—f— _ Q A .C a. .‘¢ .. y o Ly :« A» v§ .,\ —: am ”A,” 't. p: ‘- “; fl 5 0‘ § . N.‘ “H Wig v; 5‘ .Q ~ ‘9' V fl .M. 5» L. A. “a. J. v. ..‘ a: A. «y. S ;. aw. v.. H! .» .¢ . . .t o ‘ y «L. ~ s 5 fl» .1 t H]. .3 a. . _~ .3 .,. .. .4 . .3 a» a... e s . v‘ .Q e «3 ul. 5» Ab a3 . AC v; a E .3. at an “A. v - V . n4,u PL M. . A “A; o s S 0 e .3 ._ . a . .C 3: e .u . . In. . . A . «a .C .. .4. .. T. a» at Cm. .a. ‘. v. . ..& ‘ 3. n. .3 v . . v .. . 2. .. u .. . e C C E C u. 2: I. . g -. ‘ .3 an .a. q p. .: nu. .nyu v~ Vs \‘J a» C» .. .... .. -.. .3 .. .s‘ .. «In .3 ‘7 g .. ¥~ :. .u. .\. .s. .1 .5 85 ~ 5 b ‘ :s «\- S. v . .. . in . . . .. . .I : . .3 n a .3 v. 2. C . a u. .. u .n . .\~ .n . u 5 . nu ~ g : s u . 3. \Ph NV ABSTRACT MODELS OF PRE-DORSET CULTURE: TOWARDS AN EXPLICIT METHODOLOGY by Albert Arch Dekin, Jr. '9/: J Past archaeological methods and techniques have not resolved several long-standing problems in Eastern Arctic prehistory. Moreover, the recent resurgence of once- discarded models leads to the conclusion that our research has not resulted in significant improvements in our under- standing of processes of Pre-Dorset behavioral variation in space and time. Archaeological research has been con- ducted under implicit paradigms with imprecise methods and techniques. Variations in archaeological interpretations of the Arctic Small Tool tradition have resulted from the lack of a generally accepted paradigm and from variations in archaeological data, techniques, and methods.’ Such in- terpretiveAproblems are characteristic of our study of Pre-Dorset structures, the Pre-Dorset migration into the Eastern Arctic, and the subsequent processes of formation of regional variants of Pre-Dorset culture. The explanation of behavioral change in archaeologically-known populations requires a precise chronological framework. Using available radiocarbon dates from the Arctic Small Tool tradition, making implicit pos- sible sources of variation and adjustments in individual . . ’ ...- g o ‘z'y‘: vie A ...,~‘ yo~vv U h . o -.v~p1 \~-Q . O . - fl; b_--..V.»L' .: t. a. ' 4'- Q. a-‘v.b‘.—...‘- 5 a 4 u. .n...‘...‘..\, .p ‘ U o ~~~< bn .q.,.,,‘ I— . P‘ 4 ‘ "-I -\- yvv-‘vgne . . .‘ H}... "ou. _. . . '."” ~~ u'u..- .‘. . ,— . ”Ha—t“, ‘ 3r, . \ g.” a. .n;\ A: u, ’0 ..VVo U. 3 . “‘ ‘ I “‘v.' A - ADV. 5»... 'V“..—_ es»..._.:, *- .. .. ,4 ‘K I q- v§5_.‘..~F “ \ to .‘d ‘v. ' ..-"“ J“ . “-‘a £3. 6"...“ ‘I"‘ ‘§ 5 4 Qk‘.e a. “a . . y 0 - .. a LA- ' . EA... :. § ¥‘:e Q > “‘ ~. F» 51"" - H.~:.a ‘q‘ : "W. C. s. s "u ‘U‘ '0 F l n- u \ boo.“ “1| . s'.5:- I-flhn ‘a-c . y f - £ T\* \" V‘ I I" ‘U ‘ Q‘ _ F’CDK' A \‘ , 3«=S , D d:- N-‘ .- ‘ . k J‘ U . a. .~ . 5‘ 'fi'.‘ ‘n‘c‘ “x Albert Arch Dekin, Jr. dates, these data are treated precisely and a provisional chronology is established. Techniques for the evaluation of significance of difference between individual dates are used to establish the probability of contemporaneity of several components of the Closure site (Kqu 11). Because human behavioral systems are our major means of adaptation to our environment, any Changes in the en— vironment of prehistoric Arctic p0pulations are potential sources of variation in human behaviors. Data on prehis- toric environmental change during the time of Pre-Dorset occupation (4050-2750 B.P.) are synthesized to establish trends in environmental change and a marked shift in climate at SE- 3500-3600 B.P.. Precise field techniques and data depiction from the excavation of the Closure site (Kqu 11) resulted in data on rock and artifact distributions manipulated using an elliptical data structure and statistical tests of signi- ficance to test hypothesized structural attributes. These methods produced a hypothetical model of Pre-Dorset tent structures at the Closure site, which is suggested for further testing on comparable Pre-Dorset data sets. The value of precise and explicit field techniques and analyti- cal methods is demonstrated. A model of the migration of Arctic Small Tool tradi- tion peOples as a diverging horizon is derived deductively from theoretical dispersal processes in other species in which there is increasing behavioral variegation with . «"lan v~~ use.“ ‘.:u8b>dfi .tqu I a DAQpA‘ ‘F :v 5: uCD~CU Vu u ' I o "v'fifln H“n: -w..&v'up 00....- n, 3 9|. (J‘ . , ' V Y:""" ‘q Don, ,. 'T“:'} 5" an; .. . . ‘.-. A‘ .. ‘. > u': N" ‘ ““u- svuu. . J . , . "“' 3'. bu .. ..'V“ ‘ -~. .- ‘ ‘.¢ - A: e. \‘\~ A HVV‘ - u i-.;1 "”2I5- ".1; ~.....,-_: - . 5.“. ’Vq- A. \- -._,_- U. u ~h--‘ . 5H: ’ :2" “A ’ “o... 5v‘.: b “y, . {- C--1 _ 1 had A‘- I ‘ LL‘J‘ :‘H R ‘ R- 5..-..Le :a. ‘\ ‘ JV . n. _‘ ‘s;": S“ a” 3":‘5: ", F‘At ~Q:_ern Ava» _ 3 C ‘0’ .1 Albert Arch Dekin, Jr. dispersal into an unoccupied ecological niche. This model is tested on available data from the Arctic Small Tool horizon, using artifact variation in burins, microblades and end blades. In spite of the limited data available on variety in these artifacts, the data on morphological and dimensional variegations are in accordance with the predic- tion of the model and it is suggested for further testing. An elaborate model of the relationship between environ- mental change, social scale, and technological change is de- rived from other studies of environmental change, social change, and economic development. This model is tested with data on the development of regional variants of the Eastern Arctic Small Tool tradition. Portions of the model accurately predict the macro-fragmentation of the Arctic Small Tool horizon following the demonstrated climatic change ca. 3500-3600 B.P.. This study demonstrates the impact which implicit as- sumptions and imprecise methods and techniques have had on our understanding of prehistoric behavioral processes in the Eastern Arctic. The advantage of precise field techniques for the testing of hypotheses is demonstrated. The use of explicit and precise processes of model building and testing leads to a greater methodological SOphi- stication and to the increased potential for significant theoretical contributions. Models of Pre-Dorset structures, of the Arctic Small Tool horizon, and of the impact of en- vironmental change on cultural systems may be developed and " A- (. ' ‘ ...—. a..- .‘n - a ov-uu ‘ot'u Cuv .\ . -. n~.b.-.‘ ,. .1 v ~‘~ui‘.d" .y _ u . . -. ~~_v—~ ‘v“ t A ’ QIGIOUO“‘ V..‘:: - Albert Arch Dekin, Jr. tested. Greater technical precision and methodological ex- plicitness are necessary if we are to benefit from the time depth and ecological dimensions of archaeological data to contribute to the more general explanation of human be- havioral processes. ‘u— 0-- u I“v0c' _-,..,.. - . J'\.‘..