UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Shadow and substance : a computer assisted study of Niska and Gitksan totem poles Shane, Audrey Mackay 1978

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata


831-UBC_1978_A8 S53.pdf [ 23.19MB ]
JSON: 831-1.0094378.json
JSON-LD: 831-1.0094378-ld.json
RDF/XML (Pretty): 831-1.0094378-rdf.xml
RDF/JSON: 831-1.0094378-rdf.json
Turtle: 831-1.0094378-turtle.txt
N-Triples: 831-1.0094378-rdf-ntriples.txt
Original Record: 831-1.0094378-source.json
Full Text

Full Text

SHADOW AND SUBSTANCE: A COMPUTER ASSISTED STUDY OF NISKA AND GITKSAN TOTEM POLES by AUDREY PATRICIA MACKAY SHANE Dip. I.D. University of Manitoba, 1942 B.A. University of British Columbia, 1974 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS in THE FACULTY OF GRADUATE STUDIES Department of Anthropology and Sociology We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA September, 1978 © Audrey Patricia Mackay Shane, 1978 In presenting th i s thes is in pa r t i a l fu l f i lment of the requirements for an advanced degree at the Univers i ty of B r i t i s h Columbia, I agree that the L ibrary sha l l make it f ree ly ava i lab le for reference and study. I fur ther agree that permission for extensive copying of th is thesis for scho lar ly purposes may be granted by the Head of my Department or by his representat ives . It is understood that copying or pub l i ca t ion of th is thes is for f inanc ia l gain sha l l not be allowed without my writ ten permission. Department of Anthropology and Sociology The Univers i ty of B r i t i s h Columbia 2075 Wesbrook Place Vancouver, Canada V6T 1W5 Date October 4. 1978 MAP 1. TSIMSHTAN VILLAGES, BRITISH COLUMBIA A f t er George Via c Donaid N a t i o n a l Museums of Canada. Social anthropologists only occasionally turn their attention to art, and then usually with some uneasiness. There is the lurking suggestion that this interest, however indirect, might on the one hand undermine their s c i e n t i f i c approach, and on the other c a l l forth from their colleagues one of the current terms of disparagement: ethnologist or ethnographer, with museum or "cultural" leanings. I realize of course that there are outstanding exceptions; but i t s t i l l seems necessary to emphasize that art is a legitimate topic for anthropological consideration, and one which has not had the attention i t deserves. Ronald Berndt, 1958 ib ABSTRACT This thesis attempts to distinguish varying styles in a particular set of massive carvings from the Northwest Coast of North America, the totem poles of the Niska and Gitksan. The method of investigation is based on the use of hierarchical clustering and multi-dimensional scaling computer programmes. These programmes are of a type used in ecological, geological, and archaeological studies. Their purpose is to establish a numerical taxonomy from which inferences may be drawn. The data used in the study are based exclusively on photographs, and i t is possible to include artifacts no longer in existence. There is an ethnographic record against which the success of the methodology is measured. It i s concluded that there are four distinctive styles of carving and organizing the totem poles. Two of these are attributed to the Niska and two to the Gitksan. A rhythm of order i s demonstrated in the placement of figures on the poles. It i s concluded that the taxonomy gives positive support to the hypotheses of previous investigators in regard to clan formation: originally there was a two-fold rather than a four-fold division among these Tsimshian groups. Traits associated with i i individual artists are not defined by the programmes, although associated traits preferred in certain locations are described. i i i ACKNOWLEDGMENTS I wish to acknowledge-with deep gratitude, the support of many friends and colleagues during the course of this study. First of a l l , thanks are due to Dr. Harry Hawthorn, Mrs. Audrey Hawthorn, and other members of the Department of Anthropology and Sociology, University of British Columbia, who made i t possible for me to embark upon the third career of my l i f e . As chairman of my committee, Dr. Marjorie Halpin has contributed encouragement and constructive criticism based on her thorough knowledge of Tsimshian l i f e and art. Other members of the committee, Dr. K.O.L. Burridge, Dr. James Caswell, and Dr. R.G. Matson have offered advice and assistance i n their several specialized fields. A former member of the committee, Wilson Duff, was generous in sharing his unique insights on Northwest Coast art although he remained sceptical on the subject of blending art and computers. I am indebted to the University of British Columbia for the support of a Fellowship prior to my appointment to the staff of the Museum of Anthropology. A l l of my colleagues have been helpful in countless ways. Special acknowledgment is due to the Director of the Museum of Anthropology, Dr. Michael M. Ames, for his active encouragement of research on the part of his staff. iv Lastly, as is usual, I wish to acknowledge the warm and invaluable support of my husband and family. They have borne my transition from wife and mother to student with seldom f a i l i n g patience. TABLE OF CONTENTS Page ACKNOWLEDGEMENTS i i LIST OF TABLES v LIST OF ILLUSTRATIONS v i INTRODUCTION 1 Chapter I. THE TSIMSHIAN PEOPLE 4 II. STYLE, ART, AND ANTHROPOLOGY 14 III. METHODOLOGY . . . 25 IV. DISCUSSION . 37 V. CONCLUSIONS 61 EPILOGUE 70 APPENDICES . . 108 BIBLIOGRAPHY 120 v i LIST OF TABLES Table Page 1. Key to the Identity of Gitksan Totem Poles . . . . 37 2. Key to the Identity of Niska Totem Poles 40 3. Poles in the Order of Dendrogram Clustering, Appendix II 41 4. Percentages of Poles Raised by Clans, by Decade . 51 5. Niska and Gitksan Poles Clustered by Carving Characteristics 51 6. Preferred Character Traits, by Location 58 vxi ' LIST OF ILLUSTRATIONS Figure Page 1. Vectors 1 and 2, Multi-dimensional Scaling, Dendrogram, Appendix II 48 Plate I. Figure 1. Pole 1 71 Figure 2. Pole 84 Figure 3. Pole 2 II. Figure 1. Pole 66 72 Figure 2. Pole 6 Figure 3. Pole 93 III. Figure 1. Pole 3 73 Figure 2. Pole 22 IV. Figure 1. Pole 54 74 V. Figure 1. Pole 56 75 VI. Figure 1. Pole 92 . . 76 Figure 2. Pole 7 VII. Figure 1. Pole 26 77 Figure 2. Pole 12 VIII. Figure 1. Pole 46 78 Figure 2. Pole 38 Figure 3. Pole 45 IX. Figure 1. Pole 61 . . . . . 79 Figure 2. Pole 61, detail X. Figure 1. Pole 9 80 Figure 2. Pole 18 Figure 3. Pole 47 XI. Figure 1. Pole 43 81 Figure 2. Pole 10 Figure 3. Pole 33 v i i i LIST OF ILLUSTRATIONS—Continued Plate p a g e XII. Figure 1. Pole 23 82 Figure 2. Pole 44 XIII. Figure 1. Pole 24 83 Figure 2. Pole 32 Figure 3. Pole 31 XIV. Figure 1. Pole 68 84 Figure 2. Pole 29 Figure 3. Pole 34 XV. Figure 1. Pole 67 85 Figure 2. Pole 79 XVI. Figure 1. Pole 4 . 86 Figure 2. Pole 48 Figure 3. Pole 89 XVII. Figure 1. Pole 78 87 Figure 2. Pole 13 Figure 3. Pole 27 XVIII. Figure 1. Pole 60 88 Figure 2. Pole 39 Figure 3. Pole 49 Figure 4. Pole 20 XIX. Figure 1. Pole 71 89 Figure 2. Pole 37 Figure 3. Pole 55 Figure 4. Pole 62 XX. Figure 1. Pole 8 90 Figure 2. Pole 25 XXI. Figure 1. Pole 16 91 Figure 2. Pole 17 Figure 3. Pole 14 Figure 4. Pole 15 ix LIST OF ILLUSTRATIONS—Continued Plate Page XXII. Figure 1. Pole 30 92 Figure 2. Pole 36 Figure 3. Pole 91 XXIII. Figure 1. Pole 35 93 Figure 2. Pole 87 Figure 3. Pole 95 XXIV. Figure 1. Pole 80 94 Figure 2. Pole 83 Figure 3. Pole 85 Figure 4. Pole 86 XXV. Figure 1. Pole 11 . . . . . . . . . . . . 95 Figure 2. Pole 72 Figure 3. Pole 73 XXVI. Figure 1. Pole 63 96 Figure 2. Pole 96 Figure 3. Pole 76 XXVII. Figure 1. Pole 58 97 Figure 2. Pole 59, detail , . . , Figure 3. Pole 77 XXVIII. Figure 1. Pole 28 98 Figure 2. Pole 51 Figure 3. Pole 88 XXIX. Figure 1. Pole 90 99 Figure 2. Pole 74 Figure 3. Pole 81 XXX. Figure 1. Pole 19 100 Figure 2. Pole 53 Figure 3. Pole 94 XXXI. Figure 1. Pole 21 101 XXXII. Figure 1. Pole 40 102 Figure 2. Pole 52 , x: LIST OF ILLUSTRATIONS—Continued Plate Page XXXIII. Figure 1. Pole 5 103 Figure 2. Pole 41 Figure 3. Pole 64 XXXIV. Figure 1. Pole 65 104 Figure 2. Pole 82 XXXV. Figure 1. Pole 75 105 XXXVI. Figure 1. Pole 42' 106 Figure 2. Pole 57 Figure 3. Pole 50 XXXVII. Figure 1. Pole 69 107 Figure 2. Pole 70 Map Map 1 Frontispiece. x i INTRODUCTION The purpose of this thesis is to demonstrate the u t i l i t y of numeric taxonomic classification by means of a computer, as a tool in determining relationships between artifacts of a complex nature. Specifically, i t is hoped that such a method can provide significant and reliable results when applied to sculptural forms for the purpose of distinguishing a particular style of an individual or group. Furthermore, in drawing exclusively upon photographs for a data base, i t is hoped to encourage more studies of artifacts not physically ~ available to the researchers; in the case of this thesis, some of the material described has been dust for nearly f i f t y years - yet images of i t remain, potentially of value. The data to be used are photographs of Niska and Gitksan totem poles drawn from the publications, of Marius Barbeau,. the f i l e s of Charles Borden, the f i l e s of Wilson Duff, and. the f i l e s of the National Museums of Canada. The selection for further investigation of a group of objects already well known and widely published may seem redundant. It should be emphasized, however, that none of the written information w i l l be used as part of the data base. This information, then, exists as an independent gauge of the r e l i a b i l i t y of the method. Conversely, lack of " f i t " between the ethnographic record and the results of computer clustering w i l l pinpoint problem areas. The choice of artifacts for investigation also reflects a deep 1 2. sense of frustration, which is surely shared by a generation of researchers in Northwest Coast art. There are only two general sources for the study of totem poles. These are the Totem Poles of the Gitksan, Upper  Skeena River, British Columbia (1929), and Totem Poles (1950), both published by Marius Barbeau. Without them, a body of monumental sculp-tural work of international importance would be l i t t l e known. Yet the organization of information in these books is haphazard, repetitive, and daunting to a l l but the most persistent readers. The less per-sistent w i l l not notice a number of inaccuracies and contradictions; worse s t i l l , they w i l l accept some of the conclusions without question. Although I remain unconvinced of the accuracy of some of Barbeau's theories and a r t i s t i c judgments, i t would be ungracious, and certainly less than f a i r not to acknowledge our debt both to him and to the Tsimshian ethnographer, William Beynon, his collaborator. The data used for the thesis consist of ninety-six totem poles from the iNass and Upper Skeena Rivers of British Columbia. These poles w i l l be described by means of a character l i s t . A "character" may be any aspect of the pole which is judged by the investigator to be relevant to the study. Such a character must be capable of being scored as "present" or "absent". I have chosen to restrict these characters to those descriptive of form rather than iconography. This choice is based on the assumption that an a r t i s t , or group of artists may be more readily identified by means of the manner in which their work is conceived and executed rather than by the content of the work. 3. The quantitative techniques employed to transform the score of each pole (in terms of the presence or absence of any given character) into a measurement of similarity, w i l l be described in detail in a later chapter. I believe a technical analysis should precede and complement any intuitive judgmentetof relationships in art styles. I also believe such an analysis may be greatly assisted by the use of a computer in appropriate cases. Particularly when dealing with a data set both complicated and subtle in i t s variations, the sorting capabilities of a computer are invaluable. Only the investigator, with intuitive judg-ment, can determine the meaning of the results of such a sorting process. Employment of a quantitative methodology makes i t unnecessary to rely wholly upon the reputation of the investigator, as has been the case too often in the past. It i s hoped that the thesis w i l l : 1. Establish sub^-populations of poles which may be defined as belonging to a particular configuration which constitutes a style. 2. Define some clusters of traits which may be associated with a particular artist or group of artists. 3. Demonstrate that the placement of figures on poles is not random. 4. Illustrate the potentialities inherent in applying a new method to old knowledge. The end result should provide a definition of the style of Niska and Gitksan totem pole carving which is both empirically based and more precise than any now available for this plastic art (Barbeau, 1929, 1950; Wingert, 1951). CHAPTER 1 THE TSIMSHIAN PEOPLE The artists whose styles I hope to define are responsible for some of the finest massive carvings in the world, totem poles. Totem poles were once an essential part of the l i f e of the Tsimshian, whose home i s on the Northwest Coast of North America.(Tnpl939, -L A.L. Kroeber published his definitive work Cultural and Natural Areas  of Native North America, in which his stated purpose was to examine both environmental relations and historic relations of culture areas or geographical units. On the basis of culture t r a i t s , patterns, and geography, Kroeber identified a unit now known as the Northwest Coast. This area stretches from Yakutat Bay in Southeast Alaska to Trinidad Bay in Northern California. Kroeber's definition is now a standard in the literature. While the connection between the northernmost Tlingit and the southern Hupa may appear tenuous, similarities can certainly be seen along the continuum which l i e s between them i f this i s placed in con-text with cultures of the rest of the Americas. Nevertheless, the image which usually comes to mind when the Northwest Coast is referred to i s that of a complex which includes feasting, masking, totemppoles, and types of social organization which belong particularly to the area lying from Campbell River on Vancouver Island, and the adjacent mainland northward. Part of this forms the distinct Northern Maritime Province, 4 5. restricted to the Tlingit, Haida, Tsimshian, and Northern Kwakiutl (Haisla), a l l of whom share the important concept of matrilineality as well as other traits which tend to set them apart from the rest of the Northwest Coast (Drucker, 1955: 109). The Tsimshian people, whose culture is the focus of this paper, have been described by Boas (1916) and, following further f i e l d work, by his pupil, Garfield.(1939). These papers are the only published descriptions of traditional culture available to us, and i n some im-portant respects they, have been further refined by Halpin (1973). Halpin has presented a clear and useful outline of the Tsimshian and their social organization: which -is a distinct improvement on previously existing outlines. With her permission, the following discussion w i l l contain extensive quotations from her.dissertation.. There is confusionn and'inconsistency in the literature regarding li n g u i s t i c , cultural,.and geographic divisions of Tsimshian-speaking people. The following classification is proposed and used herein to avoid ambiguity and clearly distinguish l i n g u i s t i c (e.g., Tsimshian) and major (e.g., Coast Tsimshian) .and, minor (e.g., Lower Skeena Tsimshian) '•"'cultural-geographic units or divisions: Tsimshian is a linguistic designation for people speaking two related languages (Rigsby, 1969): Coast Tsimshian, IandJ Nass-Gitksan. These two language groups are divided into three broad cultural and geographic divisions: Coast Tsimshian, liv i n g along the lower Skeena River up to and including i t s canyon, and the coasts and islands from the mouth of the Nass south to Milbanke Sound; Niska, l i v i n g along the Nass River and Portland Canal; [and] Gitksan, living along the upper Skeena River or i t s tributaries. The Gitksan are often referred to in the literature as the people of the Upper Skeena; 6. ' Tribes: The Tsimshian lived in some twenty-six local groups usually referred to as tribes. Each tribe customarily occupied a single winter village, often of the same name. The tribes [whose poles are discussed] are lis t e d below according to the preceding geo-cultural divisions. . . The spelling of t r i b a l names is based upon Marius Barbeau's f i e l d orthography and English translations are approximations derived from his f i e l d notes, except where another source is given. The accepted names and spellings of present Tsimshian bands are given in brackets (from Duff, 1964: 18-20). . .[Coast Tsimshian omitted here] Generally speaking, the four Niska tribes have not retained clear and separate identities to the same degree as have their Coast Tsimshian and Gitksan neighbours. There was considerable population movement in this area, and the people who comprised the tribes and lived in the different villages are not always clearly distinguishable on the basis of present data. . . The people of the lower Nass called themselves the g' itxat' i^n and were divided into two tribes: the g'itxat' i.'n (proper) and the g* it g ' i^'Eotix, who were a small offshoot population that moved a short distance upriver to the village of antegwalE '. The g'itxat'i/n did not conform to the usual pattern of l i v i n g in a single winter village, and they and the g* itg* ig'E/nix experienced a number of population shifts so that their history of settlement is complicated. Since the late decades of the 19th century they have been living at the two modern villages of Kincolith ("place of scalps") and Greenville or laxg.aldzap ("on deserted village s i t e " ) . The four villages listed below for the g'itxat^i/n were 19th century "totem pole" villages. . . 