabulation of "butterfly" designs gives the same kind of result. The converse of this experiment is to take a single design and tabulate its names. Thus, the Pomo "quail tip" design is found elsewhere under the names bushes, pine cones, mountains, squirrel foot, and foot. This suggests that we must allow for the borrowing of both designs and names independently, or at least for the former.

Now while this is very good argument against the wide application of the design name theory of origin, it does not by any means prove that in the beginning the decorator did not copy from nature, for subsequent and repeated borrowing would completely disassociate the names. On the other hand, the steady growth of this art would produce a conventional naming system of whose existence we have good evidence in the published studies. Also, the acquisition of textile decoration requires the comprehension of simple steps, or elements, before mastery can be acquired over complexes. It is inconceivable that decorative art began with the most complicated designs and developed into the simplest; and although it sometimes happens that designs do degenerate to mere dots and bars, yet there is no reason for believing that the whole of decorative design was evolved in this way.

Unfortunately, we lack similar studies for pottery decorations, but the objective analysis of certain local types by Fewkes^[35] gives us ground for suspecting an analogous relation of names and designs. It is clear, however, that in pottery decoration we have different technical conditions; yet, one must assume that beginners would start with very simple forms, as in textiles.

However, in the art of most peoples we find a few designs that rise to the level of true symbols. Among the best known New World symbols are the cloud terraces of the Pueblo peoples and the "whirling logs" or swastika^[36] of the Navajo. The list is, however, very short, but in addition we find many degrees of symbolic association as among the Arapaho, where current designs were often chosen by an individual to stand for some personal interests peculiar to himself. Again, not a single case of real symbolism has so far been reported for the many basket makers of California. Its strongest development is in the Southwest which is, perhaps, the center of its northern dispersion. Among the Navajo we note that because of their sacred character the true symbols are not used in blankets, and in the Plains we further note that the conventional and æsthetic relations are practically never modified to meet the demands of interpretation; it is always the latter that is sacrificed. All this indicates that we are dealing with decoration primarily, upon which is occasionally grafted some symbolism. The facts are that practically all of the religious art of the

Fig. 49. True Symbols. The first represents the clouds, or "cloud terrace" of the Pueblo Indians; the second, the swastika, or "whirling logs" of the Navajo

New World is highly realistic and, therefore, stands apart from the art of ordinary decoration.

In conclusion, we may recall our initial question: Is the pattern name at its inception symbolic, or even representative? We can safely say that in most cases it was certainly neither. The suggestion is that symbolic art is primarily realistic, and that many true symbols may be explained as derived from pictures; but true symbols are relatively rare in the geometric designs we have studied and we have consequently no good reason for assuming that many of these as a class were once realistic. In short, the problem is an historical one. We have seen that geometric art is sometimes under pressure from realistic art and perhaps is always so. Hence, the feeling that its designs should be representative may universally arise and so account for all these design interpretations as secondary phenomena.

---

1. Spinden, 1913. I; Kidder, 1915. I.
2. Holmes, 1888. I.
3. Schmidt, 1905. I, p. 330.
4. Kidder, 1915. I; James, 1914. I.
5. Fewkes, 1898. I.
6. James, 1914. I.
7. Hough, 1907. I.
8. Fewkes, 1914. I.
9. Boas, 1903, I; Kroeber, 1908. II.
10. Kroeber, 1905. I.
11. Speck, 1914. I.
12. Speck, 1914. I.
13. Boas, 1897. I.
14. Emmons, 1907. I.
15. Boas, in Emmons, 1907. I.
16. Emmons, 1903. I.
17. Swanton, 1911. I.
18. Tozzer, 1907. I.
19. Lumholtz, 1900. I.
20. Spinden, 1913. I, p. 148.
21. Lumholtz, 1904. I.
22. Holmes, 1888. II; MacCurdy, 1911. I.
23. Joyce, 1914. I.
24. Mead, 1906. I.
25. Von den Steinen, 1897. I; Koch-Grünberg, 1908. I.
26. Whiffen, 1915. I.
27. Kroeber, 1905. I.
28. Von den Steinen, 1897. I.
29. Haddon, 1902. I.
30. Boas, 1903. I.
31. Barrett, 1908. I.
32. Wissler, 1904. I.
33. Dixon, 1902. I.
34. Kroeber, 1902. I.
35. Fewkes, 1914. I.
36. Wilson, Thomas, 1896. I.; Matthews, 1902. I.