~ MODELS OF PRE-DORSET CULTURE: TOWARDS AN EXPLICIT METHODOLOGY by Albert Arch Dekin, Jr. A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Anthropology 1975 a“ © Copyright by ALBERT ARCH DEKIN, JR. 1975 . as..-,‘. 7““ . "'¢...’.3 .c: .. V n. D % H’ (n ‘ o ACKNOWLEDGMENTS If the research presented herein contains elements of lasting value, they stem largely from the mentors, col- leagues, and students who have influenced me. The short- comings result from my failure to follow their advice or to fulfill their expectations. Above all, Moreau Maxwell has provided me guidance, knowledge, Opportunity and freedom to pursue what some may see as high risk research. More than anyone, he has fostered a respect for the variety in artifacts which form the basis for archaeological research. Charles Cleland, Charles Hughes, and Iwao Ishino have continually forced me to seek the larger picture of human behaviors and have ikept me from seeing burins as self-replicating populations. I)ouglas Byers taught me the advantage of precise archaeo- ilogical field techniques and my students taught me that Such precision should suit the problem being investigated. iElmer Harp, Jr., kindled my original interest in the Arctic and has fanned its flames ever since. The institutional support and cooperation of the National Science Foundation, Michigan State University, The State University of New York, College at Potsdam, and the National Museum of Man, National Museums of Canada are ii . v.4.- I ' '\ .0. A I c ' 0 1. ”DO!“ A. 5". oo.~V.A t.) 0: vb i.ol’ J. .‘fl‘.A L”?! U..- (A. gratefully acknowledged. I owe a special debt to my family, especially to my wife Ruth, whose tolerance and efforts through the evo- lution of this dissertation are much appreciated. iii ‘U .._‘n.un-‘ “ \ c. .uMquO‘b" 1'1" ‘V .51... L‘ a Arctic TABLE OF CONTENTS List of Tables . List of Figures. INTRODUCTION . Chapter I. PART I. THE INTELLECTUAL BACKGROUND The Study of Pre-Dorset Culture and the Arctic Small Tool Tradition . . . PART II. THE SPACE/TIME FRAMEWORK II. Dating the Arctic Small Tool Tradition. III. The Changing Environmental Setting. PART III. MODELING THE PRE-DORSET IV. The Data -- Their Collection and Analysis . V. Pre-Dorset Tent Structures and Internal Activity Areas From the Closure Site. VI. The Arctic Small Tool Horizon: A Behavioral Model of the Dispersal of Human Populations Into an Unoccupied Niche. VII. An Ecological Systems Model of Culture Growth, Atrophy, and Stability in the Pre- Dorset VIII. Discussion: The Conceptual Setting . IX. Recapitulation. BIBLIOGRAPHY . iv vii 79 108 164 201 238 263 304 322 336 ‘0‘ . : I .¢,.: V \"-":'0' ‘ .' m~.v ‘_ I ‘VNO': - -\ dumb--~ n ' \1.~2*" _ .' UMQ“' . u . - ”Vabn .fi Al“‘-» ‘ a .. 3 A '- . ' UAJga. ’ I .n“ , i. 4 Table 10. ll. 12. 13. LIST OF TABLES Page Summary of ASTt Field Research: Western Arctic. . . . . . . . . . . . . . . . . . 45 Summary of ASTt Field Research: Eastern Arctic. . . . . . . . . . . . . . . . 72 Radiocarbon Dates--Western Arctic ASTt. . . . 91 Radiocarbon Dates--Eastern Arctic ASTt. . . . 92 Radiocarbon Dates from Pre—Dorset Sites Analyzed in this Thesis . . . . . . . . . . 99 South Baffin Island Pre-Dorset Radiocarbon Dates, With Adjustments . . . . . . . . . . lOO Tables of "t" and Probabilities of Significant Difference Between Radiocarbon Dates of Table 5 . . . . . . . . . . . . . . . . . . 101 Summary of Eastern Arctic Environmental Changes, 4000-2500 B.P. . . . . . . . . . . 159 Minimum.Sample Sizes to Insure Finding One Item From Populations With The Following PrOportions at the 90% Probability Level. . 191 Statistical Tests of Significance of Artifact/Rock Distributions . . . . . . . . 228 Measurements of Selected Chert Artifacts, ASTh . . . . . . . . . . . . . . 252 Metric Parameters for the Sample of Burins From the Closure Site . . . . . . . . . . . 253 Metric Parameters for Burin Samples Within the Closure Site. . . . . . . . . . . . . . 253 . L -coU " I "V" onv '1 i... -C_I a ~v- no” 3..t,v v1 ' Li) 'Y) ,.. Table Page 14. End Blade Variety in the ASTh. . . . . . . . . 257 15. End Scraper Variety in the ASTh. . . . . . . . 259 16. Variety in Sarqaq Assemblages. . . . . . . . . 297 17. Variety in Pre-Dorset Assemblages. . . . . . . 298 vi q. t“"v'- - 0...... .1 L‘J aDVfi-Y‘. a; A. ".=- 8"-“ n... “‘d&‘ ‘ I I 9" A ‘ :I.“'un. ‘. *3- “ . "M. ‘VM 3: I l L A v v SL.‘;~. U Figure 10. 11. 12. l3. 14. 15. l6. 17. LIST OF FIGURES ASTt Sites from Table 1. ASTt Sites from Table 2. Portions of the Suess Curve for Tree Ring Calibrations . Probabilities for Values of "T". A Provisional Chronology of the Arctic Small Tool Tradition . . . An Adjusted Chronology of the Arctic Small Tool Tradition . . . . North American Arctic--Tundra and Sea Ice. Mean 500-mb Contours (in Km) Surface Currents of the Eastern Arctic . Distribution of Arctic Char and Lake Trout . Distribution of Arctic Cisco and Lake Whitefish. Distribution of Arctic Grayling and Longnose Sucker. . . . . Distribution of Musk 0x. Distribution of Ringed Seal. Distribution of Bearded Seal . Distribution of Harbour Seal . Distribution of Harp Seal. vii Page 46 73 94 98 104 105 114 116 119 124 125 126 127 131 132 134 135 “5 1a [.11 “Q .1. (.4 D .c g_, kg) LA) to &.l' LA. 0‘ LA) (1‘ A - J \u-hrga.‘ DUAAD'.“AO‘ a . \b..n-' A Vuau‘ VI C I A :‘Q 3 “94C. C "‘.. A 5-13.. 3 A""v--' ..ch"b‘ . \::‘en "- ‘v -\ ‘ .. “NJ: 73 ‘ 5. ‘V . ‘n'fl 1‘ Ag-“ ‘ ““\‘ A“): 11 L t ‘ 9’ Q l~", . ‘ ““4 . ‘L K“ 4 A “J: 1‘ A L- Figure Page 18. Distribution of Walrus. . . . . . . . . . . . 137 19. Constraints on Archaeological Data for the Study of Human Behavior . . . . . . . . . . 166 20. Probability Curves for the Absence of Artifacts With Specific Random Sample Sizes and Population Pr0portions. . . . . . 188 21. A Model of the Nested Hierarchies of Units of Human Behavior and of Units of Archaeological Analysis, and of Their Interrelations . . . 195 22. Nested Units and Levels of Analysis . . . . . 197 23. Lake Harbour and Vicinity . . . . . . . . . . 203 24. Kqu-ll, The Closure Site . . . . . . . . . . 204 25. Key to Symbols Used on Distribution Maps and Profiles. . . . . . . . . . . 210 26. Kqu ll-B . . . . . . . . . . . . . . . . . . 211 27. Kqu 11-10. . . . . . . . . . . . . . . . . . 212 28. Kqu 23 . . . . . . . . . . . . . . . . . . . 213 29. Kqu 11-6+30. . . . . . . . . . . . . . . . . 214 30. Kqu 11-7 . . . . . . . . . . . . . . . . . . 215 31. Kqu 11-8 . . . . . . . . . . . . . . . . . . 216 32. Kqu 11-6+50. . . . . . . . . . . . . . . . . 217 33. Kqu 11-6+50, Plus-Minus Profiles . . . . . . 218 34. Kqu ll-6+50, Left-Right Profiles . . . . . . 219 35. Kqu 11-8, Ellipses Constructed . . . . . . . 223 36. Kqu 11-6+50, Ellipses Constructed. . . . . . 