1. g* i t 'iks ("people of iks." 1 an exclamation) 2. kwunws^q ("where people sleep" when travelling) 3. ang*edE/ ("where they catch eulachon with rakes") 4. g'itlax*a /us ("people on the sandbar"). . . The people of the upper Nass weretthe g* it g n w i l i k s ("people staying temporarily," referring to their movement down the river at eulachon fishing time). They were divided into two tribes: the g* itw^nksi'lk, ("people of the place of lizards") [Canyon City] who lived at the canyon, and the dom-inant g* itlaxda.'mks, ("people on the place of springs") [Gitlakadamix] who lived a few miles above them. Around the turn of the century the old village of g* itwgnksi^lk burned down and the people settled at Gwinaha. They have since moved to Canyon City. The Christians of g* itlaxda/mks moved to the missionary village ai >ya /nc ("early leaves") and, when i t was flooded ca. 1918, they re-turned to g'itlaxdafmks, which was then given the name ai^ya^nc. Recently, the g*itlaxda.mks have moved to the modern village of New Aiyansh. . . 7. The seven Gitksan tribes each occupied a single winter vil l a g e , six of them on or near the Skeena and one, k'jtwgn-fcku-'l, to the north on the "grease t r a i l " to the Nass. About 1880 another small tribe, the anlag.as^mdE'n, joined the k*isg.ag.a/s. Three Christian communities were founded between 1890 and 1910. Glen Vowell, the only one s t i l l in existence, drew i t s converts from k*ispayaks, k*isg.ag.a.s, and qaldo/; Andimaul ("where they fi s h with hand lines") mostly from k'itsggu'kla; and Meanskinisht ("at the base of the big mountain") from k* itwang. E / and k*itwgnlku / jl. g'it* anma'ks, at the site of the white settlement of Skeena Forks or Hazelton (founded in the late 1860's), att-racted people from neighbouring villages and i t s originally small population now exceeds a l l of the others; the k'isg.ag.a.s have now completely amalgamated with them. The people of qald o / have amalgamated with k*ispayaks, and many of the k* itwgn3:ku/:> 1 moved to the Nass in the late 19th century to l i v e at Aiyansh and Kincolith. . . Descent Groups Houses: The basic social unit in a l l three divisions of the Tsimshian was a corporate matrilineage called a "house" (wglp) and named after i t s highest-ranking chief's name. Larger lineages fissioned into branch lineages, each named after i t s own house chief, but subordinate to the highest-raking chief of i t s major branch or segment. The house as a matrilineal descent group was not coterminous with the household in which lived members of the matrilineage plus their affines and children belonging to other lineages. The larger matrilineages often occupied more than one dwelling, which were named according to a different system: dwelling names were inherited as crests (ayiiks), of which high-ranking lineages usually owned several. Branch lineages could draw upon the stock of house names owned by the parent house. The house was the principal resource-owning corporation in Tsimshian society. Its resources included fishing, hunting, and gathering territories or l o c a l i t i e s , which were exploited under the direction of the house chief. The house also owned a stock of ceremonial privileges: names (of several taypes), crests, myths, songs, and feast prerogatives, which were also under control or stewardship of i t s chief. We know that houses fluctuated widely in size, at times resorting to adoption to prevent extinction, at other times growing so large that they fissioned into two or more separate houses. But there are no descriptions in the Barbeau notes or {until recently] in the published literature, of intra-house composition and dynamics [Adams: 1973], The lines of fission mentioned in traditional narratives were between brothers. 8. Clans: Each house and i n d i v i d u a l belonged to a l a r g e r , exogamic, s t i p u l a t e d , m a t r i l i n e a l group which i s u s u a l l y c a l l e d a phratry i n the l i t e r a t u r e , but which I am c a l l i n g a clan. There were four, each represented i n a l l three d i v i s i o n s (the names of two of them change with the Gitksan), although not a l l four clans were to be found i n a l l t r i b e s . The names and p r i n c i p a l crests of each clan are as follows: . . . TABLE I TSIMSHIAN CLANS AND THEIR PRINCIPAL CREST ANIMALS Coast xcgianha'da* l a x k * i b u / g'isp^wudwa'da la x s k ' i ' k Tsimshian (?) ("on the (?) ("on the wolf") eagle") Niska I I I I I I I I Gitksan laxse.'l** I I g " i s j . a.'st I I (g.anha /d^ i n ("people of Kitsegukla) the fireweed") Crests Raven Wolf Killerwhale Eagle Frog Bear G r i z z l y Beaver g.anha /da may be derived from the T l i n g i t Raven clan name ganaxadi ("people of ganax") ( G a r f i e l d , 1966: 19). ** l a x s e / l may be derived from the Tsimshian word, for the T l i n g i t v i l l a g e at Cape Fox: l a x s e l ^ . Clan exogamy was extended to the corresdponding clans (moieties) of the neighbouring T l i n g i t and Haida, for which the four Tsimshian clans were grouped i n two p a i r s , as follows: TABLE I I TSIMSHIAN, HAIDA, AND TLINGIT CLANS AND CRESTS Haida Ravens T l i n g i t Wolves Crests T l i n g i t Ravens Haida Eagles Tsimshian g r i z z l y raven Tsimshian g * ispawudwa'dj g*isg.a/st k i l l e r w h a l e frog g.anha /d9 l a x s e / l laxk"ibu' wolf bear eagle beaver laxsk'i.'k 9. Adams has complicated the picture sketched by Halpin by proposing that lineages are restricted solely to people who can claim a mother's mother in common, stating that this is due to a belief that matrilineal grandparents are reincarnated through their descen-dants. Furthermore he states "Houses are not descent groups" (Adams, 1973: 31). This seems strange in the light of his earlier statement that "The cultural forms we observe today are vir t u a l l y the same as those reported for the Coast Tsimshian (Boas, 1916; Garfield; 1939) " (Adams, 1973: 22). Garfield refers to a l l the members of a house as part of a lineage, and furnishes a diagram whose members are clearly not restricted to those with a "mother's mother" in common (Garfield, 1939: 276-279). Pending c l a r i f i c a t i o n of Adams' definition, i t seems best to adhere to previous proposals. Origins of the clans, as explained by the histories of the Tsimshian, are quite diversified. Unlike the histories of the Kwakiutl and Bella Coola, who reckon descent from both parents and who are specific about the establishment of their villages by a particular semiedivine sky being (Garfield, 1951), the Tsimshian accounts emphasize the possession of common prerogatives obtained from ancestors, which descend through the maternal line. Thus their stories of migrations, beginning in some accounts with the migrations initiated by a Flood or a disastrous "lo c a l " snowfall which overwhelmed their ancestral home, Temlaham, often mention that wandering groups settled for a time with other groups with whom they shared clan a f f i l i a t i o n s and major crests. The 10. Kltwancool tribe, according to their histories, is a group of people of different origins; one group t e l l s of leaving their early home on the sea-coast, travelling up the Nass, and then moving into their present territory (Duff, 1959). A very important factor in the lives of a l l three Tsimshian divisions was the annual eulachon run which took place in early spring. This f i s h , which is also called the candlefish, i s extremely greasy, and may be rendered to provide an o i l much prized not only by the Tsimshian, but also by other tribes. A l l Tsimshian groups had traditional fishing stations along the Nass, and the gathering of the people there was the occasion for intensive trading, not only among themselves but with v i s i t i n g Haida and Tlingit. Trading also took place by means of the "grease t r a i l " an overland route which linked the Nass and Skeena rivers by way of the village of Kltwancool. The existence of this t r a i l no doubt was partly responsible for the links which existed between the Niska and Gitksan - links which were perceived to be closer than those which either group had with the Coast Tsimshian. This is a land which like any land trodden by human beings has been enjoyed, exchanged, fought over and loved for millenia. Archaeologists have verified the early occupation of the site at the mouth of the Skeena which we now c a l l Prince Rupert - and this was at least five thousand years ago. "These traditions are confirmed by archaeological record, which indicates that the Tsimshian have occupied Prince Rupert harbour and the surrounding coastline con-tinuously for at least 5,000 years " (MacDonald and Inglis, 1976: 34). 11. MacDonald cites the evidence of artifacts to argue that, contrary to some other proposals (Barbeau, 1929; Garfield, 1939), the Tsimshian can be assumed to have occupied the site continuously during that five thousand year span. Like their neighbours the Tsimshian depended heavily for their livelihood on salmon fishing, seal and sea otter hunting, eulachon fishing, and, in the inland areas, the hunting of mountain goat and mountain sheep. This was supplemented by the gathering of a wide variety of berries in season, as well as some root plants. In the good years, the hunting and gathering of the spring, summer, and f a l l allowed the people to store enough food for the winter, as well as to build up the surplus needed for trade, capital, and to support the expenditures necessary.to maintain and enhance social prestige. Although i t is customary to characterize this land as a land of plenty, and i t may have been so in contrast, for instance to the territories of the northern Ojibwa, i t was not invariably so. For reasons s t i l l obscure, the salmon run cannot always be depended upon; and the traditions of the people contain many references to starvation. Nor should i t be forgotten that even when the bounty of sea and land was available, i t took strength, knowledge, and courage to gather such resources. Both the words of the writer and the drawings of the art i s t are inadequate to convey the physical reality of this country, the setting for so many bold adventures and supernatural experiences. The sea both nourishes and k i l l s ; sparkling, rich in treasures, but 12. always ready to suck the unwary into the whirlpool abode of an e v i l s p i r i t or enfold him in a robe of bone chi l l i n g fog for endless days. The forest is the source of shelter, fuel, and clothing, but i t s heavy dank undergrowth also conceals dangerous animals and even more dangerous ghosts. To the successful hunter the mountain offers the prized fat of the mountain goat as well as i t s meat, horns, and hair - i t also offers freezing winds, storms, and icy slopes. As well as being bountiful and dangerous, this is a beauti-f u l land. It is a land of mountains whose open meadows are speckled with flowers, somnolent in the heat of late spring. It is a land of forests whose cedar fronds are pungently scented, and whose yellow leaved cottonwoods bend against the blue sky in the brisk wind of autumn. It is a land whose t i d a l f l a t s of water sculptured sand are gentle beneath the bare feet in the warm dusk of summer. Above a l l i t i s a land of artists who are keen observers of the world around them. When winter came i t was time to retreat to the winter villages, to carve, paint, and participate in the sacred season. In a l l of these activities the cedar tree played a central role. Integral to l i f e , both secular and sp i r i t u a l , was the red and the yellow cedar. Houses, canoes, boxes and clothes, a l l were of cedar, easy to carve and beautiful when finished. Totem poles, too, were of cedar. Totem poles could be raised as free-standing memorials, or as frontal poles for houses. It was not an endeavor which occured every day; i t was the culmination of years of planning by a chief and his lineage. A l l were part of a social system in which the lineage 13. gained prestige as their chief gained; the system was not static. An essential part of the social fabric involved the assumption of high ranking names and their accompanying responsibilities and p r i -vileges. Public validation of such an assumption was necessary; the validation was accomplished during the course of a ceremonial feast, and the raising of a totem pole. The raising of the pole was the v i s i b l e and permanent record of the validation of the rank and t e r r i t o r i a l claims of a lineage. Preparations for raising a pole were lengthy and expensive—one of the major considerations was the choice of a carver. According to Barbeau: The privilege of carving the pole and rendering specific ceremonial services for a l i b e r a l stipend f e l l to a smaller cir c l e of strangers, who may be termed a l l i e s or relatives by marriage. Not every a r t i s t , though a stranger, could be invited to carve a pole, as has often been supposed even among ethnologists. Far from i t . He must, indeed, be selected from among the "fathers" of the deceased or his heir; in other words, he must be either one of the "fathers" of the members of this family or one of their immediate relatives according to native computation. The "fathers" always belonged to another I clan] as no one was ever allowed to marry within one's own [clan] even with members . . . wholly unrelated. The carver whose services were sought was as&a rule the best available from among the "fathers." When he lacked the required a b i l i t y , he himself appointed a substitute who did the actual work while he "stood over him," as the saying goes. He otherwise assumed the credit of the work. This sometimes made i t d i f f i c u l t to find the name of the actual carvers, after the lapse of many years (Barbeau, 1929: 7). to distinguish the work of such a carver, or group of carvers, is one of the aims of this thesis; therefore the next chapter w i l l consider the important question of style. CHAPTER 2 STYLE, ART, AND ANTHROPOLOGY The study of style i s an integral part of the study of mat-er i a l culture. What is material culture? According to Kubler (1965: 9): The term includes both artifacts and works of art, both replicas and unique examples, both tools and expressions -in short a l l materials worked by human hands under the guidance of connected ideas developed in temporal sequence. From a l l these things a shape in time emerges. A vi s i b l e portrait of the collective identity, whether tribe, class, or nation comes into being. This self image reflected in things i s a guide and point of reference to the group for the future, and i t eventually becomes the portrait given to posterity. Style may be understood f i r s t , then, as part of the "shape in time" of a particular culture. In his classic paper on the subject of style, Schapiro (1962: 297) says: Although the attempts to explain styles as an a r t i s t i c expression of a world view or mode of thought are often a drastic reduction of the concreteness and richness of art, they have been helpful in revealing unsuspected levels of meaning in art. They have established the practice of inter-preting the style i t s e l f as an inner content of the art, especially in the nonrepresentational arts. They correspond ' to"the'conviction of modern artists that the form elements and structure are a deeply meaningful whole related to metaphysical views. As part of culture, style should be of interest to anthro-pologists; and indeed i t was an issue of considerable importance in anthropology nearly a century-ago. What does "style" mean? Art historians may use i t as a general term, embracing both form and content, and may further define i t as being used in two different senses, depending on 14. 15. whether or not i t is placed in a particular historical context. If i t i s not so placed, i t may be considered "roughly equivalent to a ''complex recurrent type' . . . A style in this sense i s not limited to any one period, place, or people in history; i t may occur in many times and places" (Munro, 1970: 236). An example of the use of the word "style" in this sense i s evident in the terms "archaic style", "classic style", or "decadent style", widely applied in many different contexts. In the second sense, "style" means: An interrelated set of traits which i s characteristic of the art produced in a certain place and period, by a certain social group or individual a r t i s t . Style in this sense i s definitely linked with some historical provenance - chronological, geographical, social, or individual, or perhaps a l l of these. Works from that source are said to mainfest the style in i t s most authentic, typical, pure, or highly developed form, although imitations of i t or similarities to i t may be found elsewhere (Munro, 1970: 237). Style in the present thesis should be understood in the second sense, and should also be understood to deal primarily with form, not content; morphology, not iconography. The problem of the dichotomy of form and content has been dealt with by Panofsky (1962). He distinguishes "three strata of subject matter or meaning" (Panofsky, 1962: 16) which may be paraphrased as follows: primary or natural subject matter, both factual and ex-pressional, or the world of a r t i s t i c motifs whose enumeration con-stitutes a pre-iconographical description; secondary or conventional subject matter which relates motifs and combinations of motifs (compositions) to themes, the enumeration of which constitutes an iconographicalj analysis'.in the narrower sense;'.and int r i n s i c meaning 16. or content, the world of the discovery of symbolical values (manifestations of the underlying principles basic to the attitude of a nation, a period, a class, a religious or philosophical persuasion), or iconography in a deeper sense. The third clause represents a synthesis of form and content (Panofsky, 1962: 5-8). Form in the sense of pure shape, or relationship between such shapes,is a mere component of the pre-iconographical description. Panofsky's argument for this placement of form is that i t is impossible to give an adequate description of a shape without some knowlege of i t s context. His argument may be valid for the European corpus with which he deals, but i t is not necessarily valid in the context of primitive art.(Panofsky, 1962: 9). Application of Panofsky's methods presupposes access to a wide variety of literary records, as well as the exercise of a high degree of intuition. The highest controlling principle of inter-pretation rests on the art historian checking what he thinks is  the intrinsic meaning of the work of art against what he thinks is the intrinsic meaning of as many documents as he can master (Panofsky, 1962: 16). This is precisely what the anthropologist does when he" interprets^-the"meaning of camcomputer iclassif ication; the difference l i e s in the intervening steps, in other words in the basis for interpretation. Levi-Strauss (1963: 91) takes the view that "in 17. structural analysis i t i s impossible to disassociate form; from content". He has also been quoted in translation (Rossi, 1974: 69) asj-saying "form and matter have the same nature and are amenable to the same analysis. The content derives i t s reality from the structure, and what is called form is the 'placing in structure' of local structures, which is the content (1960: 21-22)." This concept post-dates the well known paper "Split Representation in the Art of Asia and America" (Levi-Strauss, 1967: 239-263) in which Levi-Strauss uses the term "style" in the sense of a distinctive association of decorative motifs, while at the same time saying that i t i s a "graphic or plastic projection of a reality of another order" (Levi-Strauss, 1967: 257). McLuhan's dictum "the medium is the message" also deals, probably, with form and content. One way of interpreting his cryptic phrase is suggested by Culkin (Crosby and Bond, 1968: 181); "The form of communication not only alters the content, but each form also has preferences for certain kinds of messages. Content always exists in some form and i s , therefore, to some degree governed by the dynamics of that form'." It is apparent that there is no clear general consensus on the meaning of the terms "style", "form", and "content" at the present time, although the trend of thought appears to be toward the synthesis of form and content. It was style mainly in Munro's second sense (1970: 237) 18. which concerned anthropologists in the closing decades of the nineteenth century, and the opening ones of the twentieth. In a discipline s t i l l relatively unspecialized, ethnologists and archaeologists remained closely in touch, often being wrapped together in a single person; the right hand was usually aware of what the l e f t hand was doing. Archaeologists were enthusiastically re-constructing ancient c i v i l i z a t i o n s , with much of the evidence being based on style, as witnessed for instance by the work of Arthur Evans at Knossos, and that of Flinders Petr.ie in Egypt. Much of Petrie's work, unlike Evans', has stood up well under the onslaught of succeeding generations equipped with superior technology; i t is oft-interest that some of Petrie's work bas based on principles of seriation: I w i l l return to seriation in the next chapter. Ethnologists were f u l l aware of the pragmatic value of style in archaeology, and of i t s possible application in studies of culture definition and culture change. Potentially of equal interest was the study of the role of the artist and the arts ast-communication and value reinforcing devices. Moreover art style was a component of the human experience worthy of study in i t s e l f . Somewhat surprising, however, i s the direction many of these studies took. Much thought and debate was given to the problem of the development and progression of design motifs in general (Haddon,1895; Holmes, 1886; Balfour, 1893; Boas, 1908). The fact that the argument could be cast in these terms reflects the preoccupation of anthropologists at the time with evolutionary theory. 19. One of the better known proponents of the "realism to abstraction" school of thought in design development was A.C. Haddon, marine biologist turned ethnologist. In discussing Haddon's theories Herskovits (1966: 244-245) has pointed out Haddon's developmental series purporting to show arrow designs from the Torres Straits moving from a recognizable crocodile motif to an unrecognizable abstraction, and comments: Does not a series of this sort, then, demonstrate how variations of a design indicate the way in which convention-alization must have developed out of realism? . . . Like a l l other attempts to devise developmental formulae of universal applicability, this one f a i l s both because of i t s faulty method and i t s failure to take a l l the facts into account. Haddon's arrows . . . were arranged in accordance with a scheme that existed in the mind of the student. The postu-lated development has vali d i t y only in terms of the hypo-thesis. A l l the objects belong to the same epoch and do not represent a developmental series in the sense that one item can be shown to have preceded the other. Many f r u i t f u l lines of investigation claimed the attention of early ethnologists—questions of social and mental, rather than material, culture. While art continued to have a place in the well rounded ethnographical description or monograph, i t s study became peripheral to the mainstream of thought in the decades after 1930. The causes of this decline probably relate to increasing special-ization within the discipline of anthropology; there was an increasing sense of the urgency and interest of ethnographic fieldwork, and although art style continued to be a useful adjunct to archaeology, in that context i t was not a matter for theoretical discussion. Not only primitive art studies, but the broader subject of 20. material culture has suffered from a relative decline in interest. Sturtevant (1969) has used a chronology in discussing material culture studies which reflects the shift in the academic base of the discipline of anthropology from museums to university departments. This sh i f t , practically complete by 1920, has emphasized the fact that important aspects of anthropological research, such as linguistics, kinship terminology, religion, and the family usually proceed satis-factorily without any reference whatever to material culture. According to Fenton: Studies of technological subjects have never really gone out of fashion and they persist in new guises and serve different ends. The commonplace that one hears in the profession—"there isn't much interest in material culture studies any more"—really says that fewer anthropologists proportionately are involved with materials per se; but the remark overlooks the important fact that other anthropologists are interested in the ideas that l i e behind the objects and are searching for data to analyze systematically and to treat s t a t i s t i c a l l y [emphasis added] (Richardson, 1975: 21). As the great era of gathering material culture is passing, i f not already gone, anthropologists may turn their ingenuity toward making more innovative use of what has been available for a long time: museum artifacts. Fenton has stated this idea very well: Ethnographic collections become of increasing importance as cultures vanish from the world. Since many of these contain specimens that are undocumented, we can learn much from the archaeologists about how to study them. Museums of Europe are f u l l of what I c a l l "Ethnological Chippendale": the specimens are what someone says they are; but where the^documentation exists to prove their authenticity, we can extrapolate from them to similar pieces, in the manner of art historians who can teach us much.(Richardson, 1975: 31). In the last decades of the twentieth century, we may be 21. entering an era of "networks" in research in which the inter-relationships of material culture and social and mental culture w i l l be perceived as a valuable study. The work of Claude Levi-Strauss in this f i e l d is of major importance. It has been deliberately excluded from consideration in this study since the relationship between transformations of iconography and variations in content is not of immediate concern. However i f we accept Levi-Strauss' position on form and content, then even the very restricted domain of style in the sense of form and form relationships, the domain of this study, inevitably has content. Some indication of this content may be made apparent by quantitative methods of classification. Levi-Strauss has said that "the principle underlying a classification can never be postulated in advance. It can only be discovered a posteriori by ethnographic investigation" f(Levi-Strauss, 1970: 58). A similar idea l i e s at the heart of this thesis. Before a principle can be discovered a classification must exist. This classification must arise from the data i t s e l f , insofar as this is possible, thereby allowing the museum artifacts to "speak." A numerical taxonomy w i l l provide just such a classification, as I w i l l discuss in the next chapter. Since this is a new approach to the old problem of describing artifacts and classifying them, I hope i t w i l l add a new dimension to studies of museum artifacts. A mathematical approach to the data may be one of the few methods of bridging the gap between material, mental, 22. and social culture studies in a credible manner. The quotation from Berndt which opened this paper reflects a state of affairs which may be rectified i f a new generation of students w i l l take the trouble to investigate the sophisticated world of modern technologyc-with a".view to -applying-new methods '-to^problems in anthropology. Such methods must be used very precisely, in con-junction withpwellst-ried methodsmofhthe past. The old method used in this study i s , of course, s t y l i s t i c analysis: the search for patterns in form. In the area of specific interest in this thesis,^the:-Nbrthwestr:Coastv there-have been several works which combine keen observation and analysis with clear description. Holm (1965) has described the principles of design organization and in addition (1972) has analyzed carving styles in some detail, particularly with reference to the humanoid head. He considers and describes these styles inva geometrical frame of reference. Kaufmann (1969) has contributed a major study on style in a r g i l l i t e carving. There is a growing body of journal articles, anthologies, and books on specific aspects of art in anthropology, indicating that the subject is alive (Brasser, 1975; 'd iAzevedS.v 1975; Duff, 1975; Forge, 1973; Halpin, 1973; Holm, 1965, 1973; Jppling, 1971; Otten, 1971). A paper by Inverarity, "Observations on Northwest Coast Indian Art and Similarities Between a Few Art Elements Distant in Time and Space" (1972), analyzed arm and leg positions on totem poles. The arm and leg positions used in this thesis are derived from his work. Disappointingly, the study which actually has the least to offer 23. is Tsimshian Sculpture (Wingert, 1951). It is true that this is described by the author as a preliminary attempt, but he also states that i t provides a methodology for developing further research in Northwest Coast style considerations. This is a dubious proposition. According to Wingert, Tsimshian totem poles "demonstrate emphatically" the present of a t r i b a l style. The distinctive elements of the style are described in the following passage: Important among these elements are the basic con-ception of a carved pole, the organization of the design, the principles of rhythm and balance, the sense of scale, the system of proportions and the sculptural technique employed. The concept of the pole as a vertical shaft with a curving sur-face is strongly expressed. The arrangement of the figures  in a series of clearly defined superimposed horizontal divisions Jemphasis added] confirms visually the convex curvature of the surface, while the rhythmic repetition of figures or parts of figures, often in an alternating system and usually with variety, carries the eye upward and stress v e r t i c a l i t y and height. Scale is further conveyed by the use of two sizes  of figures of practically the same proportions, one very small, the others very large [emphasis added] (Wingert, 1951: 89). While an emphatic t r i b a l style may exist, i t s identification from the cited description would hardly be clear-cut. In fact, only the passages underlined by the present writer provide any tangible assistance. Try, for example, reading the f i r s t sentence of the quotation without the phrase "of a carved pole." The sentence s t i l l makes sense. Or, to further determine whether or not Wingert is providing, as specified, a meaningful description of a Tsimshian totem pole, substitute for the deleted phrase "of a carved pole" the phrase "of a Buddha", or the phrase "of the Parthenon." There is no loss of descriptive power, because there is_ no descriptive power. If this were an isolated paragraph one could accept i t ; unfortunately i t is typical. It is also the sort of writing which creates general scepticism about the validity of art studies. In a literary sense i t may be beautiful, but i t advances our knowledge of or insight into the subject under discussion not at a l l . After twenty-seven years, this i s s t i l l a major reference on Tsimshian sculpture. To determine whether some improvement is possible, this thesis has turned to a different approach - quantitative techniques. CHAPTER 3 METHODOLOGY What have quantitative techniques to do with artifacts, art and style? I believe that just as everything in the natural world, from the construction of a nautilus shell to the movements of the stars may be described in mathematical terms, so i t i s with the cultural world of mankind: an inner harmony, which manifests i t s e l f as a style, exists, and is capable of description in mathematical terms. Such a description may then be translated into words, providing a secure armature to support aesthetic perceptions of the individual, whose intuitive judgments may or may not be reliable. Nor should statistics alone be considered i n f a l l i b l e ; but when the two approaches can be brought together, complementing one another, the result should be an enrichment of knowledge. The belief that a mathematical approach may be useful in anthropological studies is scarcely new. Johnson (1972) has drawn attention to the role of Franz Boas as an innovator in his use of a matrix of similarity values as an aid to organizing and analyzing anthropological data. Interestingly enough, in a paper published in 1895, Boas used an elementary Q technique of seriation and clustering based on folktale motifs to pinpoint the Tsimshian as a relative new-comers in their area of the Northwest Coast. MacDonald (1976) argued that Boas' conclusion was wrong. Johnson was unaware of MacDonald's argument. He describes Boas' method as a "seriation by inspection" 25 in which relationships between neighbouring ethnic groups are compared to determine which ones were "dissimilar to their neighbours and hence, by inference, recent arrivals in their area of occurrence." Johnson (1972: 315) also notes "Boas' crude treatment of his frequency data - he failed to standardize his element counts correctly - and his unsystematic use of the matrix i t s e l f . " In 1899 Flinders Petrie used principles of seriation in a matrix to establish a chronology of Egyptian material (Petrie, 1899). As early as 1908, S.A. Barrett (1908) used similarity scores to organize linguistic data. Nevertheless, quantitatively based studies have only begun to achieve wide interest in perhaps the last fifteen years, coincident with the rise of interest in systems theory and also coincident with a tremendous amplification in the capability of computers. Sheer drudgery is involved in manual mathematical computations; the information yield may prove to be quite out of proportion to the effort. The advent of the computer offers the possibility of reversing this situation. But with the opportunity also comes a number of problems. Appropriate choices of study, data base, and mathematical technique are by no means automatic. Nor is the philosophic approach to such studies unvarying. Ten years ago a conference called "Computers and Their Po^ tential Applications in Museums" was held at the Metropolitan Museum of Art. The published proceedings (Metropolitan Museum of Art, 1968) indicated that there are many exciting p o s s i b i l i t i e s . The areas which appear to have received the greatest attention in the inter-vening time are those of cataloguing and the establishment of data 27. banks and networks—basic necessities for large scale studies. Application of computer technology to the elucidation of problems has not made equal progress in applications to modest, but complicated investigations. More than twenty years ago, Kroeber (1956: 330-331) said: I continue to believe that s t a t i s t i c a l correlations can be used in s t y l i s t i c analysis, but that their main function is to convince those who are, by nature or inexperience, s t y l i s t i c a l l y insensitive. S t y l i s t i c qualities and their patterns and inter-relations must be taken in through the senses and digested through the subrational process sometimes called intuition, but for which the term "perception with aesthetic feeling" is adequate. A mathematical approach, being abstract and rational, seems best deferred u n t i l the pioneering job of analysis has been pretty thoroughly done by exercise of perception. On the other hand, f i n a l proof, in the sense of formal s c i e n t i f i c proof, has apparently to be quantitative. I believe that most qualitative data and perhaps even s t y l i s t i c ones, can also be interpreted quantitatively in the end; but certainly not so profitably at the beginning of inquiry. The approach of this thesis differs sharply in holding that the quantitative analysis should come f i r s t following in this regard David Clarke, whose f i e l d is archaeology. Speaking of material culture systems, he says: The intuitive and arbitrary "spotting" of key attributes in a system, before proper,analysis, leads to the arbitrary definition of group boundaries based on "type-fossils" which are typical. Such set markers may be ultimately established but the "typical attributes" or the key attributes must be the terminal result of the analysis - not the premise of the opening gambit (Clarke, 1973: 138). An example of the "type-fossil" approach is supplied by Burkitt, who advises archaeologists to: Group the artifacts into families, taking some obvious charact-i s t i c as a criterion of a family. In practice i t is found that the numberc-»of othese families is small, and: indeed-in these days they have become perhaps rather stereotyped, and i t is only rarely that a new and distinct family is recognized (Burkitt, 1963: 29). 