224 37. A Mbdel of Pre-Dorset Tent Structures From The Closure Site. . . . . . . . . . . . . . 227 38. Igdluluarssuk . . . . . . . . . . . . . . . . 235 viii Figure Page 39. Description of Measurements taken on Burins. . . . . . . . . . . . . . . . . 255 40. An Ecological Systems Model of Culture Growth, Atrophy, and Stability in the Pre- Dorset . . . . . . . . . . . . 270 41. A MOdel of Vectors of Environmental Change . 275 ix I: is a 3;. ‘0 . he-‘.. Cu, ‘1‘ ‘ers. .. ‘ “ *.g (I . fivco.“ A...“ '.‘-s I . LUQJ S . ‘.“:~: . ‘H "5&0: m -... . ~‘~-.§IAA r5‘»5‘c fl: '5; L quAA . c u .. _ ‘ - “H.101“ 0 Cr- .. .1; . A‘.‘ I p““ ac“ .\.-1 5 Vb “.4 C t " E2, 1 ‘ “' ‘ 9 V . A“ A a H M' tad“... .f\b»lon arid I a.“ “Eat of 03.. “- .‘rt -‘Y ' . ‘3“. ' "a“ «\J‘ nta1 INTRODUCTION It is a difficult task to assess the present state of our understanding of Eastern Arctic prehistory and to evaluate an ever increasing number of alternate hypotheses. The last decade has demonstrated that we are learning more and more about less and less, as increasingly sophisticated methods of analysis are applied to smaller and smaller data sets. We are faced with increasingly diverse interpreta- tions of the same data and a corresponding inability to formulate research strategies which would reduce the inter- pretive chaos. There is a general lack of explicitness and precision in the conduct of archaeological research in the Eastern Arctic and it is difficult to determine the sources of variation in archaeological interpretations. The lack of explicitness and precision has led to the selection of data and analytical techniques based on im- plicit assumptions and biases which are not made available for evaluation or discussion. Thus, there is great varia- tion in: excavation techniques; criteria for data recording, depiction and description; amounts of data published in support of conclusions; methods of analysis; criteria for the evaluation of radiocarbon dates; and data on the en- vironmental setting. Since the criteria on which these l 3? LVSJ ‘ vv- ‘5‘ 325 -o 0' sad p-Iann . ..'O‘: 0 tag- --J~ ‘ c a» v. C. —4 c.“ ‘fisé- ‘5“ q H. b \ 5 Intere 4 V»; 2 selections are based have not been made explicit, differ- ential selection is a significant source of variance. This lack of methodological specificity has plagued much of archaeology, but we cannot continue to use implicit and imprecise methods if we wish to make non-trivial contribu- tions to our understanding of human behavior. This research has two goals: 1) to demonstrate sources of variance in Arctic archaeological interpreta- tion resulting from differences in the data, their collec- tion and analysis. These differences often result from a lack of precision and explicitness. 2) to model Pre- Dorset behaviors at three levels of abstraction, using ex- plicit and precise methods of data collection and analysis. My interest in this problem stems from recent research and historical studies which contributed to my conclusion that we had not increased our understanding of the Eastern Arctic prehistory during the last decade, and that we were beginning to learn more about less. The recent revival of concepts once thought to be of limited utility was further indication that our studies were not building a body of understanding on which to base my own (and future) re- search. It became apparent that one reason why we lacked sophisticated understandings of these data was that we lacked any rigorous treatment of data, method and theory. This thesis is a demonstration of the impact of im- precise and implicit archaeological techniques, methods, I i V: fl 4 D.‘ Ogu IIS 0 sub v“: 0 fit ‘9 O Q :3: rreater exp. 3 1 ' g Oughab. q “_‘ "'-V‘d\vb in .Au :13 :33: tne anal : :12 may take :L‘h..“ H ' f pane .r.e ...:,-~_ .~ ‘ ‘ .~:.51‘ ‘8 H An .- -.. “V i:"~- t. . .u: 'dS-er" O6 3 and theories on the present state of our knowledge, a plea for greater explicitness and precision in archaeological research, and a demonstration of such methodologies in the formulation and testing of models of human behavior at several levels of analysis. By making the treatment of these data as explicit as possible, others can see where and how the analysis proceeds, and why. While such a pro- cedure may make this study more open to criticism, it should pave the way for fruitful discussion of archaeo- logical methodology and place archaeological explanation in the Eastern Arctic on a more firm foundation. PART I THE INTELLECTUAL BACKGROUND Chapter 1. The Study of Pre-Dorset Culture and the Arctic Small Tool Tradition The 3:; w... H‘“n;." . '5': C‘....._,_ .0.- The Study L: .~a Ffifioc ‘ ‘-¥b- u 1‘ Q».... N ‘ a r 4 ;:":‘:I“' JDCM " J «7.1, 5' .“ UAU:§‘ 5 ‘..e 1". .iv. B.P. Ip- .'\ “AA.- ‘.ze CotAN-c- ..‘,A.‘ i d I fi A::: ,‘ltt L u“‘. :9. ‘ "N... 0 ~ - ~u..~..33. ‘ne ’ ' '2»: p : Nona“ ad am | 3“}, : A Q t .45.; yabt art: A... ['3 :Ha:‘ ~5u5£blv ‘0 ,. 4 ‘3‘ s Q vi; c ‘ ‘ . {4&3. f C 'F.‘_, K ‘ Ma Ade“ u.‘ u.‘ . x: A? ‘V‘lc r.-, “°fi-f Chapter 1 The Study of Pre-Dorset Culture and the Arctic Small Tool Tradition Three Concepts The study of the early prehistory of the Eastern American Arctic is dominated by three concepts and associ- ated models: 1) Paleo-Eskimo; 2) Pre-Dorset; and 3) Arctic Small Tool tradition. These concepts have been generally used and accepted as heuristic devises for ordering the widely distributed and often little known cul- tures across the American Arctic during the period 5000 to 1000 B.P. The concept of Paleo-Eskimo was suggested by Steensby (1917) to differentiate between two cultural strata of Eskimos. The Paleo-Eskimo were adapted to land and sea- ice hunting and lived in snow houses, originating from the inland Indian cultures in the Central Arctic and spreading both East and West. In Alaska, this earliest Eskimo pOpu- lation was subjected to the influences of Pacific cultures, ostensibly including Japan, leading to the Neo-Eskimo de- velopment of open sea kayak hunting and to the use of the umiak. Their descendants were considered to be the historic Eskimos. v- - 9. . ,:‘:~-‘O cvc ’.'V v'. .-.~.-d “Qufl. ' I O 1 0 sun. q-“ -—»'A&: o...‘- Laud .C‘CAI by . C1338 07.6 1:": '.-.' 4.5.;'-"" ‘ 0' DU-..