28. To reject Kroeber's approach is not to deny the value of "perception with aesthetic feeling." The exercise of such perception in the analysis of large numbers of complicated artifacts may ignore significant structural relationships of a modest and subtle nature. There are many methods of achieving practical results. Kaufmann's (1969) study of Haida a r g i l l i t e carving was mentioned in the previous chapter. This is based on an extremely careful analysis of a combination of form and iconography, using seriation by frequency of occurence to provide a chronology of style. It provides a practical guide to the approximate dating of museum a r g i l l i t e artifacts from 1820 to 1910, and in addition offers an anthropological interpretation of the data. Kaufmann also isolates a number of "internals . . . which seem to have a l i f e of their own .[which] would correspond to Kubler's notion of an a r t i s t i c infrastructure " (Kaufmann, 1969: 154). Her conclusion i s that structure changes with symbol, at least in her data. Her basic method consisted of calculating the percen-tage occurence of selected characters and arranging these in a number of separate tables; radically different from numerical taxonomy, which calculates i t s results from the whole spectrum of characters or from an unbiased sample from such a spectrum (the term "character" w i l l be defined below). Numerical taxonomy is a method of classification. It proceeds on these assumptions: generalizations about the taxa cannot be made before one has recognized the taxa; that taxa cannot be recognized before resemblances between organisms lor whatever the discrete object of investigation happens to be; this i s usually referred to as an OTU, operational taxonomic unit] are known; and that these resemblances cannot be estimated before organisms and their characters have been examined (Sneath and Sokal, 1973: 5) 29. Estimation of resemblance begins with the choice of characters. A logically irreducible character is an attribute of the OTU which the investigator considers fundamental - and a basic subjective judgment must be made here - but there are c r i t e r i a for inadmissable characters. For example, meaningless characters for an estimate of resemblance between totem poles would include the negative numbers assigned to their photographs; logically correlated characters would include measurements of both the height of one segment of a pole, and the height of one half of one segment of a pole; tautological characters would include measurements of the height of a figure, and the t a l l -ness of a figure. In addition, tautology may be suspected i f the scores on two characters are found to be perfectly correlated; in this case one character should be rejected before computation. Invariant characters, while legitimately descriptive in themselves, do.not add information, and should be excluded. An example of this would be to use "roundness" as an attribute, or character, of totem poles, although this might be a logical and useful attribute for some other set of OTU's. It is obviously not possible to eliminate a l l personal bias from the basic character.selection, nor to determine absolutely the necessary number of characters. In their discussion of the requisite number of characters, Sneath and Sokal (1973) note that there is no generally valid answer. However, they also note that empirically, when large numbers of characters are measured, they obey an apparent principle of inertia; as more and more characters are added, i t takes an increasingly large number of such characters, with quite different 30. information, to alter the estimate of similarity. On the subject of the problem of different investigators defining identical characters in different ways, or making observations on the same data base using varying numbers of characters, Sneath and Sokal cite a study by Sokal and Rohlf (1970) in which of 17 common factors in a particular data base, 16 were found by an experienced taxonomist classifying the data on 119 characters; 14 common factors were found by the "least tradi-tional" taxonomist using 53 characters; and a l l 17 were found by an inexperienced person using 62 characters. Thus i t appears probable that the choice of at least 45, and no more than 100 characters by any given investigator w i l l yield reliable results. A recent study by F.R. Hodson (1971) is very pertinent to the problems both of analytical approach and choice of method in this thesis. First of a l l , i t i s addressed to a comparison of the results of intuitive classification and classification by hierarchical clustering. The data used were photographs of brooches recovered from Swiss grave sites of the La Tene culture, circa 250 B.C. Hodson had established a variable character l i s t by observation on the original material; this l i s t was scored and clustered. Then the photographs were submitted to a number of other professional archaeologists and an anatomist, who agreed to arrange the brooches chronologically by making intuitive judgments on the basis of style. The arrangement made by computer clustering was independently verified by stratigraphy and radio-carbon dating. There was, of course, no question of the computer providing any interpretation of the meaning of the clustering. Hodson noted: 31. Archaeologist 4 classified the brooches basically into a series of interrelated clusters. Archaeologist 3 and the Anatomist into a more or less linear development without discrete clusters, Archaeologists 1 and 2 into clusters that are themselves related in a linear fashion . . . Thus there are considerable differences in the general interpretation of the structure within the set. Specific interpretations of individual brooches cannot be discussed in detail, but i t may be noticed that some brooches are consistently allotted the same neighbours by most analysts (cf. brooch 10), while others are not consistently grouped at a l l (cf. brooch 5). Judged between themselves, then, i t may be said that despite a tendency to agree for some brooches, there is a surprising difference of opinion on others and on the general interpretation of the structure behind the group (Hodson, 1971: 650). In actual fact, the Anatomist's solution was closer to the true chronological arrangement than: that of any of the Archaeologists'; the separate solutions of two of the Archaeologists contained grave errors. While i t might be considered that this indicates nothing more than that some archaeologists should never be called upon to make decisions in matters of style, I believe i t highlights the assistance we can expect to receive from s t a t i s t i c a l techniques such as hierarchical clustering and multi-dimensional scaling. The particular techniques chosen for this study, then, have been demonstrated to have application in somewhat similar situations (Hodson, 1971; Matson and True, 1974; Matson and Lipe, 1978; Flynn, 1976). These techniques avoid the assumption of arbitrary key a t t r i -butes. Since each application of hierarchical clustering and multi-dimensional scaling is varied to f i t the particular case, the method used w i l l be described in detail'.-.' 32. There are a number of different ways to approach the computation of a measure of resemblance. On this occasion the method of choice is a coefficient of dissimilarity (distance), expressed as zero for identity and an undefined positive value for maximal distance, or disparity; in this case, Jaccard's coefficient was used. The un-standardized values are arranged in a matrix; hierarchical clustering of these proceeds by " f i r s t grouping points, objects, variables or observations which are 'closest' together to form a cluster, then com-bining these facets and repeating the procedure u n t i l there remains but one item or cluster " (Flynn, 1976: 30). Differences in methods of forming clusters are also important. Ward's error sum of squares was the method chosen. This determines the proximity of two clusters by forming new cluster groups and calculating the sum of the square of a l l distances from one point to another with the new cluster. The solution which minimizes this sum of squares is the one used to form new clusters. For a discussion and comparison of this and other methods of hierarchical clustering, see Flynn (1976). Hierarchical clustering provides a solution along one dimension. It is also possible to consider the data by means of a multi-dimensional scaling technique providing n number of dimensions in diminishing order of importance, which may be plotted graphically as a series of two dimensional diagrams. This technique has also been used for this study, by means of Torgerson's metric multi-dimensional scaling technique (Matson and True, 1974). For a discussion of the u t i l i t y of this, see Matson and Lipe (1978). 33. The photographs used for the observations were drawn, as indicated in the introduction, from various sources. Poles raised after 1925 were not included. Of the ninety-four poles of the Gitksan available for study, eighty were judged to be adequate for use on the basis of at least one f u l l length view and one detail being recorded. These poles were originally located at Kitwancool, Kitwanga, Kitse-gukla, Kispiox, and Hazelton. To them were added sixteen poles from the Niska, representing Gitlaxdamks, Gitwinsilk, Angede, and Gitiks, for a total of ninety-six poles in a l l . A character l i s t totalling seventy-one characters was esta-blished by making observations on the ninety-six poles. These characters are listed in Appendix I. Characters could be scored as present, absent, or no comparison. In order to provide information on the com-parative sizes of the poles, as well as to allow for the inclusion of information on the general type of figure occupying each position on the pole, the f i r s t forty^-eight characters represent twelve re-petitions of four choices - humanoid, animal, bird, other - which were scored beginning at the top of each pole. The term "animal" i n this case encompasses a l l land and sea creatures other than humanoid. A contentious problem was presented by the frequent presence of figures with composite attributes, e.g. a human with a large beak. While these have names in the literature (although the exact identity of any particular one is a matter for considerable dispute), the best solution in such a case seemed to be to score "present" on both "humanoid" and "bird" in the same position, thus differentiating them from both humans and birds. "Other" would be scored as present in 34. the case, for instance, of a figure regarded as a canoe - i f the canoe contained a human, "humanoid" would also be scored present for that position. Some figures absolutely defy characterization as humanoid, animal, or bird - they became "other" even though superficially they might appear to be a candidate for one of the other categories. If a pole had only six figures, the remaining six positions (of a potential twelve) were scored as "no comparison". A pole with a figure at the top, then a plain shaft, with two or three figures at the bottom, was assigned a number of empty "no comparison" spaces proportionate to the size of the figures at the bottom. The remaining characters chosen, numbers forty-nine to seventy-one, record aspects of the whole pole, as well as aspects of individual figures on the pole. No absolute measurements could be used, due to the varying scale of the photographs and the distorting effect of perspective. However some relative measurements were used, since proportion is an extremely important component of style. For example, the proportion of the humanoid eye orbit to the head was established in each case by means of dividers. For the second stage of the analysis, a second character l i s t was established, which dropped the characters pertaining to position on the pole, and added more which described the figures themselves. After scoring was completed, the results were transferred to computer punch cards in two forms: that suitable for "Q" type examination, and that suitable for "R" type examination. "Q" technique uses totem poles as OTU's and defines them in terms of the characters, thus grouping together poles with similar characters. "R" technique uses characters as OTU's, grouping them together into bundles of traits which tend to occur in association on poles. In both cases, whether using "Q" or "R" technique, the results of an hierarchical clustering w i l l appear in the form of a dendrogram. This is a re-presentation of similarity along one dimension. Units of closest similarity w i l l be paired, with other items added to the cluster in the manner described above. Well defined clusters w i l l create a "tree" shape. As the clusters become larger, the. level of similarity at which the units cluster, drops. The arrangement of clusters in a standard dendrogram is not significant along the vertical axis. What is important are the relationships of units within clusters and of small clusters within larger ones. The multi-dimensional scaling program extracts vectors (or factors) in order of importance. These vectors may then be graphed as co-ordinate points on a series of two-dimensional plots. The following operations, u t i l i z i n g these procedures, were undertaken. F i r s t , "Q" sort data using ninety-six OTU's, the poles, and seventy-one characters were clustered to determine whether meaning-f u l sub-populations which included figure placement would be formed; thus indicating whether the choice of figure placement was or was not random. Second, "Q" sort data using ninety-six OTU's and forty-seven characters, a l i s t from which figure placement had been removed, were clustered to determine whether meaningful sub-populations related to an artist or a r t i s t i c group would be formed. Third, "R" sort data using a l l the poles from the Nass, com-bined with poles from Kitwancool, and Kispiox, were clustered to establish the "bundles" of a r t i s t i c traits found in this area, where Barbeau had hypothesized a great deal of mingling on historic and geographiesgrounds. In each case these data sets were also submitted to multi-dimensional scaling. The results of these operations w i l l be discussed in the next chapter. CHAPTER 4 DISCUSSION If distinctive styles of totem poles can be identified, we would expect to find them clustered together in one or more of the following ways: geographically, chronologically, by schools of a r t i s t , or by individual a r t i s t . The result of an hierarchical clust-ering operation is a dendrogram, a graphic representation; of similarity along one dimension. Although shaped like a tree, i t should hot be read in terms of genetic relationships, but in terms of level of simi-l a r i t y between units. The f i r s t operation consisted of an hierarchical clustering of the poles on the basis of characters which are lis t e d in Appendix I. The resulting dendrogram (Appendix II) produced a number of well defined clusters. A key to the identity of the poles is in Tables 1 and 2; in the following discussion they w i l l be referred to by the numbers used in that key. TABLE 1 KEY TO THE IDENTITY OF GITKSAN TOTEM POLES Date Pole Location .(Barbeau, Carver Pole Carver Illustration Number '1929) (Barbeau, 1929) Clan Clan (Barbeau, 1929) 1 Kitwancool J1875 Sqateen (Nass) A C 111:1 2 Kitwancool 1875 [Nass] A ? 111:2 3 Kitwancool 1885 Nees-laranows A C 111:3 4 Kitwancool [1890] Nees-laranows A C IV :1 5 Kitwancool 1875 Pees A A IV: 2 37. 38. TABLE 1—Continued Date Pole Location (Barbeau, Carver Pole Carver I l l u s t r a t i Number 1929) (Barbeau, 1929) Clan Clan (Barbeau, 6 Kitwancool 1885 Nees-laranows A C IV: 3 7 Kitwancool [1916] Qaqhl A A IV: 4 8 Kitwancool 1865 Nass carver A5 ? XII: 3 9 Kitwancool 1880 Arhtseeprh (Nass) A 5 B XIII :1 10 Kitwancool 1900 Harhpegwawtu A 5 B XIII:2 11 Kitwancool 1860 Nass carver A 4 ? X:l 12 Kitwancool 1870 Hrtseeyae (Nass) A 4 D X:2 13 Kitwancool 1890 *Hlamee 4 A X:3 14 Kitwancool 1888 *Haesemhli-yawn? A XXV :1 15 Kitwancool 1860 Nass carver ? XXV: 2 16 Kitwancool 1875 *Haesemhli-yawn A XXV: 3 17 Kitwancool. 1895 Hlamee ~ _ ,* c 4 A XXIV:3 18 Kitwancool 1875 Haesemhli-yawn c A XX:2 19 Kitwancool 1885 Hlamee c . A XXII:1 20 Kitwancool 1905 Hlamee Co A XXIV:1 21 Kitwancool 1865 Haesemhli-yawn X 4 A XXII:6 22 Kitwancool [1870/80]*Nass carver? C 4 ? XXII:4 23 Kitwancool [1890] [Gitksan] ? 24 Kitwancool [1865] A ' ? 25 Kitwancool [1865] ? A 4 7 26 Kitwancool 1880 Sa'anrhkwanks (Nass) A B X:4 27 Kitwancool 1895" Weegyet A l B X:5 28 Kitwancool [1890] ? 7 XXII:5 29 Kitsegukla 1900 Weegyet A i B V:l 30 Kitsegukla 1895 Hlamee A 3 A V:2 31 Kitsegukla 1890' Kwaw'amats A 4 B XII :1 32 Kitsegukla 1885 Sqayaen At D XI: 6 33 Kitsegukla 1890 Hlamee A 2 A XI:7 34 Kitsegukla 1885 Hlamee A XVII:4 35 Kitsegukla 1880 William Nass B 3 B XVII:5 36 Kitsegukla 1875? Haesemhli-yawn B 3 A XX: 1 37 Kitsegukla 1890 Negutsrael A XXI :1 38 Kitsegukla 1895 Hlengwah B A XIV: 2 39 Kitsegukla 1910 Qaqhl A XIV: 3 40 Kitsegukla *1900 Geesarhkees A XVIII:4 41 Kitsegukla 1895 Kwawdzabarh B 2 A XVIII:5 42 Kitsegukla 1925 Wawralaw B C XIX :1 43 Kitsegukla 1895 Kwawdzabarh B A XIV: 4 44 Kitsegukla 1890 Tom Campbell B A XV:2 45 Kitwanga 1890 Gitrhawn A D VI:3 46 Kitwango 1900 Kwawdzabarh A A VI: 4 47 Kitwanga *1880 *Nass carver A 7 VII :1 48 Kitwanga *1890 Nass carver A 7 VII: 2 49 Kitwanga 1920 Harpegwawtu A B VII: 3 39. TABLE 1—Continued Date Pole Location (Barbeau, Carver Pole Carver Illustration Number 1929) (Barbeau, 1929) Clan Clan (Barbeau, 1929) 50 Kitwanga 1865 Nees'awaelp c j D XXVII:1 51 Kitwanga 1875 *Hlamee? c 1 A XXVII:2 52 Kitwanga 1850 ? D ? XXXIII:1 53 Kitwanga 1875 Negutsrael D A XXIX:3 54 Kitwanga 1870 Nass carver D 1 1:2 55 Kitwanga 1900 Negutsrael? D A - XXVIII:4 56 Kitwanga 1880 Geesarhkees D A XXVIII:2 57 Kitwanga 1914 Qaqhl Do A XXVIII:3 58 Kitwanga 1900 Tom Campbell A VII: 5 59 Kitwanga 1920 Harpegwawtu? 4 B VII: 6 60 Kitwanga 1907 ? A2 1 IX:1 61 Kispiox 1875 Haku cl A XXVI:4 62 Kispiox *1890 Haku B/C A XIX: 5 63 Kispiox 1870 Waws emlarha e A B IX: 4 64 Kispiox 1880 *Waws emlarhae B/C B XIX: 2 65 Kispiox 1855 Nass carver B 1 XIV: 5 66 Kispiox 1895 Tsugyet I* C XV: 3 67 Kispiox 1870 Waws emlar ha e B VI: 2 68 Kispiox 1915 ? ' A ? 69 Kispiox 1850 Qelran (Nass) , B C XVI: 2 70 Kispiox *1870 Naeqt B A XVI: 3 71 Kispiox 1900 Tsugyet B C XVI: 4 72 Kispiox 1855 Gurhnahaks C A XXII:2 73 Kispiox *1885 Tsugyet C l A XXIII:2 74 Kispiox 1855 Kwikihl'wans B l A XVI: 5 75 Kispiox 1875 Hakst B A XVII:2 76 Kispiox 1880 Raenem? C2 C XXI:4 77 Kispiox 1900 Raenem? A l C XIII:3 78 Kispiox 1892 Larhwilemhot, A 1 C V:3 79 Kispiox 1875 Waws emlarhae C ? XXII:3 96 Hazelton 1865 ? A 7 IX: 2 * I disagree with this attribution. 40. TABLE 2 KEY TO THE IDENTITY OF NISKA TOTEM POLES Date Pole Location (Barbeau , Carver Pole Carver Illustration Number 1964) (Barbeau, 1964) Clan Clan (Barb eau, 1964) 80 Gltlaxdamks 1875 Neeskyinwaet C C P- 440 81 Gitlaxdamks Hladerh D1 C NMC-70687-B 82 Gitlaxdamks 1860 Oyai C NMC-70687-B 83 Gitlaxdamks 1870 ? c 4 7 P- 450 84 Gitlaxdamks 1865 Arhtseeprh? A B P- 74 85 Gitwinsilk 1892 Weesaiks C B P- 237 86 Gitwinsilk 1892 Qaderh B C P- 438 87 Angede 1850 Oyai C C P- 227 88 Angede. 1860 Oyai A C P- 174 89 Angede 1850 7 D ? P- 48 90 Angede 1860 Oyai C C P- 228 91 Angede 1870 7 C 7 P- 229 92 Angede 1880 Weesaiks? C 7 P- 231 93 Gitiks 1870 Oyai D C P- 51 94 Gitiks 1870 Aqstaqhl D C P- 43 95 Gitiks 1870 Oyai D C P- 23 The f i r s t item of interest in the dendrogram is the strong tendency of Niska poles to cluster together, confirming that a substantive difference exists between the poles of the Niska and the Gitksan, the two divisions with which I am concerned. There is a close association between poles number 84, 81, 90, 88 and 94; and between poles number 83, 80, 91, and 85. This can be seen both in the dendrogram and in Table 3 below. The order reflected in the various sub-clusters was at f i r s t d i f f i c u l t to discern. Georgraphy certainly appears as a factor in the case of the Niska poles, as i t may be in other cases to be noted below. It does not seem to have any general significance for the UBC Computing Centre Programmer F O R T R A N C O D I N G F O R M Date Page. of. S T A T E M E N T F O R T R A N S T A T E M E N T IDENTIFICATION 1 2 3 4 5 6 7 8 9 10 i i 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 LOCRT-fQN POLE , -NUMBER, p>l R 1 > Hi Mf- N Bi) li <m Hn Ht) V Ha, m UL VFR. BL RA %. y_ \D Ho Ho IH V Ki TI */< :o 01 fS6 5 / /? \\- 1 5± RI f\ i fit i -IL I.ZLI MR N Ki Iff vc 0( )L 8 C 4- '/£ y . / 4-a / • i ) Itt J If in T l Ml iff" '/// QL Hi/ 'Jk-L Kt * — t 51 n \Q 5 B 9/ 5<J H$ T 6 6 ul Bi LSI Rl Hu w v Hu 1ft y Ho Iff/ y Bi v> 'JJ. RN JLLL IN % Hu tit Iff V % 7N 91-'US. Kt 5/ vc X /t J-7 5 3 5/ Ri t 7 5 Bi b m Ho b 8 Ri > r > IF i % K OI R % Wl bp >£R fin m h W Kt n IH N> 10 )L K •5 /? Z w % Bi RI Htt H(\ V Bi Rfl Hu k RP Hat M h 'EH B Ri ) Kt TK 'HI >t Oi >L n •0 r? ;c 1 H Oi HE, ? Ho iff V ft 'IN 9N mi Oi W R, 5,i ?/> Oi 'Ht. X Ht HI H Ht IN, IN Kt n 'HI ic oc >L n 5 /? d B, RL % f/ft i HA IHI }L flii IR R> tin fa mil Ki T\ Vfi t\> rf r It '0 /? If lA j-6 ', Hi ID Hu Ml- Hi M Ho Nil 'y 9H 'Hi JL Hti V % \J> *7 'HI ic 01 W, o /? ',H BI -Lid* Ho Ht ,f 1— 'ii — Hi it'/ Nl -J /Y % Li 1.1 Nt j— n RI\ IN 71 M-L Ki t—fD S i it c X/ 75 b 1 Bi> ID % IN U fit >lf1 RL .LSA Ho L-LL tit ±3— 'ti UJt JL±L Lull Li— if* LLM t.Jr~. Kt T * ^  JI 'li <?/ /< i i K . L/ '? 3 xs -t-U. EL 6 0 /?/ ) Hu Nt 'N 'It// 'i Bi h /HI 1 % Iff ?L HA IN, 91 Ht 'IN ffL Ki Sf 7C it 0 /) X' IE* 6 3 SV Ho "HH Y Hi 'tfti N Hi /it 'ii Ht NR y Ht, Ht, N % 'N, Ki 5 ?/< >X li 10 c -A /8 1 > 6 F>l Ri :h M fa W Bi RI Hu It N Hi RI. KI 'Hi ic oc 'I / JO 3 A ^ f/. 4-m h m Ri B \ Rt) RL A fi w m 'li flu P/> & J <EI{ KI n F(I VI ;i 7 / ?( S z ?/ r 4- O {to ttflNd lot fiH, 'a m Hi INI ?N BI Ri- KI rs EG u H /, kc ft < ?5~ 3 f) 3/ Rl Hu IH y m ii RI KL] Iff v< •H // >0 i) Kf\ Kl y s 5 Z Rt <iL Or Ml R. a R fit Ri Hi iff) ti KI rs Eb 01 ;L ? / it >0 1 x; )FC fit Ri ? Bt w % ttl HI. HI Ht 'HL Hi Nl IN Ki Ki Li n a Ft •ft, u Yi 0 A 3 '.ft x^ )£l d / B>r Ri ) Hi Itt ft? dt RJ ) RI iff 'Hi Bi HI } <L f) li M 15 /I + IN ,EL 3 z ) Or HB ft l & m b ti& Ht IN 8t u 'ffE R Hu ID n BI Rl HL 'MJ U Bi R. 0 f?A 'II1, HI D Ki n IRi oc L 3C 4- A 4- '-Hi t( >FL 1 3 & y >ii RJf > Hu Nt N Ki n ILC il V A / './? 8 Bl p< m ft a rfti % 0 m TR o- -Hi rft h m 'a OA Hi IH ?/y HA HI) Ht Ho Hi IN Ki IR 0 ?/ n ti r, 4.' !,/? ; / a 'J % Ht V Ji fit tu i/ti 3/ m Ht 'IN BL fa m 8i (If) HL HI Uh /?, T\ \ffj ti( ii — A x 1 f 5 — I — ~\ " | " -4~ TABLE 3—Continued 42. UBC Computing Centre Programmer _ F O R T R A N C O D I N G F O R M Date . Page of S T A T E M E N T c N O . c 0 N T F O R T R A N S T A T E M E N T • I D E N T I F I C A T I O N 1 ! 2 3 4 5 6 1 i 9 10 11 12 13 H 15 16 1? IS 19 20 21 22 23 25 26 2; 28 30 31 32 33 M 34 FFf, 35 9A 36 / V 0 36 76 39 e,\ 40 42 <3 44 45 46 4 7 48 LQCBT/ON DA TE Li W i f 1 NFIHPFZ ft m He Hi *k Wi Mi T£ U Hi v Hi r. K l i y 5 /=} 1 h EL 3 0 • Hi r ini X-9A / Kin ' 1 f A! A // 1 ' /? 1 t-. 11 Lf Xi EL -ft Hi ft, VP / Ht H, fA HL n H )F f/ A r. T l J \ L ft li // 4-Lh ?EL 3 3 > Hi l- V/t IP L ft // f/A F ft n '10 y< ll I. 5 C iff >8l 0 Si ; > 'HI R 0 W FR 0 r// c\ 0 T% a 'HA 7s ft; y/F kt fit X ib 5 ft LH (RL 1 HL 'A? 0 "ft ft V/i / < > 77/ 0 Oi ~R Ch WF- kt A/ 7"K 77 w \£ w 01 Ki 7L -/= — 'ft 0 5 (1 c -f IP IRL 3 HI 'ft fl Ho 1M b FH£ ?< m Hi V mf / (rl 4 3 II\ ft V o-rH> Fit fft 1 T Ki A /f 0 P» f ft fit <h Hi, M fAi Or, Si ? / » Iff, Hi •Mi vy T . if IN w 1 L /A \T 3. r ? > A Ft (. Hi i I • ? &4 t i * 'I! :.v HF R % m 4 £R H\ lit iff-? ft: Yfi y ^ / t r '< *' r 6 LA, • 0 ft n — ft an MA lit ft) R Hi f4 "iff H T 7/1 ft/ 111 ft* Iff / ft m HP L £1 )p if 0 c 'fit 9 ^1 ~ • • ~l I % Fit ?/y fit '/V m Hi IN / HI V/F (ft 1 Ht H ft/ V/t W /I Ki 71* fit IC 00 1 If ,l7 A 4 LP 'jfl Hi\ II? t ,n ! f f fit if*- its Kt r • it VI ri • /S / 8 0 0 IR AS •Ft r t — — Ht ft/ Hi w 1 ' /Ft • i J 4 /// % Kt r; m u 1 ft) t [A KIK f •MPS H HI HHft Him P i) V 4v, m X tfi Jh t ft. R, ) HL /// Ki Th T-m CL ft-li >o Al l_ R -- f8 0 B 5 LA EL ft> R 1-1 7/ -It Hit. » A % Hi m h *&} Hf tin ,'>•:? kt 5 ft Z J / J it? 3 • <\ / r ' - A • 1 'til--' c \ '\t. (1 u I T U 7 01 / Ht •H R y Ho Hp HF NJ % Hk IH -?. ? A / HL HL 'M Fit n He •V h Kt {1 D m p, '& )0 OL L 1L 'B 6 5 A 5 'J EL - r Hf a. y ?A Hi 'Ft 6 5 A 5 '/? EL I 5 — h ON s e m b, m % IH ?L h & Kt o7 >x 19 /. — > R Iff xs EL (0 8 T utsL uompuiing cen t r e Programmer F O R T R A N C O D I N G F O R M Date Page_ of. S T A T E M E N T F O R T R A N S T A T E M E N T IDENTIFICATION 1 2 3 4 5 6 7 8 9 JO n 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 .38 39 40 41 42 43 44 45 46 47 48 L 0CAT/DAf DBTF CI US Z ?_ CLF }N • POLE MJHRBR. Bl Hi \lh ) -Hi Rl :R m <n IH ex f? n » ft U t Hi 'Hi its ?A _ HL HL 'tit. H6 % (L % m 'Ht IEA 0 m Ri Kt If? R ft o P iff 5 6 '-'V _ Ul rl Ti 'ii -H (/? Iff It LA* 5 R "IT >-R Lid El 8 4-HA \l/HRt Bt Rj ) Hi Ht 'li HL 'Ht w r m m IR, :— Ho St n 9X OF Hi ,5 IS 6 p m IAL ft 1 r m 91 : % <m m M IH HA i HA fit IftL Hu m k '/?li LJJL LLJ. •JJ Hi IQ lYl /i 16 / * r TH.L. 'A rlsi. U 9 0 in JJL ' ff 7 lL ILL Ik J ISJJ -ij- -/-'Hi Hi Hi Ml Bi Rl Ha Ht IH Oi 'HF e Bi Rt) Ht ig pp p H \& Q Ft 'fi X' PL 8 8 1-1,1. m Bi '—Li, Hi -LI— eJJ— CV—{ % m k M BL 'M Hi Hk RJS m Hj 3i Iff, FL m •h w 9i r K S It uu. Vi 0 0 -LI 'J- m 9 4-Hi Hi Rl RA 'If, ft i Fli iff 'fl, Pi y/i YH / Kt TL IHI IC Oi ){_ ff I 0 f) )£L 5 0/ XL -LLL £<• '•Ui a H 1 L. i «. <-/-8 z <?/. >e. 9Sf 4 1 5i Rl 5i Hi & 2_ EL A 0 tft tHi its ft* m 'ft L r Ki 'Hi IC Oi )L It ti •5 c LR '/Ri r 1 8 U-fJ. J-L -LLL i—^» i.U* hk Hf\ w h Il/ltli c ?, R> lit) HI Rt III Iff l_ 0 rHi -ti Ri 'it L /?/ V/l IH • Ki TU 'Fll 'C 7C L /£ >B 5 1 i 'ff Xft //?/ i ? 'ft Hi IN ii / '•'J-1 lit W I. HA Itfi Jft/t l{ Hit 'fi, 9L Q, rH Ff > <r, TI * II /R 'Ot IL Ih 6 5 -J. l 'fi 'u/l LLIL IRL i 2 / * Llif L—It ' I.I LU\ JJJ1* III if, I £ L. d Vl IL. a m 1 fit )f/t 'ff I Ki XL 61 IH 'R li '£ \5 5 c 'A s 0 LLL Iff Mrs— 1 ™ f-J.l Rt ' ft Vff, IJ-'ft i Ht IIS iui iftH b w h '(If HI f ' n 'IC i X n l V If ,s c —A] ' R AS** IRL z -t-J Jx <\ rJ-L this -I— > Hu W h m M h H£ M W k int t, '</A 'fli HIII iftti h 1Ef\ Ki TI /Ft vc OC V ff 5 / IJ R-5 'ft is EL Q foe/ m fa m "•=1 m J Hi 7JA IL 9/S -4 fa tft b u. • k Lt (ft 1U KL g If 5 J.— -l—A R — u IH 3 t Rj > ffi fa v> -OS (Hi i-C Oi n >/ I—c h -fj C IRI 1 T Bi SJ lot C TM & m 2_ 1 3i, VI X >_ Hi HL UL HJ Hi ?AI -a m <I T~' >-£. 31 u K 'fi 11 n /) / 3 s us'. 3 5. K\ i ^ SJ /\ • • ~n. \5. B u fiS'i j s J.Li UBC Computing Centre Programmer . _ F O R T R A N C O O I N G F O R M Date Page. of. UBC computing cen t r e Programmer F O R T R A N C O D I N G F O R M Date Page. of. S T A T E M E N T F O R T R A N S T A T E M E N T IDENTIFICATION 1 2 3 4 5 • 6 7 8 9 10 l l 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 LOCBJWM. DflTF r I.USTPR C/BA 1 ' - P AlU OL m S£A a yj. IL UJ & m 11 <£ •n & & 'JL CL L a Hi V7. n /A LL •y h Hi If/ a i Hi i/// /fi / HL 'M ? A Kl Iff Y6 IS f 1 9 D 'H KS Klf\ 5 Q-l IJM ift 1^ C/-4 7/y Kt 7" Jl ?A q a $ 0 •t— p '.ff t$ KlK 5 5 <... dL HL 'n w TfFk 'H IN ft) \l Ho Ht fH -\-L Kt 1  1 H—L °l t-S-l 0) ( X — (fi 0 H, 11 >£ ?Si f '-A no jj- //_/. * J.U. L/JZ .LJ-R P \ /?/ y? p>, ' ff R r Si .a 19 fc ) o ft rl->5 fa ? 8, (M rH / U-t /\ J f J t il h Hf), A w LZ-< /? 'Ri • B, R, ) Ht 'If 'ft. 'IC JJht 'X /<* '9 0 Pit H, )ty. IV u ?, P>l &4 fi J 0 Y.L-fit iUJi R 0 Ri > p. 'AT. 0 -U— JJ-i t-t-i A PI 0) C 15 IS y/i IS. «J7 6 ? '—V d/ f?, •) Bi o f) Ki. K/ l£ Vl SI )X ' p 2, 5 2 1 R 'Zl, 4 2, •SI 1X1 J. •-</ Hi Ml H) 1/ Hi 9/ / ff (o 5 I— i 'Ft f Kh i 'JJ. 1J-, J-L A- J-L .tx. f L- , si-l. 11 & 'Ht -h H, y/ Ht If P> 'R 1) H If- iV rf l< P Klk 5 JHN R f) 0 ~H r. ? o El ? ,5 °/( 0 o rt: ft5) / r~ 1 J. '/y , 1 '.L~—< -i— ft, /> Hi >H v/ ^/ 7" <// 'At rfl '2 0 -H F) 5 9 r /1 / V. M I W W ht ?( \h Fi I Hi J-L Iff 91 V a./ 7— 5/ ?/( /-YJ ;X 0 0 Li— H IP X.< 1 Lm F.L 7 ? ~H. T Bi R, ) Hi IH 9t ( ~H Hi 'H ?i\ • HI *PI 91 ( Ht Y/t (ff, • r / Kt Ht T\ /Y rf » '0 o 3 H :R X5 FI 5 Ht >H, ?f\ 1 Bi —V-RI » Ht ih Hi V Hi. '/ft ft] I -f-f-, S-7 7 Pf 1) -—1~ /f )t '5 3 \G BS f Hi RI Ht >M Q! / Ht IH 9i\ 1 El T. i ff li ff 'Ft Ft J-L — •i-L ./-/ /—I- f-i- 4— — ___ 46. entire set of poles. Chronology and art style do not show any significant patterns. There i s , however, a meaningful association of poles by clans, and by arrangement of figures. Table 3 presents the poles in the order in which they are clustered by this dendrogram. The table includes information on date, clan a f f i l i a t i o n , location, and position of the various figures on the poles. From the l e f t hand margin, the l i s t of figures for each pole begins at the top of the pole, and moves downward on the pole as the l i s t moves to the right. Lines represent blank spaces on the pole or i t s termination at less than twelve figure positions. Only the f i r s t twelve positions (forty-eight characters), those specifically concerned with types of figures, are represented although seventy-one characters contributed to the clustering. Barbeau's informants distinguished a series of sub-clans to which Barbeau attached descriptive names. For convenience of discussion they are simply referred to by letter: A and i t s subdivisions for Laxsel, B and i t s subdivisions for Gisgast, C and i t s subdivisions for Laxkibu, and D for Laxskik. While the character l i s t had twelve positions related to clan crest figures, the figures were not specific. The literature, both published and unpublished, states that very specialized crests were in the possession of sub-clans (Halpin, 1973; Barbeau/Beynon f i e l d notes). At the same time i t is assumed that these specialized crests developed from a stock of 'generalized crests common to the main clans. At a very superficial level, i f the order reflected clan a f f i l i a t i o n i t might be expected that there would be four clusters, simply representing 47. for instance, the importance of animals to the Laxkibu, or that of birds to the Laxskik. Rather than four major clusters the dendrogram is s p l i t at a very low level of similarity into two large clusters. Within, these clusters there are other clusters which are related quite closely internally. The f i r s t of the two major clusters contains nine sub-clusters. There is a preponderance of Laxsel (A) poles, with a tendency to cluster with Laxskik (D) poles. Two of the sub-clusters are composed exclusively of Gisgast (B) and Laxkibu (C) poles. Sub-clusters one, two, and three are dominated by separate villages. Sub-cluster nine is heavily dominated by Niska poles. The second major cluster contains sub-clusters ten to twenty. This major cluster is composed very largely of Gisgast (B) and Laxkibu (C) poles, clustering together. Sub-cluster fifteen brings together a number of Niska poles with a group from Kltwancool. Other sub-clusters do not display any definite geographical trend. The potential significance of the clustering by clan and village w i l l be discussed in the concluding chapter. The same data set was also submitted to multi-dimensional scaling, using Torgerson's metric programme. 89.84 percent of trace was extracted by 12 roots. The f i r s t two vectors are illustrated (Fig. 1) and discussed. Once §gain we find a strong tendency for Niska poles to cluster together, possibly in conjunction with a chronological trend. Directly in the centre there is a group of poles pre-dating 1875, from Kitwan-cool, Gitlaxdamks, and Kitwanga, which are of A and D clan a f f i l i a t i o n : 4 a >6S .63 i 0.O •53 "60 *4>: »43 I I ' • to 6f 4« 8£ 4* •35 4SK •8a .80 • 3 6 •90. 1 4* »4fc -4a. •S4 »34 i—i o 0 0 49. numbers 1, 8, 81, 11, and 52. In the lower right hand quadrant there is a looser cluster, also pre-1875, but with the addition of two poles dated at 1880: a l l of these are also of A and D clan a f f i l i a t i o n , numbers 93, 9, 56, 84, 54, and 24, from Kitwancool, Gitlaxdamks, Gitiks and Kitwanga. In the upper right hand quadrant is the previously mentioned group of Niska poles: numbers 85, 87, 88, 89, and 94. Across the top, in both upper quadrants, is a scattered cluster of pre-1875 poles of the B and C clans, numbers 50, 70, 69, 65, and 72, from Kitwanga and Kispiox. Scattered around a l l these clusters, but definitely trending toward the l e f t , are poles of increasingly later date. I believe we may infer from this that there were at least four "nuclei" of related early poles. The large number of outliers on the plot, of later date, also suggests a trend toward loosening of tradi-tion. To sum up, the results of an analysis of the ninety-six poles in terms of seventy-one characters defining pole organization, as well as a r t i s t i c t r a i t s , were the confirmation of a distinct difference between the Niska and Gitksan poles; a positive suggestion that the content and organization of the poles reflects an older two-fold clan division of which more below; and a confirmation of the close association of some of the poles to which Barbeau assigns early dates. Following this general investigation, the ninety-six poles were submitted to hierarchical clustering in terms of a character l i s t from which the f i r s t forty-eight characters, those-dealing with pole positions of figures, had been removed. The remaining characters were re-numbered one to twenty-three, and others describing further a r t i s t i c 50. traits as well as arm and leg positions of the figures were added to make a new character l i s t totalling forty-seven characters (Appendix III). This l i s t was used to re-score the ninety-six poles, and to produce both a dendrogram and multi-dimensional scaling plots. The dendrogram obtained from clustering this data set had six major clusters, (Appendix IV). The clusters are illustrated below in tabular form. The underlying organization reflected i s both chrono-logical and a r t i s t i c , and in most cases the methodology substantiates Barbeau's dates, by tending to cluster poles of the same decade. These dates were obtained from a variety of informants. When alternate dates were given, Barbeau simply chose between them by assessing one informant as more reliable than the other, although he noted the con-f l i c t i n g opinions. At other times, the dates represent Barbeau's guesses. Generally, the dates are rounded off by decade, and should not be considered absolute. If the clustering i s accepted as valid, a number of poles which Barbeau was unable to date now may be dated by association. Also we'may use thev>clustering^ to^compare .the rates at which clans were raising poles, and by inference have some idea of the power and wealth of the clans in relation to one another. Thus, i f we consider the poles of each clan for each decade by percentage, i t is obvious that in clan D power was concentrated before 1875; in clan C in the decades before 1885; clan A shows a fluctuating rate u n t i l 1910, and clan B, although much smaller, also was able to main-tain i t s efforts u n t i l 1910 (Table 4). 