‘ .A'. ‘1 D L.) :5112-3 and 3&0 --“ 3.1.263 an EX: EC: 4:22:33 is '33 t': 5 In 1950, Helge Larsen made this concept explicitly archaeological, relating it to technology: "A Paleo- Eskimo culture would thus be a culture in which chipped flint implements are preferably used, and a Neo-Eskimo culture one in which ground slate implements are pre- dominant" (1950:186). In the Eastern Arctic, Paleo- Eskimo and Neo-Eskimo have served as a useful distinction between early coastal-tundra cultures whose technology in- cluded an extensive reliance on ground slate. The dis- tinction is both temporal and cultural, with considerable spatial overlap of both stages. Paleo-Eskimo cultures in the Eastern Arctic have been defined by reference to specific distinctive cultures in Greenland (Sarqaq--see Meldgaard 1962; Independence I and II--see Knuth 1967) or as develOpmental to later cultures (Pre-Dorset--see Collins 1954b). Recently, the inconsis— tent application of these concepts as unifying ideas lead McGhee to revitalize the concept of Paleo-Eskimo to refer to those early cultures in the Eastern Arctic which stemmed from an early migration, or migrations, and which developed into several distinctive adaptations to dif- fering environments before the development and migrations of the Thule culture of ca. A.D. 1000 (McGhee 1973). Taylor had recognized the strain on the concepts of Pre- Dorset and Sarqaq in 1968, suggesting that we substitute the term "Carlsberg culture" to encompass those Canadian- Greenland cultures previously called Independence, Sarqaq, a:i ?:e-Dorset k :13 DECO-2.6 23.: -v I ‘ IA.‘ - Dug ' A. 3.053: sue bands 0 ”I ‘vnfi‘ us. 3.5.. .-.as‘.-:a as a :25. :.;S "”8 S‘\I.4 C" ‘ 6” -au- l fi‘.-0‘. . 'A‘ "" AA. ""“ A.C xx. “‘0. . 3": Earl . . , w ‘19,“ s' . ‘u h 5323‘“ S A :x. “a. ‘l 5. I F. ‘ ‘ g fi Q‘- “\‘5 : § ’“ ‘ ‘ 1.. c. l . \‘l 6 and Pre-Dorset (Taylor 1968:85). Whatever the title, it has become apparent that a new model is necessary to de- scribe the relations among these early cultures. The Arctic Small Tool tradition was first defined in Alaska as a result of studies of core and blade technolo- gies. The studies of MacNeish (1954:252) in the Yukon and interior Alaska demonstrated the presence of a series of boreal sites containing polyhedral and tongue-shaped cores and blades struck from them, several varieties of scrapers, choppers, and large projectile points. "This early in— terior northwest North American cultural pattern seems to be distantly related to another early pattern along the Arctic coast, often called the Paleo-Eskimo. . . . Both of these patterns may have developed from some as yet un- defined early (Mesolithic) Paleo-Siberian cultural com- plex" (MacNeish 1954:252). This contrast in stone technologies was also observed by Irving, who noted that "the boreal forest sites dis- cussed here seem to hang together, and lack the types as- sociated with early man which appear in the Denbigh Flint complex. They seem to belong to a line of development different from that found thus far in the Eskimo area" (1955:382). Irving formalized the similarities in tundra and coastal burin-blade industries into the "arctic small-tool tradition" (Irving 1957:47) and contrasted it with MacNeish's boreal forest sites at Pointed Mountain and the ‘A..' IQA‘ n 3:”: as Itil a: 55:23:38 OI 31". VA... _‘ ‘:°5:C..Ce :0 .L. ~..-_ 0 D (ID (I) (.1. r? O (L (T) (I -3, . 4.; 3" P v u.‘..3"| 0‘ “a i-E . Q £3er 1:1 A: . ' ‘ “35:1: 3» A.E§33r 7 Campus site. Irving's prominent reference to "early man" and to the Denbigh Flint complex reinforces the interpre- tation that this tradition was believed to have great time depth as well as extensive geographic range, being repre- sentative of an "extensive continuum through space and time" (Irving 1953:71). Pre-Dorset was first defined as a cultural entity with reference to those cultures in Eastern Canada which were believed to develop into Dorset culture, representing a migration of peoples closely related to the Denbigh Flint complex in Alaska (Collins 1954b:304). This definition resulted from the discovery of sites earlier than classic Dorset sites which contained burins, burin spalls, micro- blades and other artifacts similar to those believed to be early in Alaska. Thus, the first application of the term Pre-Dorset was as if it were a residual category produced by the timing and extent of previous archaeological in- vestigations in the Eastern Arctic. It is the develop- mental aspect of the term that has lead to its criticism (Noble 1971; McGhee 1973) and to its rejection as a widely useful term for comparative studies. It is significant to emphasize the points of reference for the development of these concepts. 1) Paleo-Eskimo. Steensby's attempt was essentially ethno-historic in ap- proach, attempting to explain the observed diversity in historic Eskimo adaptations as a result of the movements of people with different technical knowledge and u u .' ...~~Osu'\ 9‘ 3;.353uCAob {3" l H ' IA. "?V' V':“ HTML”) In: I ‘ ' ‘. ' v -; ~‘n ..::.:’:3.05.»35 C .. v. ._ :_ p... q . , ..:. .. .3.- O In 0 ‘ c. . 1 vAA.>. ,1 . ‘3'. ..a.‘.. ‘x ' a v . . . ' ;"',"‘ Il-np A. dU-ACO°.‘ . -.‘. ~\ - 5" 9 ' ~ " h- -1 .n..v.. ‘ UGDG Av. 11 54.:0: a: FEAR" an, b4 "o.l 3:25: ‘“ ‘ ‘¥-. “"Ru‘ .- I ‘ i"-A‘ 0 “‘n ‘V . b .l‘» r : §Aj~ -q-A - d :35- . y‘a 1 AK. ‘&«EI C l :I: "R . ”‘ "‘11‘5 : - 5 § A, . o: 5 h ‘A ~0 it a“; Hg ‘-‘I b Qulres of ‘.‘ *L: "Q‘a h k nae-W” . - b ‘l I ‘w‘La‘,’ “J dE‘J ‘.. ' n‘ WI nLES (33‘ 5 1 .‘Ons a. 1A,» 'A\‘$ dc. . L “Eli-Fr- a: 5 J ‘1 ~ '\ “if. 5 Cali”, . 8 subsistence patterns. The application of this model to archaeology was a fortuitous result of the attempt to fit archaeological data in support of his considerations of historic Eskimo migrations. 2) Arctic Small Tool Tradition. The Arctic Small Tool tradition was formulated by specific reference to the Denbigh Flint complex, then believed to represent early man in Alaska (Giddings 1951, 1954, 1955). The long time depth associated with the idea of "tradition" resulted not from.extensive chronological inference, but from the asso- ciation of basal fluting with precision flaking and a burin- core-blade technology, all presumed to be indicative of early migrations to the New World. Thus, the initial time depth to the Arctic Small Tool tradition came by deduction from.association and assumption. 3) Pre-Dorset. The Pre-Dorset was a by-product of the research into Dorset origins, when it became obvious that earlier cultures in the Eastern Arctic lacked many of the traits of "classic" Dorset, but were apparently ances- tral to it and were technologically related to the early cultures of the Denbigh Flint complex in Alaska. Again, this concept was formulated to describe ill-defined inter- mediary developmental stages between relatively well-known cultures (Denbigh and Dorset). Evidence from.Meldgaard's excavations at Igloolik (1962), Maxwell's analysis of Dorset development (Maxwell 1967), Noble's investigations of the Canadian Tundra tradition (1971) and from McGhee's v€««1aissa::e O . 5v... ~n~~q‘\‘ "‘ . m" 5.13:3. 