51. TABLE 4 PERCENTAGES OF POLES RAISED BY CLANS, BY DECADE Up to but not % of A % of B % of C % of D Total % Including Poles Poles Poles Poles of a l l Poles 1850-1875 21% 19% 39% 50% 28% 1875-1885 13% 19% 34% 25% 21% 1885-1895 35% 23% 26% 8% 26% 1895-1910 18% 28% 8% 8% 16% 1910- 10% 9% 8% 7% The most important aspect of the clustering of poles from this dendrogram in Appendix IV is the insight i t gives into the styles of the Niska and Gitksan carvers. In Table 5 below, the poles are shown in the way they clustered, with dotted lines between significant sub^clusters. A solid line separates.the major clusters. TABLE 5 NISKA AND GITKSAN POLES CLUSTERED BY CARVING CHARACTERISTICS Pole # Location Date Carver Pole Clan Cluster 1 1 Kitwancool 1875 Sqateen (Nass) A 84 Gitlaxdamks 1865 Artseeprh (Nass) A 2 Kitwancool 1875 Nass A 66 Kispiox 1895 Tsugyet B 6 Kitwancool 1885 Neeslaranows A 93 Gitiks 1870 Oyai (Nass) D 3 Kitwancool 1885 Neeslaranows A 22 Kitwancool [1870] Nass c 4 54 Kitwanga 1870 Nass D 56 Kitwanga 1880 Geesarhkees D 92 Angede 1880 ? C 52. TABLE 5—Continued Pole # Location Date Carver Pole Clan 7 Kitwancool 1916 Qaqhl A 26 Kitwancool 1875? Sa'anrhkwanks (Nass) A 4 12 Kitwancool 1870 Hrtseeyae (Nass) A 4 46 Kitwanga 1900 Kwawdzabarh A 38 Kitsegukla 1895 Hlengwah B 45 Kitwanga 1888 Gitrhawn A 61 Kispiox 1875 Haku c 2 Cluster 2 9 Kitwancool 1880 Artseeprh (Nass) A 5 18 Kitwancool 1875 Haesemhli-yawn C 47 Kitwanga 1860 Nass A 43 Kitsegukla 1895 Kwawdzaharh B 10 Kitwancool 1900 Harhpegwawtu A 5 33 Kitsegukla 1890 Hlamee A 4 23 Kitwancool [1890] [Gitksan] c 4? 44 Kitsegukla 1890 Tom Campbell B 24 Kitwancool [1865] A 5 32 Kitsegukla 1885 Sqayaen A 4 A 3 31 Kitsegukla 1890 Kwaw'amats 68 Kispiox [1915] jKwaw'amats]? A 29 Kitsegukla 1900 Weegyet A 1 34 Kitsegukla 1880 Hlamee B 2 67 Kispiox 1870 Wawsemlarhae A 1 79 Kispiox 1875 Wawsemlarhae C Cluster 3 4 Kitwancool 11890]? Neeslaranows A 48 Kitwanga *1875 Nass A 89 Angede *1850 D 13 Kitwancool 1890 *Hlamee A 4 53. TABLE 5--Continued Pole # Location Date Carver Pole Clan 78 Kispiox 1892 Larhwilemhot (Nass) A 1 27 Kitwancool 1895 Weegyet A 4 60 Kitwanga 1907 ? A 3 39 Kitsegukla 1910 Qaqhl B 49 Kitwanga 1920 Harhpegwawtu A 20 Kitwancool 1905 Hlamee C 71 Kispiox 1900 Tsugyet B 37 Kitsegukla 1885 Negutsrael B 3 55 Kitwanga 1900 Negutsrael? D 62 Kispiox 1880 Haku B/C Cluster 8 4 Kitwancool 1865 Nass A 5 25 Kitwancool [1865] A 5 16 Kitwancool 1875 Haesemhli-yawn c 4 17 Kitwancool 1900 Hlamee c 4 14 Kitwancool 1888 Haesemhli-yawn c 4 15 Kitwancool 1860 Nass c 4 30 Kitsegukla 1895 Hlamee A 1 36 Kitsegukla 1875 Haesemhli-yawn B 3 91 Angede 1870 7 C 35 Kitsegukla 1880 William Nass B 2 87 Angede 1850 Oyai C 95 Gitiks 1870 Oyai D 80 Gitlaxdamks 1875 Neeskyinwaet C 83 Gitlaxdamks 1870 Artseeprh c 4 85 • Gitwinsilk 1892 Weesaiks c 86 Gitwinsilk 1892 Qaderh? B Cluster- 5  — •• - -11 Kitwancool 1860 Nass A 4 72 Kispiox 1855 Gurhnahaks C 73 Kispiox *1885 Tsugyet? C 54. TABLE 5—Continued Pole # Location Date Carver Pole Clan 63 Kispiox 1870 Wawsemlarhae A 2 96 Hazelton 1865 7 A 76 Kispiox 1880 Raenem? C 58 Kitwanga 1900 Tom Campbell A 3 59 Kitwanga 1920 Harpegwawtu A 3 77 Kispiox 1895 Raenem? A 2 28 Kitwancool 7 c 1 51 Kitwanga 1875 *Hlamee c 1 88 Angede 1860 Oyai A 90 Angede 1860 Oyai C 74 Kispiox 1855 Kwikihl1wans B 1 81 Gitlaxdamks 1860 Oyai D 19 Kitwancool 1885 Hlamee C 53 Kitwanga 1875 Negutsrael D 94 Git: iks 1870 Aqstaqhl D 21 Kitwancool 1865 Haesemhliyawn c 3 40 Kitsegukla 1885 Geesarkhees B 2 52 Kitwanga 1850 7 D •Cluster 5 6 Kitwancool 1875 Pees A 41 Kitsegukla 1895 Kwawdzabarh B 2 64 Kispiox 1875 *Wawsemlarhae B/C 65 Kispiox 1855 Nass B 82 Gitlaxdamks 1855 Hladerh D 1 75 Kispiox 1875 Hakst B 1 42 Kitsegukla 1925 Wawralaw B 2 57 Kitwanga 1914 Qaqhl D 50 Kitwanga 1865 Nees1 awa elp c 1 69 Kispiox 1850 Qelran (Nass) B 70 Kispiox 1882 Naeqt B * I disagree with this assignment. 55. Within the f i r s t major cluster shown on Table 5 there are two related sub-clusters, from pole number 1 to pole number! 92 inclusive. These poles generally share the characteristics of wide noses, wide mouths, elongated heads, and deep r e l i e f carving. The distribution of the poles is mainly on the Nass and at Kitwancool; the majority of the carvers are said to be Niska. With only two exceptions the carvers belong to the C (Laxkibu) clan. While the poles themselves are not organized in any distinctive manner, the execution of detail by the carvers certainly shows a shared view of a method of rendering. In addition to the associated traits mentioned above, the majority of these poles show a bold roundness and an over-all harmony in the relationship of the segments. I believe this may be called a style in the sense of the shared manner of carving. On the grounds of the geographical location of the poles, and their attribution mainly to Niska carvers I have called this style Niska I. This series is illustrated on Plates I through VI. The last sub-cluster within the f i r s t main cluster of Table 5, poles number 7 through 61 constitute another group sharing a distinctive concept. This has to do with the organization of the total pole. In this case, although details of rendering d i f f e r , the concept of the pole surface as a continuous one to which raised figures are applied as separate motifs quite discrete from one another creates a unifying theme. The distribution of the poles is exclusively on the Skeena and the majority of the carvers are Gitksan, s p l i t evenly between A (Laxsel) and D (Laxskik) clans. It should be noted that the two 56. earliest poles are at Kitwaneool and are attributed to Niska carvers. These poles may represent a proto-type. I believe this manner of organizing the poles deserves to be called a distinctive style. On the grounds of distribution by location, and of a majority of Gitksan carvers, I have called this Gitksan I. This series i s illustrated on Plates VI through IX. The second major cluster, poles 9 through 79, represents yet another distinct idea of pole organization. Shafts are lightly incised or uncarved, while at the bottom figures whose outer surface continues the same general plane as the shaft maintain the i l l u s i o n of a continuous surface. The geographical distribution of the poles is entirely on the Skeena, with emphasis on Kitsegukla. The majority of carvers are Gitksan, representing clans A (Laxsel), B (Gisgast), and D (Laxskik). There is a considerable range in time represented, although one group, poles 10, 33, 23, and 44 are in the .same decade. I have called this.Gitksan.il; i t ,is illustrated on Plates X through XV. The next large cluster, poles 4 to 62 inclusive, do not represent any distinct style. They have a wide variety of character-i s t i c s and are generally late in date. Poles number 48 and 89 I believe to be incorrectly dated by Barbeau (1929: 47; 1964: 49). In neither case is the ethnographic data very firm. Aesthetically they both belong with the later group in which clustering has placed them. Particularly notable is-the weak execution of the arms in the lowest figure on pole 89. Generally the poles are poorly proportioned and haphazard in organization, in my estimation. They are illustrated 57. on Plates XI through XIX. In the fourth major cluster poles from Kitwancool and the Nass predominate. This cluster also contains the last of the four distinctive styles I believe have been demonstrated by this dendro-gram. Characteristic of these poles are strongly marked horizontal divisions, with frequent alternations of large and small figures. The details of execution differ widely. A long time span is represented, from 1850 to 1900. Although the geographical distribution of the poles is f a i r l y evenly s p l i t between the two areas, I have considered this fact to be outweighed by the larger number of Niska carvers; therefore I have called this style Niska II. It is illustrated on Plates XX through XXIV. Major clusters five and six are illustrated on the remaining Plates XXV through XXXVII. There are no distinctive styles, but some of their features w i l l be mentioned in the concluding chapter. It should be emphasized that none of the styles discussed above exist as a series demonstrating a l l the features mentioned, and only those features. On the contrary, I am talking about tendencies which are marked in certain: .cases and weak in others, but which nevertheless show consistency over a range of poles. The multi-dimensional scaling plot from this data set (Appendix V ) did not add any new information. 82.60cofLtrace was extracted by 12 roots. There is further confirmation of a difference in tradition between the Niska and the Gitksan, with a clustering of Niska poles in the upper l e f t quadrant. Also noticeable in the lower 58. r i g h t quadrant i s a loose c l u s t e r of poles, l a t e i n time and poor i n execution. The l a s t operation undertaken was an R technique c l u s t e r i n g and s c a l i n g of the 47 characters as seen on 63 poles; those from the Nass, from Kitwancool, and from Kispiox. The dendrogram (Appendix VI ) has two major c l u s t e r s within which are a number of w e l l marked sub-clusters. The numbers used i n thesf611owing<.discussion are those of the characters l i s t e d i n Appendix I I I . The o r i g i n a l data were examined to determine whether the sub-clusters were s i g n i f i c a n t i n terms of geography. The r e s u l t s are l i s t e d i n Table 6 below i n the order of c l u s t e r i n g . Locations l i s t e d are those i n which the t r a i t associations were preferred. Lines separate the c l u s t e r s . TABLE 6 PREFERRED CHARACTER TRAITS, BY LOCATION Character traits Location 1. Pole figures c l e a r l y divided h o r i z o n t a l l y Kitwancool 36. Arms i n "holding" p o s i t i o n at chest Kitwancool 13. Pole segments f u l l y modelled Kitwancool 31. Bird has human limbs Kitwancool 42. Knock-kneed legs Kitwancool 47. Arms hanging, s l i g h t l y bent Kitwancool 3. Discrete figures within pole l i n e Kitwancool 6. Orbit more than 33 1/3% of the head Kispiox 19. E y e l i d double plane change i n carving Kispiox 26. Prow chin Kispiox 4. Intertwined figures Nass/Kitwancool 41. Legs "hugging" the pole Nass/Kitwancool TABLE 6 Continued 59. Character traits Location 2. Series of near identical figures 39. Arms at waist, dexter over sinister 32. Straight hanging arms 40. Straight legs 3 8. Arms horizontal, sinister over dexter General Kitwancool/Nass Kitwancool/Nass Kitwancool/Nass Kitwancool/Nass 10. Lips delineated by incised line 17. Eyelid line incised 23. More than one set of segments matching 35. Arms in ''holding" position at waist 46. Figure kneeling Kitwancool Kitwancool Kitwancool Nass/Kitwancool Nass/Kitwancool 15. Pupil mainl^ebei'ow line, eyelid tip to tip Kitwancool 34. Arms akimbo, resting on knees 22. More than two segments matching 45. Knees drawn up and doubled back 33. Arms clasp drawn up knees Kitwancool Kitwancool Kitwancool Kispiox 5. Orbit less than 33 1/3% of head 28. Equilateral nose 27. Mouth and chin in continuous plane 29. Wide mouth 11. Humanoids with teeth 21. Only two segments same height 43. Knees drawn up, kneecaps turned to touch Kitwancool Kitwancool General General Kitwancool/Nass Kitwancool/Nass Kitwancool/Nass 8. Semi-angular carving junctions 14. Pupil mainly above line, eyelid tip to tip 37. Arms facing out, palms out, at chest 9. Figures hold smaller figures 18. Eyelid single plane change in carving 20. Eye circular Nass/Kitwancool Nass/Kitwancool Nass/Kitwancool Nass/Kitwancool Nas s/Ki twanc 001 Nass/Kitwancool 60. TABLE 6 Continued Character traits Location 7. Carving junctions angular Nass/Kispiox 12. Discrete figures modelled beyond pole line Nass/Kispiox 24. Eyebrows nearly meet Kispiox 30. Birds have plain wings Kispiox 16. Pupil perimeter congruent with eyelid Kispiox/Kitwancool 25. Crisp eyesocket Kispiox/Kitwancool 44. Knees in sitting position Kispiox/Kitwancool Multi-dimensional scaling of this data set did not produce an interpretable plot. . u• This chapter has indicated by means of a dendrogram (Appendix II and Table 3) that a clustering by clan reflects an earlier state of clan a f f i l i a t i o n , and that a rhythm and order exists in the placement of figures on poles. I have also indicated, by means of a second dendrogram (Appendix IV and Table 5) that four distinctive styles exist within the totem poles studied. In addition a dendrogram displaying association of character traits with location was presented. Further implications of the data above w i l l be considered in the next chapter. CHAPTER 5 .CONCLUSIONS The object of choosing a quantitative methodology for this thesis i s to provide a framework for consideration of the art of the Niska and Gitksan totem pole carvers. It i s intended to act as a base for subsequent intuitive observations and as a complement to the existing ethnographic record. It is also expected to provide new angles of vision, much as the turn of the hand produces a new pattern from the same bits of glass in a kaleidoscope. When the poles were considered in the f i r s t instance by means of both content and execution, there was strong confirmation of the assumption that a recognizable difference exists between the poles of the Niska and the Gitksan, as demonstrated by their positions in clusters in the dendrogram. Unexpectedly, the clustering also provides strong support for the assumption that the Tsimshian clan system was once two-fold rather than four-fold (Emmons, n.d.; Halpin, 1973). The high incidence of the conjunction of Gisgast (B) and Laxkibu (C) poles reflects their exogamic identification with the Haida Ravens and Tlingit Wolves, while the equally striking conjunction of Laxsel (A) poles with Laxskik (D) reflects their identification with the opposing Tlingit Raven, Haida Eagle division. Since the character l i s t used was so generalized that a "bear mother" and a "frog woman" would both have been scored as "human/other", and since the same situation i s true of other obvious 61. 62. sources of separation, we must suspect more complicated factors at work. While further investigation of these clan relationships suggested by the dendrogram is outside the. scope of this,thesis, I believe the data could be used as the basis of such an inquiry. Even more striking are the inter-relationships of figure placement demonstrated in Table 3. The f i r s t three clusters, for instance, show an association between the bird at the top of the pole and the figure in the f ourth-position; at Kitwancool, Kispiox, and Kitwanga (Clusters 1 and 2), this figure i s a humanoid. At Kitsegukla i t tends to be another bird. The series of poles 82, 89, 10, 16, 15, 86, 11, 83, 80, 91, and 85 is overwhelmingly humanoid, or alternates humanoid and animal. This leads to the inference that the rhythm of placement is one of the contributors to similarities between the poles of the Nass and Kitwancool. It also may be part of the reason for their separation from other Gitksan poles. I believe that these arrangements cannot be random. They deserve further investigation. The distribution of the poles on the multi-dimensional scaling plot for this data set underlines the cohesiveness of the Niska poles, and in some measure the strength of association between Niska carvers and the poles of Kitwancool. As well, the distribution suggests several different centres of s t y l i s t i c association. The dendrogram obtained from the clustering of the poles using the 47 character l i s t was discussed by means of Table 5 in the previous chapter. A series of common themes, elements of which naturally tend to overlap in such a complicated data set, were 63. identified. These themes suggest the following conclusions. Four of the themes are sufficiently distinct to j u s t i f y calling them sub-populations constituting a style. In other words, I believe they represent manners of carving, in the case of Niska I, or concepts of total organization of the pole, in the case of Niska II, and Gitksan I and II. Poles demonstrating these styles may be found outside the clustered series. Reference to the plates il l u s t r a t i n g Table 5 should make this visually clear. The f i r s t style, Niska I, is characterized by large scale figures whose heads are higher than they are wide. They generally have wide noses and wide mouths. Many have a prow-like chin. Their carving r e l i e f is rounded, sometimes angular, but deep and bold in proportion. Each segment of the pole is self-contained, giving an impression of horizontal divisions. The poles represent a Niska style, generally prior to 1880. The second style, Gitksan I, features a very distinctive concept of using a continuous plane as a unifying background for the figures. At f i r s t glance this appears a very simple idea, but in actual technical execution, of course, i t is not so simple. Generally the carving of the figures i s in low re l i e f and softly rounded; but high r e l i e f and low f l a t r e l i e f also occur. The details of the figures vary widely. There is also a wide variation in time, from 1870 to 1916. The third style, Gitksan II, has another distinctive concept of the whole pole. At least one third, and frequently more of the upper portion of the pole is plain, while at the bottom the figures are carved so that their outer surface s t i l l continues the same plane as that of the uncarved shaft. This made i t necessary to cut the pole 64. horizontally just above the uppermost figure, giving the figure the appearance of "sinking" into the pole. The r e l i e f carving tends to be f l a t and low, sometimes with rounded edges, sometimes sharp. Heads tend to be elongated. Other details vary widely. None of the carvers are Niska. This style occurs mainly at Kitsegukla, although there is one early pole at Kitwancool, and two others—possibly by the same a r t i s t — a t Kispiox. There i s another cluster of poles distinguished by the . computer as already noted, which display a variety of characteristics, many combined in an aesthetically unpleasing way. They are late in time, 1890-1925, and may be f a i r l y considered to represent a decline in tradition. They do not constitute a style. The fourth style, Niska II, is the one which Wingert assumed to be "Tsimshian style". The poles are divided horizontally; large and small figures of the same relative scale alternate and the scale is heavy. Many figures havehproBr-chins.