5" ' ”J W ' ‘ya-‘Av‘ ‘ 9‘ “o. A I r‘ a.;. c-t.~~~~ A ‘1‘. ’ "I :‘1" J::.. .00 ¥¢JC" ‘ 3 .- 41.va —-“‘F A»: vC'CbOD—nso-bd ”9.1., : . .:.-.: tilt-u.b' LAAMLuuCu C .1. .‘Vfl'vy: ‘ . av. :.a\.".:g tr: :. . fl...‘ P p‘ a 3 6.9 01 I; ‘i 554'. M,:V§ALESES r9,“A 5.1,: ' .3 S . ‘ 5::~. \;~~ ~~b of .L I w “‘5 Jéw" 1 x. N 9 reconnaissance of Devon and Dundas Islands (McGhee 1973) suggests that much of what we have considered Pre-Dorset is not developmental into Dorset and thus the term has been too widely applied and has misleading implications. The developmental model does not apply to all that are fre- quently included under its wing. It is apparent that the conceptual schemes and models for studying the early cultures of the Eastern Arctic are in a state of flux, lacking general acceptance or utility. The use of several of these is anachronistic and the models which they represent no longer fit with the available data. One of the purposes of this analysis is to develOp models which will reflect the present state of our knowledge and which will be useful in the generation of testable hypotheses regarding cultural processes. An Historical Perspective Because the substantive data on which this study is based are from the Pre-Dorset of the Eastern Canadian Arctic, the following discussion of the historic back- ground to this study will regard Pre-Dorset as a distinc- tive variant of an Eastern extension of the Arctic Small Tool tradition, reserving until later the evaluation of the appropriateness of these concepts. The study of Pre-Dorset culture was spared the early stages of the deve10pment of Arctic archaeology (see Dekin 1973a:15-21). The earliest published finds of what . In. --\‘\ wage fiaber r A O . . a. Abbaqbfifiq 9h: 0 urttotbfivtl 5.5L 'f‘ V’ :5“, v.4. .a.::." .5 .O‘tu ' ' "4 ..: :. 3:9 0: C “Vb, .h:.5l“. V ‘ Ma. a~ e M " J. by ”- ‘ I --Q “ Q :.'tm3.0rv an 4 4' 6“ A ~:-h,“ m .u,‘.v _ _ 4' Ana I :‘A' ‘0‘ A».CA 6::‘OrEQF V; I Tat: “‘5.‘C av. ‘N 4:: a1” " A J 5:e ' n‘. C" ' I . J 1» 'LI (1) I -- 10 were later recognized as Sarqaq artifacts and Independence houses in Greenland were conspicuous by the relative lack of attention that they attracted. Solberg's early attempt to establish a stone age culture in West Greenland (1907) lay largely unaccepted and generally regarded as icono- clastic because it conflicted with the views of the Mathiassen establishment. Thus, his contribution lay largely ignored for almost fifty years. Thostrup pub- lished one of the first series of mid-passage houses from Northeast Greenland (1911:194, 195), but they were ac- corded no great age. Pre—Thule culture in Greenland was unacceptable to the general understanding of Greenlandic prehistory and thus almost a forbidden topic for half a century. The link between these early studies and the later efflorescence of research is tenuous at best. Arctic archaeology prior to 1950 could be described as "Boasian" in that the emphasis was on the collection of data and the interpretations were built by "letting the data speak". This emphasis on empiricism and on an in- ductive approach to analysis has a long history in the Arctic, persisting even to the present day. The major analytical tool used for comparisons was the trait list supposedly defining trait complexes, which formed, through time, traditions (as trait complexes with time depth). Perhaps the greatest conflict during this period was be- tween the migrationists and the diffusionists, although neither group formed a cohesive faction. We must keep in . 1 ' ' ' ‘... run? H. ”0‘ ...Mb -.L d5 .3 AQQ V ‘L r ‘_‘e 3 ~ 533:“. 5" ll mind that this pre-1950 period in Arctic archaeology was before radiocarbon dating and that culture history was built on developmental sequences of technology with little regard for theoretical discussions regarding the nature of ethnic groups or cultural systems. "Culture" was almost synonymous with "trait complex" and the literature is con- spicuous by the imprecise use of the concept "culture." While this terminological conservatism kept the archaeologists close to their artifacts, several were quite liberal with their use of artifacts and artifact types as guide fossils, lending significance to the pre- sence or absence of certain apparently significant arti- facts (such as microblades) from collections of minute size (often as low as 30 fragmentary artifacts created as tools). Credit for initiating the modern stage of our studies of Pre-Dorset cultures must rest with J. Louis Giddings, whose excavations of the Iyatayet site at Cape Denbigh led to the establishment of the Denbigh Flint Complex (Giddings 1949, 1950, 1951). The presence of almost ex- quisitely flaked side and end blades with presumed "Old WOrld" forms of artifacts (burins and microblades) and several projectile points with basal thinning reminiscent of "fluting" then presumed to be of great antiquity in North America, led Giddings to suggest great antiquity for the Denbigh artifacts, believing them to be the products of "Early Man" in the Arctic. His early papers make Q 0 . 35510215 [313: 7.: " JRVVA‘fifi" -v~ 8'2: it'toV:o-CA. ' O o '14:: Afiy". .‘ 3. Anb.‘ g..._, u p O O 2::5 were 50;}: /:.Q . . ’- ... " V g . \VVokb.\‘ ‘131 ' u" V ,1 . N H.“ N §9JL .“O 3.1‘ 5‘ A "sle b eVAjI-‘WA 12 obvious that he regarded this complex as much earlier than even developmental Eskimo culture and his initial rejection of radiocarbon dates of three to five thousand years be- fore present (Giddings 1955) confirmed his intent to search for evidence to establish such antiquity. Almost immediately, other sites with similar arti- facts were found in interior Alaska in Anaktuvuk Pass (Solecki 1951; Solecki and Hackman 1951; and Irving 1951) all of which were considered representatives of an ap- parently early complex. The Eastern Arctic was also shaken from its state of complacency by the finds of Hans Mosegaard from Sarqaq in West Greenland. These were reported by Jorgan Meldgaard in 1952 who suggested that . .the Sarqaq material can be interpreted as evidence of an Eskimo culture, closely related to the earliest Eskimo cultures in Alaska, ‘which appeared in West Greenland after wander- ings without lengthy stOps; i.e., without de- velopment of local types in the eastern areas, contrary to what happened to the Dorset culture (Meldgaard 1952:299). Meldgaard saw the Sarqaq artifacts as the remains of an Eskimo culture related to the Denbigh Flint Complex and later cultures in Alaska, but not directly related to the then fairly well known Dorset culture, which was believed by Meldgaard to be at least partially contemporary with Sarqaq (1952:229). Dorset culture sites had been found across a wide ex- panse of the Eastern Arctic from.Newfoundland (Harp 1953) "’\'..'.‘2“' 1" :: .Iv..ut‘:v 9‘ . u . a VA! :1 :7 s W: aQUA. LU: - n 1 3 “sz van!!! a I- . Jan.:5‘ aAAC t-Ad 4 I A ' I O‘D'I'A flt‘finhu' v ”a: o:.3.-\t.ab... " r 11. - V 1 ntrioV b<)L . c ' v0 I . . ’-' 3"” ~~ .9. v I '“' ‘ ‘ uAA o_~.,. ‘. ' o :.‘\4 -. ’AA. v.99“ e the .. . . m... "‘~ nona- ‘ DAle b ...‘ u; 451.. also 51““13 C3 36:512. :aZE‘: 13a. ~?3l, 1§:. tnerefore L b :‘IE Source f. ”I Q ’ “.012" o. _ A . DC‘.’ “a H I "wing '“ U I a s m a au- ‘ dE-HEES::V 5 S l ‘ “Se Ir 4 c-fis ‘ ‘1 fi“ re - I 13 to Northeast Greenland (Knuth 1952) and these were seen by some archaeologists as being related also to both the Sarqaq and Denbigh sites, although the exact nature of this relationship was unclear. Henry Collins summarized "Archaeological Research in the North American Arctic" (1954b) in a paper which has stood the test of time and is remarkable for the degree to which the conclusions remain generally accepted. His re- marks on the relationships between Denbigh, Dorset, and what he called pre-Dorset bear quoting. On the other hand, there are indications of a cultural connection, despite a great time gap, between the Denbigh Flint Complex, pre-Dorset and Dorset-like cultures in Canada and Green- land, and the typical Dorset cultures of these regions (Collins, 1951, 1953 a,b, 1954a; Knuth, 1952; Meldgaard, 1952; Harp, 1953). There are also signi icant resemblances between some of the Denbigh implements and those of the much later Ipiutak culture (Giddings, 1951; Collins, 1951, 1953b, 1954a; Harp, 1953). It appears, therefore, that the Denbigh Flint Complex was one of the sources, perhaps the principal source, from which Eskimo culture developed. Though the Denbigh Complex and later culture stages related to it seem to have extended from Bering Sea to Greenland, it was not entirely, or perhaps even primarily an American phenomenon. Recent reportings by Russian archaeologists have described Mesolithic sites in Siberia con- taining burins, lamellar flakes, and other stone implements like those found at Denbigh. These Siberian sites do not stand in isolation; rather, they are part of the Eurasiatic Upper Paleolithic-Mesolithic continuum. This sug- gests that the pre-Eskimo Denbigh Flint Complex as known in Alaska may eventually be revealed as an easterly extension, on American soil, of a widespread Eurasiatic culture of Mesolithic age from which the earliest forms of Eskimo culture were derived (Collins 1954b:298-99). Both Collins and Meldgaard had recognized the general . .I 0 ' m ...n nv‘o‘ : ‘ 3‘-.3obv‘es w"' . '.".1 3 ..T ... a.n:101 C.1-. - I P I 'a-- ‘- AI. ’5‘, n,- ‘:.a A. Vb 'AOebt . 5:. T. a “h A . , , hf IB::aI\ I um U- "..' ,. f v" ”3:,‘c‘sn art; F $122M): Ste a N‘ '33 -‘-“ i "1 n‘ns' A.‘ g 5‘ 4. fi ‘ 7‘s: §":‘ ‘5‘): 0|: *' ‘ F 4: "‘Ectl‘] r0 :; r ’13! ".~'- -v R “M‘1‘ ‘Hgs Ir u i. 5 .:~. ‘L‘.Q a V d .‘J l4 similarities among artifacts across the American Arctic at a level of culture prior to the Thule horizon, but the re- lation of these cultures to later Eskimo cultures was un- known. The researches in the interior of Alaska and the Yukon by MacNeish and Irving led to the distinction between two types of core and blade technologies, one apparently tied to the boreal forest and the other to the tundra and the sea. As a result of Irving's researches in the Brooks Range (1954, 1955) he pointed out some of the differences between early industries of the boreal forest (e.g. the Campus site and Pointed Mountain, N.W.T.) and what he prOposes now to call the "arctic small-tool tradition", represented at the Denbigh type site and sites in the Brooks Range (Giddings, 1951; Irving, 1953, 1954) (Irving 1957:47). However, Irving did not follow through and specify the de- fining characteristics of this tradition, and it is clear that its definition to Irving was as contrasted with other Western Arctic core and blade sites. The initial categor- ization of the Arctic Small Tool tradition was the lumping of Denbigh-like sites in Alaska, and did not include any non-Alaskan sites. Giddings acquired a small collection from the Thyazzi site in northern Manitoba, which he saw as relating directly to his own Denbigh Flint complex in the West, and to sites in Greenland, Alaska, and the Siberian Neolithic (Giddings 1956:266). Giddings was apparently cautioned by negative evidence from creating any larger categories of o .. :: :0. V ‘ 9 a - S. a ’ L..o\'"‘5 ‘ ' g 1 ‘| -‘,As pucrch- : -C , -- . v Q . "I-" ‘ five" r0 9 v tuurbbb V . #v .Ah- I 1:2 I 17.638 CI ' .. In a: :1] I326, L; :' In . ‘ao ‘ I?! 1'... “.--‘sua-- .‘ \, U:F\.n: u “.1... :13: a b.‘ “315...; g: A - “ Scov . ., s:‘“ "It-u ..‘ ‘N‘. I . h.‘ & \" _‘ §,- LO 5 D: O" U I A.‘ ..‘5 C “c , A v 15 ". .useful to speak of sites as 'burin sites, finding it sites,‘ 'microblade sites,‘ and 'side blade sites,’ with respect to emphasis on (not merely the presence of) one or more of these distinctive technics" (Giddings 1956:266). At any rate, Giddings demonstrated the extremely widespread distribution of sites with similar artifacts, and hinted at the possibility of tracing diffusion of technical traits through time and space. ‘MacNeish also filled in a gap in the distribution of these sites by his excavations in the northern Yukon at the Engigstciak site (MacNeish 1956), where his New Mountain complex was seen as related rather closely to Denbigh and similar sites in Alaska as well as to pre- viously discovered sites in the Canadian Arctic and Green- land. "Thus it may well be that Early New Mountain and Irving's Brooks Range material are ancestral to other Arctic micro-tool cultures" (MacNeish 1956:100). Meldgaard's report on Mosegaard's collections from Sarqaq precipitated a renewed interest in Greenlandic archaeology. Knuth initiated a series of excavations in Northeast Greenland where he at first believed he had found a variant of Dorset culture. His subsequent expedi- tions produced evidence for two cultures believed distinct from others found in the Eastern Arctic, Independence I and Independence II, withthe remains dated to four thou- sand and three thousand years before present (1958:570). The larger size of the Independence I lithic artifacts, . - ' ‘ ' 0- n1 ‘.' daao t.“ 5455 0‘ 5 .uu 0‘ the rename or: q I '- L.-;.. .‘ uan ‘ '3 a L U- ...vv 5 .Ql 7 .arsen an; ,3" . ~33: excavatf O O .5129 OI the 4. 