-.arid "Tlingit" l i p s v ( f o r an example of this l i p formation see Plate XXVI: 2). Birds frequently have smoothly carved ovoids on their wings. Most of the animals have extremely large ears, often containing figures. There is a wide range of dates from 1865 to 1895. Nine of the poles represented are by Niska carvers. This is a Niska style. The next group is extremely interesting. It contains quite a number of poles with early dates. If i t has an underlying theme i t is that of varying combinations of small eyes, noses, and mouths. Much more important, however, is the fact that i t contains some very 65. i n d i v i d u a l poles whose components l a t e r occur i n other contexts. Numbers 72 and 73, for instance, have fi g u r e s at the bottom of an uncarved shaft, from which they are* separated 7:by a sharp, cut.-The f i g u r e s , however, are not contained within the plane l i n e of the upper shaft. One of these poles also has a hole surrounded by four small f i g u r e s . This motif recurs on pole number 21 at Kitwancool, the famous "Hole i n the Sky" pole, but as part of the d i s t i n c t i v e s t y l e I have c a l l e d Gitksan I. Pole number 96 at Hazelton, which Barbeau dates at 1865, displays an astonishing range of motifs: prow-chin; T l i n g i t l i p ; Kispiox brow (for an example of t h i s d i s t i n c t i v e sweeping brow formation see Plate XXXVII:1); and i t c e r t a i n l y appears to sink into i t s shaft .Although I consider that i t i s l a t e rather than early,.'if i t i s early i t i s a remarkable production. The Niska poles i n t h i s group share the heavy scale and decorated ears noted for other poles, but the figures are heavily intertwined. In t o t a l , t h i s group does not represent a s t y l e . The l a s t group contains some c l u s t e r s of e a r l y poles, and one as l a t e as 1925. If they can be said to have a common theme i t i s that of broad heads, wider than they are high, with large eyes. Five of the eleven poles are from Kispiox, and four of these have the Kispiox brow. Although i t i s tempting to consider t h i s a s t y l e , I do not believe there are enough associated features to merit that term. The group does r a i s e some i n t e r e s t i n g questions, as noted below. 66. There are in a l l nine poles in the data set of 96 which definitely exhibit the Kispiox brow. Seven of these poles are actually in Kispiox, which is what led me to apply this particular label when the trait came to my attention. These poles are not attributed to the same art i s t by Barbeau, nor do I believe they are by the same person. Nor are they a l l attributed to Kispiox carvers: two are Kispiox, two Nass, one Quldo, two are dubious. Poles number 22 and 67 are part of Niska I and Gitksan II, respectively, on other grounds. The other seven, however, are in the last two rather indeterminate groups. Is i t possible that they represent a proto-type of a style? If so, i t raises the further question of whether such a development began with the inspiration of a single pole? Pole number 69 was carved in 1850, according to Barbeau, making i t one of the earliest extant, and perhaps a source for other carvers. Only a very detailed study of the nine poles, with the accompanying ethnography, could answer these questions. Since totem poles have the most secure provenience, as to location, that could be desired, they are ideal for the purpose of determining whether a style is inspired by v i s i t i n g a r t i s t s , or whether the visitors absorb and use what they see before them. The R technique dendrogram as noted in the previous chapter did produce some bundles of associated t r a i t s . I am not in a position to say these associated traits are preferred by any a r t i s t , or group of a r t i s t s : I can say that they occur more frequently in some areas than in others. Generally, they confirm Barbeau's hypothesis of mingling in this area. 67. The study has defined a number of styles of poles, but the artists remain hidden. None of the dendrograms or multi-dimensional scaling plots produced close clustering of a l l poles attributed to any one a r t i s t . There are a number of artist s , according to Barbeau's attributions, who are well represented in the data. Poles number 14, 16, 18, 21, and 36 are attributed to Haesemhliyawn, a Laxsel carver from Kitwancool who is said to have worked between 1840 and 1880 (Barbeau, 1929: 5). Poles 14 and 16 sometimes occur in f a i r l y close association, for example on the dendrogram for clustering poles with 47 characters, and on the multi-dimensional scaling plot for poles with 71 characters. Visually also their association is not incongruous. However 18 and 21 do not occur with them nor with each other. Nor are they visually satisfying as the work of the same a r t i s t . In the case of this a r t i s t , the computer offers the probability that not a l l the poles attributed to him are actually his work. On the other hand i t also suggests a close association between some poles which are said to be carved by him. Conversely, poles attributed to Hlamee, numbers 13, 17, 19, 20, 30, 33, 34, and 51, remain widely scattered on both dendrograms and plots. Since the work of this a r t i s t has the greatest visual unity of any group of poles, from a subjective viewpoint, a national explanation for their failure to cluster is hard to find. Personally I would remove poles number 13 and 51 from the l i s t of his works, but this s t i l l leaves a long l i s t . A detailed study of Hlamee's poles could illuminate the problem. He may for instance have 68. copied the organization of different works he had seen. The computer has certainly not been able, in this programme, to capture the essence of his style. Even more intriguing is the corpus of work attributed to the carver Oyai: poles number 81, 87, 88, 90, 93, and 95. Barbeau has a great deal to say about Oyai "the outstanding craftsman of the Wolf clan at the canyon of the Nass." (Barbeau, 1964: 53) "The best carver of the Nass at the time - about eighty years ago." (Ibid: 11) Yet in a manuscript t i t l e d Emblems of Nobility by Barbeau which contains a l i s t by Charles Barton of notable Nass carvers, Oyai is not mentioned. Oyai is however a high ranking chief's name from Gitwinsilk, which brings us to the paradox Barbeau mentions only briefly: that of the chief who "stands over" the carver, accepting the credit. It seems a rather unusual circumstance that many of the names given by Barbeau for Gitksan and Niska carvers are also those of chiefs. But i f we put this information together with the fact that the clustering methodology has produced remarkably few conjunctions of the work of named art i s t s , we may begin to suspect that there were many more artists active than the ethnographic record indicates. In the case of Oyai, both the clustering and intuitive judgment suggest that the corpus attributed to him may have come from the hands of at least four different carvers. I would separate the poles in this way: numbers 93 and 95, with possibly the addition of 1 and 8; numbers 81 and possibly 82; numbers 88 and 90; and number 87. 69. The role of the carver of totem poles is also involved in consideration of the results of the clustering operations. Barbeau (1964), Garfield (1939), and Halpin (1973) have drawn attention to the role of the Gidzontk in Coast Tsimshian society. They were said to be carvers in secret, of the paraphernalia for dancing and i n i t i a t i o n ceremonies. Specifically they were said not to be carvers of totem poles. There are a few indications that this type of organization was active in Niska and Gitksan territories (Halpin, 1973: 90j Barbeau, 1958: 75). There is^also an implication that the carvers of secret masks might also actually be carvers of totem poles (Barton, n.d.). This matter should be pursued, since i t seems possible that the obscurity in which totem pole carvers are cloaked may indicate that their work was actually connected with that of the Gidzontk. The status enjoyed by this group, as "advisers to the chiefs in a l l matters" (Barbeau, n.d.: 2) is surely one of the most powerful ever achieved by artists anywhere, and a proper subject for further investigation. 70. EPILOGUE ART AND THE COMPUTER I believe this thesis has demonstrated the u t i l i t y of quantitative techniques as aids in defining art styles under certain circumstances. Of the four objectives outlined in the beginning, three have been met. Only traits associated with individual artists or groups of artists remain to be defined. The establishment of a taxonomy which has been demonstrated to be valid in terms of the independent ethnographic record i s the objective portion of this study. Inferences drawn from this taxonomy are my own. I believe these inferences rest on a firmer basis than those which I might have drawn had I chosen to proceed by analyzing the data in terms of attributes obvious to me. Another researcher may draw other inferences from the same taxonomy; i t i s certainly my hope that others w i l l be interested in the questions raised by my interpretations. F i g . l Pole 1 Fig.2 Pole 84 Fig.3 Pole 2 F i g . l Pole 66 F i g . l Pole 3 Fig.2 Pole 22 PLATE IV F i g . l Pole 54 PLATE V Fig. 1 Pole 56 PLATE VI F i g . l Pole 92 Fig.2 Pole 7 PLATE VII 77. F i g . l Pole 26 Fig.2 Pole 12 F i g . l Pole 46 Fig.2 Pole 38 Fig.3 Pole 45 PLATE IX F i g . l Pole 61 Fig.2 Pole 61, detail PLATE X Pig.l Pole 9 Fig.2 Pole 18 Fig.3 Pole 47 PLATE XI F i g . l Pole 43 Fig.2 Pole 10 Fig.3 Pole 33 PLATE XII F i g . l Pole 23 F i g . l Pole 24 Fig.2 Pole 32 Fig.3 Pole 31 PLATE XIV F i g . l Pole 68 Fig.2 Pole 29 Fig.3 Pole 34 F i g . l Pole 67 Fig.2 Pole 79 PLATE XVI 86. F i g . l Pole 4 Fig.2 Pole 48 Fig.3 Pole 89 F i g . l Pole 78 Fig.2 Pole 13 Fig.3 Pole 27 PLATE XVIII 88. Fig . l Pole 60 Fig.2 Pole 39 Fig.3 Pole 49 Fig.4 Pole 20 PLATE XIX 89. Fig . l Pole 71 Fig.2 Pole 37 Fig.3 Pole 55 Fig.4 Pole 62 F i g . l Pole 8 PLATE XXI Fig.2 Pole 17 Fig. 3 Pole 15 Fig.4 Pole 14 PLATE XXII Fi g . l Pole 30 Fig.2 Pole 36 Fig.3 Pole 91 PLATE XXIII 93. Fig.2 Pole 87 Fig.3 Pole 95 Fig.2 Pole 72 Fig.3 Pole 73 PLATE XXVI ft F i g . l Pole 63 Fig.2 Pole 96 Fig.3 Pole 76 .1 Pole 58 Fig.2 Pole 59, detail Fig.3 Pole 77 PLATE XXVIII F i g . l Pole 28 Fig.2 Pole 51 Fig.3 Pole 88 PLATE XXIX Fig.3 Pole 81 PLATE XXX 100. Fig.2 Pole 53 Fig.3 Pole 94 F i g . l Pole 21 PLATE XXXII 102. Fig.2 Pole 52 PLATE XXXIII 103. Fig.2 Pole 41 PLATE XXXIV 104. F i g . l Pole 65 FjLg.2 Pole 82 PLATE XXXV . . 105. F i g . l Pole 75 PLATE XXXVI 106. F i g . l Pole 42 Fig.2 Pole 57 Fig.3 Pole 50 PLATE XXXVII 107. Fi g . l Pole 69 Fig.2 Pole 70 APPENDIX I LIST OF SEVENTY-ONE CHARACTERS 1. Humanoid 2. Animal 3. Bird 4. Other 5. Character positions 5 to 48 inclusive consist of repetitions of the f i r s t four characters. 49. Pole figures clearly divided horizontally. 50. Series of near identical figures side by side in one segment. 51. Figures applied as discrete items, raised from the surface of the pole, but within the outer lines of the main pole silhouette. 52. Figures are intertwined between segments. 53. Humanoid eye orbit is less than 33 1/3% of the height of the head. 54. Humanoid eye orbit is more than 33 1/3% of the height of the head. 55. The juncture of raised surfaces with thenmain pole surface i s angular. 56. The juncture of raised surfaces with the main pole surface i s semi-angular. 57. Figures hold, or have applied to their surface, other smaller figures. 58. Humanoid or other lips delineated by incised lines only. 59. Humanoids have teeth. 60. Figures applied as discrete items, raised from the surface of the pole, but projecting beyond the lines of the main pole silhouette. 61. Whole segments of the pole are f u l l y modelled, with some portions projecting beyond the lines of the main pole silhouette. 62. More than half the humanoid pupil l i e s above a horizontal line drawn from outer point to outer point of eyelid. 63. More than half the humanoid pupil l i e s below a line drawn as above. 64. The perimeter of the humanoid pupil i s congruent with the eyelid line. 65. Humanoid eyelid line is incised. 66. Humanoid eyelid line i s carved with a single change of plane. 67. Humanoid eyelid line i s carved with a double change of plane. 68. Humanoid pupil i s circular. 69. Only two figure segments on pole are of approximately the same height. 70. More than two figure segments are of approximately the same height. 71. More than one set of segments are of approximately the same height. 108. APPENDIX II in back pocket. APPENDIX III LIST OF FORTY-SEVEN CHARACTERS 1. Pole figures are clearly divided horizontally. For an example, see Plate XXI, figures 1 and 4. 2. Series of near identical figures side by side in one segment. For an example, see Plate XXI, figure 4. 3. Figures applied as discrete items, raised from the surface of the pole, but within the outer lines of the main pole silhouette. For an example, see Plate X, figure 2. 4. Figures are intertwined between segments. For an example, see Plate XXIII, figure 3. 5. Humanoid eye orbit i s less than 33 1/3% of the height of the head. „ 6. Humanoid eye orbit i s more than 33 1/3% of the height of the head. 7. The juncture of raised surfaces with the main pole surface is angular. For an example, see Plate XXXIII, figure 2. 8. The juncture of raised surfaces with the main pole surface i s semi-angular. For an example, see Plate XXVI, figure 1. 9. Figures hold, or have applied to their surface, other smaller figures. For an example, see Plate XXII, figure 3. 10. Humanoid or other lips delineated by incised lines only. -— Eve O/zp/T 110. APPENDIX III Continued 111. 11. Humanoids have teeth. For an example, see Plate X, , figure 3. 12. Figures applied as discrete items, raised from the surface of the pole, but projecting beyond the lines of the main pole silhouette. For an example, see Plate XX, figure 1. 13. Whole segments of the pole are f u l l y modelled, with some portions projecting beyond the lines of the main pole silhouette. For an example, see Plate II, figure 2. 14. More than half the humanoid pupil l i e s above a horizontal line drawn from outer point to outer point of eyelid. 1 15. More than half the humanoid pupil l i e s below a line drawn as above. 16. The perimeter of the humanoid pupil is congruent with the eyelid line. 17. Humanoid eyelid line i s incised. 18. Humanoid eyelid line is carved with a single change of plane. WULNAQAQ AND MEN ESK JAC Q METHOD 8 3. 5410 2. 8891 2.2371 1 .5851 0 .9331 0.2811 3 . 8 6 7 0 3 . 2 1 5 0 2 .5631 1.9111 1.2591 0 .6071 - 0 . 0 4 4 9 APPENDIX II 109. » — — I 1 ( 1 ) 0 . 5 2 3 8 I.-II 14 < 66 ( 14) 66) 0 . 4 2 8 6 0 .6131 75{ 21" 12(. 75 ) 2) 121 0 .9688 0 .5200 0 . 6 7 8 5 I I I I 3( "46'f 26 { 3) :4"6"T 26) 0 .4444 1*515 2 0 . 2 8 5 7 6 H 60 r 63 ( 61) 0 . 9 553 60) 63) 0 .4000 0 . 5 1 3 7 76 { 41" 40{ 76) 4) 40) 1 .7563 0 . 5 6 2 5 0 .9830 — 38 { 38,) 0 . 6 6 6 7 52 ! 29 { 52) 29) 1.1071 0 .2222 31< 32 { 13{ 31) 0 . 5 3 7 0 32) 13) 1.3 569 0 . 4 4 4 4 I I I I 78{ 22( 45( 78) 22 ) 45) 0 .8648 0 . 6 9 7 0 0 .8768 30{ 30) 0 . 4545 67? 33 ( 67 J 33) 0 .6 233 1.0332 f I I I I I I 20( 20 ) 0 .3333 I.M..CL I I I I I I . - 79( 79) 1.1955 I 1 I \ I I I • -— — "~ —•—•' — — ~ — 23 ( 23) 0 .5094 / f J I I I s I I I I . 43 ( 433 0,3 000 I I I I I • T 44 ( 44) 2 .0631 1 I • 6( 6) 0 . 4 2 8 6 I I I • — — — — — — 93{ 93) 0 .6101 I I I * 92 ( 92 ) 0 .7896 I I I I • — — — — — — 27( 27 ) 0 . 3 8 4 6 I I I I I • 47 { 47) 0 .4242 I I I I « 54( 54 i 1.0638 I I I I • 9( 9) 0.5 000 I I I I I • 34 { 34) 0 .7123 I I 1 I I I • 24 ( 24) 0 . 3 5 2 9 I I I I I I 1 25( 25) 0 .434 7 I I I I I • 68 ( 68 J 1 .3844 I I I I • 7( 7) 0 .7421 I I I I I I 56 ( 56) 0 .3571 I I I I I I * 84 ( 84) 0 . 7 9 2 5 I I I I I I . • — — 81 ( 81) 0 .4000 I I I I I I I. 90( 90) 0 .722 7 I I II I I II • 88 ( 88] 0 .5 714 I I II I T 94( 94) 3 .7221 1 I 5 ( 5 ) 0 .2 500 I I I • 41{ 41) 0 . § 8 3 3 I I I • 1 8( 18) 1.2730 I I I I • 19 ( 19) 0 . 7 1 4 3 I I I I 1 • 21( 21 ) 0 .9515 I I I I I I • 501 50) 0 . 2 8 5 7 I I I I I * 72 ( 72 ) 2 . 0 0 4 7 { 8) 36) 17) 0 . 3 8 4 6 0 . 5 1 2 8 0 , 8 9 8 9 10_1b 36 ( 17 { 37 ( "651' 82 ( 37) 65) 82) 0 . 4 1 6 7 0 .7 415 0 . 5 2 6 3 I I i i i 89 ( 89) 1.3 489 1 1 • 1 • i • — — — — — — — —— 10( 16{ 10) 16) 0 .5 714 0 .6978 I I I i i . i i 15 ( 15 3 0 . 5 3 3 3 I I I • - 86( 86) 0 .9410 I I I I • ——————— 11 { 11) 0 .4000 ! 1 i i i i • ———————— — 83 ( 83) 0 .6 749 i i I i i i 1 X • 80( 911 80) 91) 0 .2941 0 . 5 3 2 8 85( 85) ' 4 8 1 4 8 ) 491 49) 0 . 9787 ~0TTS¥3 1.4133 28 ( 28 ) ' 9 5 1 9 5 " ) ' " 35( 35) 0 . 4 4 4 4 0 .6 33 0 0 . 4 2 8 6 1 I 87( 87 ) "5"1"{51") 70 { 70). 0 .7 079 0 .2 857 0 .9337 I I.' I I 53 ( 53 ) 5 5 (5"5" ) " ' 64( 64) 0 .4950 '0'.'3"5'7"i" 2 . 3 4 1 7 I "I I I 39 ( 62 ( 69{ 39 ) "62"")" 69) 0 .5 226 0.3'7 50 1.5 572 {  1 I I \ I I I • 42 ( 42} 0 .2500 I I I I I 1 I • 73 { 73 ) 0 .7 107 I I I I s 1 I I I 571 57) 0 .3333 / ( I I I I I \ I I I • 71 ( 71} 0 .8825 I I I I I I I I # — — *—• 59 ( 59} 0 . 1 0 0 0 I I I I I I I I • 77 ( 77 ) 1.1593 I I I I I I I I • 58 ( 58) 0 .5 257 I I I I I I I I I I .- 74 ( 74) 0 . 4 706 •I I I I II • 96( 96) 3 . 8 6 7 0 3 . 2 1 5 0 2 .5 631 1 .911 1 1.2591 0 .6071 - 0 . 0449 3 .5410 2.8891 2. 2371 1 . 5851 0 .9331 0 .2811 DENDROGRAM: VALUES ALONG X -AXIS ARE SCALED COEFF IC IENTS VALUES ALONG Y-AXIS ARE MERGE LEVELS EXECUTION TERMINATED $SIGNOFF J WUL MEN ET AL ETC HO SANS QUALITY FRTYSVN METHOD 8 3 .2028 2 .5960- 1.9891 1.3823 0 .7754 0 .1686 3 .5063 2 . 8 9 9 4 2 .2926 . 1 .6857 1.0788 0 . 4 7 2 0 - 0 . 1349 APPENDIX IV -417-r-H 1) 8 4 ( 8 4 ) 21 2) 0 . 5 0 0 0 0 .863 2 0 . 6 1 5 4 66{ 66) 6 { 6 ) 93 { 93) 0 . 8 4 2 3 0 . 4 1 6 7 0 .9801 3( 3) " 2 2 ( 2 2 ) 54C 54) 0 . 4 583 '0.73 ' lb^ 0 . 4 3 7 5 I I I I I I II 56( 56) 92 ( 9 2 ) 7{ 7 ) 0 .6718 1.2489 0 . 6 3 6 4 26( 26) 1 2 ( 1 2 ) 46( 46) 1.0501 0 . 5 6 6 6 0 . 7 1 5 9 I I I .-I I 38 ( 38) 4 5 1 4 5 ) 61{ 61) 0 .7735 6 . 6 1 5 4 1.7021 I I .-I I 9t 9) 18 ( 1 8 ) 0 . 5 6 8 0 0 .4375 I I • — — 43 { 43) 0 .9689 I I 1Q{ 10) 33( 33) 0 . 5 7 8 9 0 .772 8 23( 23) 4 4 { 4 4 ) 2 4 1 2 4 ) 0 .4444 1.2051 0 . 4 0 0 0 I I I I I 321 32) 0 .6218 ..I I I I I I 31 ( 31 ) 0 . 3 8 8 9 I I I I J I . 6 8 ( 68) 0 . 7 2 8 9 2-i i iI I I . • 29 ( 29) 0 . 4 6 6 7 I I I I I . 34 ( 34) 1. 295 8 I I I I I • I - 79 { 79) 3 .3714 I 4( 4) 0 . 6 6 7 8 I I I I - 48t 48) 0 . 5 0 0 0 I I I I . — — — — 89 { 89) 1 .0918 I I i i . — i 3 ( 13) 6 . 5 6 0 6 I I I I ' ; ; , 1 . , 78 ( 78) 0 .6651 I I I I I I . 27 { 27) 0 . 5 2 6 3 I I I I I " I.— 60 it 60) 1.0 574 I II I II . 391 39) 0 . 3 7 5 0 I II I I . r 49 { 49) 1 .4892 I I I I . 20 ( 20) 0 . 3 3 3 3 I I I I I . 711 71 ) 0 . 8 4 7 3 I I I . — 371 37) 0 .4286 I I I I I I I . 551 55) 0 . 6 4 6 0 I I I I T _ : . 62 ( 62) 1 .8963 I I I I . 81 8) 0 .4231 I I I I I . - - — 251 25) 0 . 5 4 2 0 I I I J • I : I . — 16 1 16) 0 . 4 6 6 7 I I I I I I 171 17) 0 . 6 7 1 2 I I I I I I. 14 { 14) 6 . 6000 I I II I : I . I 5 ( 1 5 ) Q .8218 I I I 301 30) 0 . 6 3 9 0 I I I I I I I I . - 361 36) 0 . 4 7 8 3 I I I I I I ' L •»• • — 911 91) 1 .1894 I I I 35 < 87 ( 95 { 35) 87) 95 ) 0 .4091 0. 4902 1.0066 t ! 7 b 80( 83 ( 85 ( 80) 83) 85) 0 . 4 1 6 7 0 . 5 3 5 6 0 . 3 9 1 3 I I." II I I 86( T i l 72 ( 86) 11) 72) 1.8343 6 . 4 0 7 4 0 .5442 II I I II II II II II II II 73{ 63( 96 ( 73) '63) ' 96) 0 . 8 2 0 4 O". 5 6*2 5' 0 .6125 II II II' I I I I II II 76( "5di 59 { 76) 58) 59) 0 . 8 9 9 8 0 . 5 0 0 0 0 . 6 1 2 7 II II I I 77( 51 ( 77) '28)' 51 ) 1 .3294 0 .4 76 2 0 . 6 3 5 1 • II II I I I II II II I I . I I 88( 90 ( 74 f 88) 90) 74) 0 . 5 0 0 0 0 .8628 0 . 3 8 8 9 II II II II II I I 81( 19{ 53 ( 81) 19) 53 ) 1 .4334 0 .8710 0 .4583 II II II II II II 94( 211 40 ( 94) 21 ) 40) 1 .0606 0.7 554 0 .4706 II II I I II 52( 5 ( 41 ( 52) 5 ) 41) 2 .2750 0 .2 500 0 .5000 (  1 I I I I • ™ . 6 4 < 64) 0 .9834 I I I I I I . . fc5( 65) 0 . 5 714 I I I I v • ' I I I . a2( 82) 0 . 6 3 4 9 ) ( I I I I I I * — 75 ( 75 ) 1 .1995 I I I I I I . 42 ( 42) 0 . 1 6 6 7 I I I I I I I . 57 ( 57) 0 . 7 7 1 4 I I I I I I I I . 50 i 50 ) 0 .2 500 I I I I I I I I I . 69 ( 69) 0 .3333 I I I I I — 70 ( 70) • + + + + + _ + +— 3 . 5 0 6 3 2 . 8994 2 . 2 9 2 6 1 .6857 1 .0788 0 .4720 - 0 . 1 3 4 9 3 . 2 0 2 8 2 .5960 1 .9891 1 . 3 8 2 3 0 . 7 7 5 4 0 . 1 6 86 DENDROGRAM: VALUES ALONG X-AXIS ARE SCALED COEFF IC IENTS VALUES ALONG Y-AXIS ARE MERGE LEVELS EXECUTION TERMINATED 0 8 : 0 4 : 0 8 7 = 4 . 0 9 5 RC=0 $5 .91 SSIGNOFF 0) • M A D E IN C A N A D A APPENDIX V 118. "34 •41 *50 •70 44 •3* * ^ •?4 ^ •8/ • 3 ^ ' -So'W •a a. •43 •4+ 5"4 *3( *2<? "5* •3 •<7'3 WUL MEN ETC HRSANS QUALITY JULY-METHOD 8 1 2. 1593 1.8023 1.4452 1.0882 0 .7312 0.3742 2 .3378 1 .9808 1 .6238 1 .2667 0 .9097 0 . 5 5 2 7 0 . 1 9 5 7 APPENDIX y i 119. + + + + + — + , + - + + + + + + 1 { 1) 0 .7241 * I T — 36 ( . 1 3 ( 3 6) 13) 0 . 9 1 0 6 0 .5 833 I I I * 31 ( 31 ) 0 . 8 0 1 0 I I I I I I . 42 ( T _ 47 ( 42) 47) 0 .5152 1.2911 3t 3) 6 ( b T 19i 19) 0 .8621 0 .7 63 3 0 .6111 26{ 26) 4 ( 4 ) 41 ( 41) 1.0542 0 . 7 3 6 8 1.3341 2i 2) 39 ( 3 9 ) 32 C 32) 0 . 5 5 5 6 0 . 5 3 5 7 40{ 40) 3 3 ( 3 8 ) 10( 10) 0 . 8 5 9 7 "I7T863" 0 . 8 1 8 2 I I .-I I i i i i i i 11 \ 17) 23 ( 2 3 ) 35( 35) 1. 1146 6 . 7 6 6 8 0 . 6 6 6 7 I I I I 46( 46) 15 ( 1 5 ) 34( 34) 1.4 360 6 . 8 571 0 .9661 I I T I .• II 22( 22) 45 ( 4 ' 5 ) " 33 1 33) 0 . 6 8 7 5 0 .9 275" 2 .2585 5( 5) 28( 28) 27( 27) 0 . 5385 ^0.6 23 6* 0 . 5 0 0 0 MIL I i 29 ( 29) 0 .7 648 i © i n n ) 0.8456 T I I I 21( 21) 0.6 800 I I I 43< 43) 3 ( 8 ) 14( 14) 0 . 9 7 5 6 6 .3 889 0 . 5 3 7 4 I 37( 37) 9 ( : 9 ) " ' , 18( 18) 0 . 6 5 9 2 0 . 2 7 5 0 I I I I I I 1 I I I I . - 20( 20) 1.278 5 . I I 7( 12 ( 7) 12) 0 . 5 2 5 0 0 . 6 6 9 6 I I I . 2 4 ( 24) 0 . 5 7 1 4 T T I I I I © 30 ( 30 ) 0 . 8 1 7 0 T I I I I I • 16 { 16) 0 . 4 8 0 0 I I I I T I I * 25 { 25) 0 . 5 6 6 3 I I I I + + -2 .3378 — + • +-1.980 8 ._+ +.. 1.6238 — + - +-1 .2667 •-+ +• 0 .9097 — + +-0 . 5 5 2 7 — 44( 44) — + 0 . 1 9 5 7 2 .1593 1.8023 1.4452 1.0882 0 . 7 3 1 2 0 .3742 DENDROGRAM: VALUES ALONG X-AXIS ARE SCALED COEFF IC IENTS VALUES ALONG Y-AXIS ARE MERGE LEVELS . EXECUTION TERMINATED 1 7 : 5 6 : 1 3 T=1.21 RC=0 $ 1 . 8 7 $SIGNOFF APPENDIX III Continued 112. 19. Humanoid eyelid line is carved with a double change of plane. 20. Humanoid pupil i s circular. 21. Only two figure segments on pole are of approximately the same height. For an example, see Plate XXXIII, figure 3. 22. More than two figure segments are of approximately the same height. For an example, see Plate XI, figure 3. 23. More than one set of segments are of approximately the same height. For an example, see Plate XXXV, figure 1. 24. Eyebrows almost meet—separated by less than the width of the bridge of the nose. 25. Eyesocket is:;crisply;:cut.For an example, see Plate III, figure'2. 26. The junction of the chin and neck has a "prow-like" appearance, due to the relatively sharp recession of the f a c i a l planes on either side of a vertical median line. For an example, see Plate XXVI, figure 2. 27. The mouth and chin l i e in a continuous plane, separated from the rest of the face by a deeply cut intersection with the cheek planes. For an example, see Plate II, figure 3. 28. The base and sides of the nose may be enclosed within an approximately equilateral triangle. 29. The horizontal dimension of the mouth appears greater than 2/3 of the width of the head. For an example, see Plate'II, figure 3. 30. The wings of birds have no decorative motifs. 31. Bird form has human limbs. For an example, see Plate X, figure 1. 32. Arm position (after Inverarity, 1) APPENDIX III Continued 33. Arm position (after Inverarity, 4) 113. 34. Arm position (after Inverarity, 5) 35. Arm position (after Inverarity, 6) 36. Arm position (after Inverarity, 7) APPENDIX III Continued 114. 37. Arm position (after Inverarity, 8) 38. Arms held horizontally at waist, sinister above dexter. 39. Arms held horizontally at waist, dexter above sinister. 40, Leg position (after Inverarity, 51) APPENDIX III Continued 41. Leg position (after Inverarity, 54) 115. 54 42. Leg position (after Inverarity, 55) X 55 43. Leg position (after Inverarity, 58) 58 n n 44. Leg position (after Inverarity, 59) 59 APPENDIX III Continued 116. 45. Leg position (after Inverarity, 62) 1 Mi 6 2 a . 6 2b 46. Leg position (after Inverarity, 63) L 6 3 -47. Arm position (after Inverarity, 2) APPENDIX IV in back pocket. APPENDIX V in back pocket. APPENDIX VI i n back pocket. BIBLIOGRAPHY Adams, John W. The Gitksan Potlatch: Population Flux, Resource 1973 Ownership and Reciprocity. Toronto: Holt, Rinehart and Winston of Canada, Ltd. Balfour, H. The Evolution of Decorative Art. London: Rivington, 1893 Percival and Co. Barbeau, Marius, and William Beynon. {Ethnographic Notes from the [ca.1915-1955) Nass and the Skeena, Fieldwork] (Manuscript in the National Museums of Canada, Folklore Division). Barbeau, Marius. "Haimas' Crab Feast." Pp. 59-64 in Raven-Clan [ca.1927] Outlaws on the North Pacific Coast. (Manuscript, National Museums of Canada, Folklore Division) Barbeau, Marius. Totem Poles of the Gitksan, Upper Skeena River, 1929 British Columbia. National Museum of Canada Anthropological Series Number 12, Bulletin 61. Ottawa. (Reprinted: Ottawa, 1973) Barbeau, Marius. Totem Poles, Vols. I and II. National Museum of 1950 Canada Anthropological Series Number 30, Bulletin 119. Ottawa. (Reprinted: Ottawa, 1964) Barbeau, Marius. Medicine-Men on the North Pacific Coast. National 1958 Museum of Canada Anthropological Series Number 42, Bulletin 152. Ottawa. (Reprinted: Ottawa, 1973) Barrett, S.A. The Ethnogeography of the Pomo and Neighboring 1908 Indians. University of California Publications in American Archaeology and Ethnology Number 6: 1-332. Barton, Charles. "Carvers of the Nass River." No page number: in 1927 Emblems of Nobility, Marius Barbeau. (Manuscript, National Museums of Canada, ."FolkloreODivision) . Berndt, Ronald M. "Some Methodological Considerations in the 1958 Study of Australian Aboriginal Art." Pp. 99-100 in Art and Aesthetics in Primitive Societies, Carol F. Jopling, ed. New York: E.P. Dutton and Co. Inc. Berenson, Bernhard. Rudiments of Connoisseurship: Study and 1902 Criticism of Italian Art. Reprint: New York: Schocken Books, 1962. 120. 121. Boas, Franz. Indianische Sagen von der Nord-Pacifischen Kuste 1895 Amerikas. Sonder-Abdruck aus den Verhandlungen der Berliner Gesellschaft fur Anthopologie Ethnologie und Urgeschichte. Berlin: Ascher. Boas, Franz. Tsimshian Mythology. Bureau of American Ethnology 1916 Annual Report, 31: 27-1037. Boas, Franz. Primitive Art. Oslo: H. Aschehoug and Co. '(Reprinted: 1927 Dover Publications, New York, 1955) Boas, Franz. Race, Language, and Culture. New York: The Free Press, 1966 Macmillan Publishing Co., Inc. Brasser, Ted. A Basketful of Indian Culture Change. National 1975 Museum of Man Mercury Series, Canadian Ethnology Service Paper Number 22. Ottawa. Brothwell, D. and Eric Higgs. Science in Archaeology. London: 1971 Thames and Hudson. Burkitt, Miles. The Old Stone Age. Toronto: McClelland and 1963 Stewart Ltd. Clarke, David, ed. Models in Archaeology. London: Methuen and 1972 Co., Ltd. Crosby, Harry H. and George R. Bond. The McLuhan Explosion. New 1968 York: American Book Co. Culkin, S.J., John M. "A Schoolman's Guide to Marshall McLuhan" 1968 Pp. 179-189 in The McLuhan Explosion New York: American Book Co. Darling, John. Tsimshian System of Land Holdings. (Unpublished M.A. 1955 Thesis in Anthropology, .^University of British Columbia). d'Azevedo, Warren L., ed. The Traditional Artist in African Societies. 1975 Bloomington: Indiana University Press. Deetz, James. Man's Imprint from the Past. Boston: L i t t l e , Brown and Co. 1971 Dixon, R.B. and A.L. Kroeber. Linguistic Families of California. 1919 University of California Publications in American Archaeology and Ethnology Number 16: 47-118. Driver, Harold E. and James L. Coffin. Classification and Development 1975 of North American Indian Cultures: A S t a t i s t i c a l Analysis of the Driver-Massey Sample. Transactions of the American Philosophical Society N.S. Vol.65(3). 122. Driver, Harold E. "Survey of Numerical Classification in Anthropology." 1965 Pp. 301-344 in The Use of Computers in Anthropology. The Hague: Mouton. Drucker, Philip. Indians of the Northwest Coast. Garden City: 1955 The Natural History Press. Duff, Wilson, ed. Histories, Territories, and Laws of, the 1959 Kitwancool. British Columbia Provincial Museum," Anthropology in British Columbia Memoir Number 4. Victoria. Duff, Wilson Indian History of British Columbia Vol.1. The 1964 Impact of the White Man. British Columbia Provincial Museum, Anthropology in British Columbia Memoir Number 5. Duff, Wilson. Images Stone B.C. Saanichton: Hancock House 1975 Publishers Ltd. Emmons, George T. Origin Kitcksian. [Fieldnotes] (Manuscript, tea.1910] National Museums of Canada, Ottawa. Flynn, Francis R.J. A Temporal Classification of Folklore of the 1976 Okanagan Indians. (Unpublished M.A. Thesis in Anthropology, University of British Columbia, Vancouver, B.C.). Forge, Anthony, ed. Primitive Art and Society. London: Oxford 1973 University Press. Fraser, Douglas, eS'.ie The'Many Faces of Primitive Art. Englewood 1966 C l i f f s : Pr.ehtice-Hall Inc. Fraser, Douglas: Primitive Art. New York: Doubleday and Co., 1962 Inc. Garfield, Viola E. Tsimshian Clan and Society. University of 1939 Washington Publication in Anthropology 7: 167-340. Seattle. 123. Garfield, Viola and Paul S. Wingert. The Tsimshian Indians and 1951 Their Arts. Seattle: University of Washington Press. Haddon, Alfred C. The Decorative Art of British New Guinea. 1894 Dublin: The Academy House. (Reprint: AMS Press, New York. 1977) Halpin, Marjorie M. The Tsimshian Crest System. (Unpublished 1973 Ph.D. dissertation in Anthropology, University of British Columbia, Vancouver, B.C.). Herskovits, Melville J. Cultural Anthropology. New York: Alfred 1966 A. Knopf. Hodson, F.R. Classification by computer. Pp. 649-660 in Science 1971 in Archaeology, Brothwell, D. and Eric Higgs. London: Thames and Hudson. Holm, B i l l . Northwest Coast Indian Art. Seattle: University of 1965 Washington Press. Holm, B i l l . Heraldic Carving Styles of the Northwest Coast. 1972 Pp. 77-83 in American Indian Art: Form and Tradition. New York: E.P. Dutton and Co. Inc. Holmes, W.H. Origin and Development of Form and Ornament in Ceramic 1886 Art. Bureau of American Ethnology, Fourth Annual Report: 443-465. Hope, Keith. Methods of Multivariate Analysis. London: University 1968 of London Press Ltd. Inverarity, Robert Bruce. Observations on Northwest Coast Indian 1972 Art and Similarities Between a Few Art Elements Distant in Time and Space. Pp. in Early Chinese Art and Its Possible Influence in the Pacific Basin. Vol..^Ili. Barnard, Noel, and Douglas Fraser, eds. New York: Intercultural Press. Johnson, Leroy Jr. Introduction to Imaginary Models for Archaeological 1972 Scaling and Clustering. Pp. 309-379 in Models in Archaeology. David Clarke, ed. London: Methuen and Co., Ltd. 124. Jopling, Carol F. ed. Art and Aesthetics in Primitive Societies. 1971 New York: E.P. Dutton and Co. Inc. Kaufmann, Carole Natalie. Changes in Haida Indian A r g i l l i t e 1969 Carvings, 1820-1910. Ann Arbor: University Microfilms Ltd. Kroeber, A.L. Anthropology; Culture Patterns and Processes. 1923 Excerpts. New York: Harcourt, Brace and World Inc. 1963. Kroeber, A.L. History and Science in Anthropology. American 1935 Anthropologist Vol. 37. Menasha. Kroeber, A.L. Cultural and Natural Areas of Native North 1939 America. (Reprint: University of California Press. Berkeley, 1947). Kroeber, A.L. Toward Definition of the Nazca Style. University 1956 of California Publications in American Archaeology and Ethnology. Number 43(4): 327-332. Kubler, George. The Shape of Time. New Haven: Yale University 1965 Press. Levi-Strauss, Claude. Social Structure. Pp. 321-350 in Anthropology 1962 Today, Tax, Sol, ed. Chicago: University of Chicago Press. Levi-Strauss, Claude. Totemism. Boston: Beacon Press. 1963 Levi-Strauss, Claude. Structural Anthropology. Garden City: 1967 Anchor Books, Doubleday and Co. Inc. Levi-Strauss, Claude. The Savage Mind. Chicago: The University of 1970 Chicago Press. MacDonald, George F. and Richard I. Inglis. The Dig. National 1976 Museums of Canada, National Museum of Man. Ottawa. Macnair, Peter. Personal Communication. 1977 McLuhan, H. Marshall, and Harley Parker. Through the Vanishing 1968 Point. New York: Harper and Row. Matson, R.G., and D. True. Site Relationships at Quebrada 1974 Tarapaca Chile: A Comparison of Clustering and Scaling Techniques. American Antiquity Vol. 39(1), January. 125. Matson, R.G., and W.D. Lipe. Seriatlon of Pueblo Ceramic 1978 Assemblages from Cedar Mesa, S.E. Utah. (In press) Metropolitan Museum of Art. Computers and Their Potential 1968 Applications in Museums. New York: Arno Press. Munro, Thomas. Form and Style in the Arts: An Introduction to 1970 Aesthetic Morphology. Cleveland: The Cleveland Museum of Art and Case Western Reserve University. Otten, Charlotte M. ed. Anthropology and Art. Garden City: The 1971 Natural History Press. Panofsky, Erwin. Studies in Iconology. New York: Harper and Row. 1962 Petrie, Sir William Matthew Flinders. . Journal of 1899 the Anthropological Institute, N.S. 29: 295-301. Richardson, Miles, ed'.P The^Human Mirror :""IMaterial"and Spatial 1975 Imagesaof Man: Baton'Rouge: Louisiana State University Press. ? r " s . ; *V" >: 2"'S •?:? >'h. . _ .. _ Rossi^ /Ino, ed.i , The sUnconsciouscini.Culture r^The^ Structuralism 1974 of Claude Levi-Strauss in Perspective. New York: E.P. Dutton and Co. Inc. Salmon, Merrilee H. "Deductive" Versus "Inductive" Archaeology. 1976 American Antiquity 41(3) : 376-380. Schapiro, Meyer. "Style" Pp. 278-303 in Anthropology Today, 1962 Tax, Sol, ed. Chicago: The University of Chicago Press. Smith, Marian, ed. The Artist in Tribal Society. London: Routledge 1961 and Kegan Paul Ltd. Sneath, P.H. and R.R. Sokal. Principles and Practice of Numerical 1973 Taxonomy. San Francisco: W.H. Freeman. Stuart, Alan. Basic Ideas of Scientific Sampling. New York: 1968 Hafner Publishing Co. Sturtevant, William C. "Does Anthropology Need Museums?" 1969 Proceedings of the Biological Society of Washington Vol. 82: 619-650. Sturtevant, William C. "Museums as Anthropological Data Banks" 1973 Pp. 40-55 in Anthropology Beyond the University Redfield, Alden, ed. Southern Anthropological Society Proceedings Number 7. Athens. 126. Tax, Sol, ed. Anthropology Today. Chicago: University of Chicago 1962 Press. Torgerson, W.S. Multidimensional Scaling of Similarity. 1965 Psychometrika 30: 379-393. Wingert, Paul S. An Outline Guide to the Art of the South 1946 Pacific. New York: Columbia University Press. Wingert, Paul S. "Tsimshian Sculpture" Pp. 73-94 in 1951 The Tsimshian Indians and Their Arts. Garfield, Viola, and Paul S. Wingert. Seattle: University of Washington Press. Wood, John J. "A Computer Program for Hierarchical Cluster 1974 Analysis." Pp. 1-11 in Newsletter for Computer  Archaeologists 9(4). University of Arizona. 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items