15:5: )0: fl,‘ Dt‘ () T M" I- 44...; for 33: II. ‘5 'v-C QM “.t“ at”! \J ‘u ‘ ‘V-‘: 16 their lack of grinding, the unusual mid-passage houses, the reliance on musk-ox hunting, and the presence of micro- blades all indicated differences from other Arctic cultures, while the lack of slate, grinding, and stone lamps dif- ferentiated Independence II from Dorset (1958:572-3). Knuth.made it clear that he characterized ". . .Indepen- dence II--which may have several phases in the northern area--as: pre-Dorset" (Knuth 1958:573). Larsen and Meldgaard (1958) and Mathiassen (1958) con- ducted excavations in west Greenland, confirming the exis- tence of the distinctive Sarqaq culture as well as the later Dorset culture. Using microblades as a horizon marker for Dorset, they forced the distinction between Dorset and Sarqaq at numerous sites in Disko Bay, where stratification at Sermermiut was used to confirm this dis- tinction. Radiocarbon dates on these two cultures indi- cated an age of approximately 3000 years for Sarqaq (Larsen and Meldgaard 1958:40) with Dorset cross-dated by stratigraphy and similarity to other dated sequences at just before 500 A.D. (1958:24). Harp conducted a survey of the Coronation Gulf lit- toral in the Central Arctic, where his Dismal 2 complex near Dismal Lake was recognized as relating to both the ‘western Denbigh-like sites and to those of Eastern Canada and Greenland, where Harp saw Dorset developing from a long and complicated cultural continuum of microlithic technology (Harp 1958:247). p I A 'N ' ' u ‘. ' ‘ ' v'. S..ea '1‘-“ ..— . u u - l'h-QI“ '5“ OJ- nAob-OV AA“ is. . . a .I...‘ of a A... h‘ . p-51. b 5-... “’"W‘as. .. .".'ol3unb . . , v . . :5" a vv-p.‘ 'g. l ."" H ..-|‘ .A. p I U ' o ,Q'V:~‘~-~.. .- § n.‘y.\c-3.. ‘_‘ ' It "'A w v I'F—E",\v .u, ”6.4:...“ _ In. 3.3311 m9” (1‘- ( -=.~é-'e as 2a: V. ‘ I ‘n A. 1 ...i R I. ‘Jc‘sh “an: 17 By the late 1950's, the increasing amounts of data frmm sites believed to have some relationship with the IDenbigh Flint complex extended across the entire American Ixrctic and the stage was set for some attempt at the syn- thesis of a somewhat more general picture of cultural re- ;Inationships. While Irving had made an attempt in this di- rection with his Arctic Small-Tool tradition, it remained ;1E<31rMacNeish.to spring into the void with his "A Specula- ‘t::ifive Framework of Northern North American Prehistory as of April 1959" (1959). While warning the reader ". . .not to believe as facts all that you read herein" (1959:l), MacNeish went on to carve the disparate data from the Arc- tic into a number of complexes and traditions. Of inter- est to us is his conception of the Arctic Small Tool tra- dit ion, as it marks the first formal statement of its characteristics. During the latter part of the deve10pment <3f the Northwest Micro-blade tradition, a new ‘tradition appears on the Arctic coast, called here the Arctic Small Tool tradition (Irving, ‘1957, page 47, footnote 4). Characteristic of 'this tradition are burins with chipped surfaces, 'burin spall tools, cuboid (and conical and tabu- lar) polyhedral cores, micro-blades (usually not ‘retouched), ripple-flaked lenticular, lanceolate .and triangular end-blades for arrows (or har- ‘poons), antler foreshafts for arrows, delicate, small neatly chipped half-moon side-blades often ‘with ripple flaking, ovoid, semi-subterranean houses with specialized central fire place (often outlined by boulders), and an economy based on caribou hunting but supplemented by a little sea- 'mamma1 hunting. The earliest manifestation of this tradition is the Denbigh Flint complex (Giddings, 1951) at the Iyatayet site on the Seward Peninsula of Alaska. Carbon-l4 dates indicate that this is not younger than 4,000 . ”>0 $07.31. CC»... '3. 31‘. 3.3 .". JC‘JEC area 0 .u A 1’: ‘. A .‘oy- .. a ' R Q ova-av ‘\ «A .4 e D u». vat» u T a :«1 ”no!“ Du): ho'.“L 7 an... "' .Lc...5. . 1- n.1' A - ‘— ‘I-l S4..-” . I ’RA" . 5.4159 5- O." :' 9 .1 '~ n in." 1.:‘c.- VI“. Ov- ' ~-u'~.€ “(1., U "‘I" no- - boa.) O“b a: VA-uqba: "~‘ht.d.. ' L ,: d. “on ad l 18 years ago (Rainey and Ralph, 1959) while sea level datings hint that it probably also is not older than 5,500 years ago (Giddings, per- sonal communication). Recently Giddings found on an old high beach level in the Kotzebue Sound area another manifestation of this cul- ture. It also appeared in the Brooks Range where Solecki (Solecki, 1951), J. Campbell, and Irving (Irving, 1953) have found similar remains. These remains from the Brooks Range I am calling the Itivlik phase. Actually some of these sites found by various archaeologists may likely represent different stages of this single tradition. However, no one has worked this out as yet. On the Firth River, the New Mountain phase (MacNeish, 1956 and 1959), esti- mated to be about 4,000 years old on the basis of Carbon-14 (Rainey and Ralph, 1959) repre- sents another part of this tradition. Here the Birth River stage with fabric impressed and cord- Inarked pottery, and the Buckland stage with den- ‘tate stamped, grooved and cord-marked pottery, :represent still later phases of this tradition (MacNeish 1956 and 1959). In the Coronation (Gulf region the Dismal II component (Harp, [1958]) are of the same tradition. At the .AAlarnerk site near Igloolik, the two earliest EStages which might be called Alarnerk I and III (Mbldgaard, 1955) dated (Rainey and Ralph, 1:1959J) as between 3,900 and 3,000 years ago, ‘Iflepresent a development within this tradition as do the Independence I (Knuth, 1958) and the Sarqaq remains (Larsen and Meldgaard, 1958; 1"Iathiassen, 1958) of Greenland. The latter has been dated as from 3,500 to 2,500 years ago (Larsen and Meldgaard, 1958). A few arti- facts from the Button Point site in the tranklin District (Mathiassen, 1927) and from tZlhe Nuvuk site in the Ungava Peninsula trojectile points (and ceramics) from the ‘Dnliddle Lena (Okladnikov, 1955) and the akitikiveem site (Krader, 1952) from the in- 1Clerior of north-eastern Siberia suggest (if the Russian dating is correct) that some of the elements of this tradition were derived Irom.the interior of north-east Asia. The IImicro-blade industry may have come from the (arth-west (interior) Micro-blade tradition already in North America, as might the Yuma <3111pping technique. The tools adapted to same 53: : .2. x' v .. I'd -.—e 0‘ 3.11 spas? were a se 1 L“. Nor: :raiitian, anti: A: .. ‘ .nn-quv ,- ..~,.‘ -~.:'CO .0 . J ‘ .- , J" .Vfidop. " new”, 4. " It. ‘- LS C13: .1- aV‘Q‘RA‘A --.‘g\'~b ol— 1 4 .qe a?” s .La,‘ v b ~~==e .ra~ U flava.’ 5:.:.S:, a A~:'P’: ."“HCM ‘. ‘1“ ., - uutb vn‘ me “a“ o ~ , a -‘ 5‘ . U‘Sbgnc b c ' "\. u ,b' 035:5 m.“ "N : .‘a- id" 0. ll: ~;1 ‘ 5&1 Otne l9 marine subsistence may have derived ultimately from the North Pacific tradition, which we will speak of presently. Here again is a case where a series of elements seem to have piled up in North America to form a new cultural tradition, and then moved as a unit across the entire Arctic and persisted in time (MacNeish 1959:8-9). However rough the boards of MacNeish's construction, he made clear that he was talking in generalities, even with regard to his use of the concept of tradition. By tradition I mean a distinct way of life as it is distinguished by different complexes of .artifacts or diagnostic traits that persist in 'time and space. The exact process by Which t:hese traditions originate, spread, change, 13ersist, and finally disappear, cannot be de- t:ermined by the present evidence. Some of t:hese traditions may reflect actual migrations c>f distinct people with distinct cultures; c>thers may result from the readaptation of a ‘vvay of life from one ecological zone to another; satill other traditions may derive by combina- t:ion of all the aboveamentioned processes as ‘vaell as many others not mentioned here. Be that £33 it may, the origin, spread, persistence and (disappearance of traditions seem to be a com- ]plicated process. However, in spite of this, the concept of tradition seems useful in de- Ilineating cultural relationships in time and space in the north (MacNeish l959:2,4) . meNeish's use of the tradition concept is first and fore- uu3£31:_ as a technological tradition which, as he points out, may be characteristic of one or several groups of pe0ple. PeZ'T‘ltlaps his greatest deviation from the concept of tradi- tion developed by the 1955 Seminars in Archaeology (wanehope 1956) is in his failure to restrict the spatial dimension and emphasize the temporal dimension (Wauchope 1956: 38-39). It should be noted that this tendency to discuss the ¥ u. " . - ‘u-O" "" “ 50:». .DJ'-V..3 tees) assuz; . u H 9 ..-- a p-v fi-I :‘b.:. bd-.i a p A. .5.‘ 09 Lvn v. Jan. A AAA; - c . 0-91-‘~. a. v... '--‘V‘5.- but so: u .o , ""‘ fl~n‘..7 h... 81:..' ‘ o .1 , OV' WIOI I~o.i..3:s :3 Q0 4' .- J. 'I. .1! 0““‘V -‘.'" ' ..D . ~ ‘t .|": ‘ ”.““') CO -..... 0 an 00 3 9.. . .‘ ~ 44"“ ‘- “\“u&3 .n. Ins . £ 5L Cons: “é :1. ,.. (J ( a) '1 U‘ C) ,J l. r f' Y) 20 distributions of artifacts and sets of artifacts (com- plexes) assuming that similar artifacts are indicative of similar "cultures" or similar behaviors was characteristic of most of Arctic archaeology during the 1950‘s and even through the next couple decades. There was a tendency to drift easily from the discussion of specific artifact dis- tr ibutions to the extent of complexes to the migrations of peoples and to the spread of cultures. MacNeish's 1959 Paper demonstrates this tendency rather well, as he dis- cill-lsses specific characteristic artifacts, sub-stages of a Culture, complexes, phases, cultures, elements, horizons and traditions as well as a culture complex. This termin- °1°gical imprecision was perhaps indicative of the rather skeizchy and widely scattered shreds out of which he was tryiang to weave his synthesis, but it is characteristic of the times that the major concern of Arctic archaeologists Was the construction of a space-time framework using what we might call site-occurrences as the data on the frame. Rad iocarbon dating was a newly found tool, and one whose 11 . Se in scattered areas served as props for cultural chro- no l<>gies across the Arctic. The "type fossil" or "guide £08 sil" approach became the accepted technique, where an a rtifact type dated in one locale was assumed to date at a s - imllar time wherever it occurred. Traditions were formed ‘37 ‘ lthout much evidence of time depth, and migrations (and 1‘ a‘rely diffusions) occurred with some alacrity (Larsen and M eldgaard 1958:71). e v'. . twat“- 3- "\r} .6: a ‘ub'us.\ 3?. Arc‘:.ae3‘.: .P a! :VOQO A. ‘ “‘b‘. 5. S '_l' (Iv p) (I) 5 I a 4 21 This pattern of thinking in Arctic archaeology has set the groundwork for research undertaken even in the present day. Archaeological problems were still largely confined to the development of a space-time framework for artifacts or artifact sets (however named or described). Chronology building was an adequate research goal. Perhaps the greatest shift in the use of concepts was in the greater use of a more precise concept of culture, as we became more aware that artifacts were made by people, and that these people formed groups and had characteristic ideas regarding technology and tool making. Even so, much 0f the research conducted during the last several decades is as if we were studying entities which were spread by life processes and which were perpetuated by genetics (see De—ki‘n 1973a:4l for further explication of this point). It is almost as if conceptual precision has been deeuled unnecessary and luxurious. Fortunately, the re- newed concern for theory and method that has been racking North American Archaeology in general is beginning to reach the Arctic, as the recent discussions at a School of Ame11‘ican Research Advanced Seminar on Pre-Dorset--Dorset I: o1>lems indicated. Shortly after MacNeish's pioneering attempt to bring a. Semblance of synthetic order out of prehistoric chaos, no Symposia were held and collected papers were edited by C ampbeu (1962b) and Hadleigh-West (1963). These collec- t ions are significant watersheds in the development of our k _4_ 22 thinking in that they caused the widespread sharing of information and allowed the presentation of up-to-date ideas to a wider audience than existed in the informal communications systems. Meldgaard reported on his excavations at Igloolik where an extensive series of raised beaches were used in conjunction with a program of radiocarbon dating to pro- duce a chronology of cultural changes in Pre-Dorset cul- ture leading to a marked chang ca. 1000 B.C. when Dorset culture apparently replaced Pre-Dorset (Meldgaard 1962: 95) . Meldgaard suggested that this change was caused by the local disappearance of Pre-Dorset people and the migra- tion of new Dorset people stemming from somewhere south of James Bay (1962:95). Taylor reported the results of excavations at Ivugivik in northern Quebec where three small sites represented a Si‘i'lgle stage of Pre-Dorset culture (1962:81). Two ground Stone artifacts and several flakes of slate were regarded as possible intrusions from a later Dorset occupation (in WhjLQh similar artifacts were prevalent), once again follow- ing a "guide fossil" type approach (Taylor 1962:88) . Ta)? 101: considered these finds as indicative of placement efa-I‘ly in the Pre-Dorset continuum and he remarked on the near~identity in burin forms between his Ivugivik speci- “lens and those of the Denbigh Flint complex as reflecting t he extent ". . .of cohesion within the Arctic Small—Tool t - radltion" (Taylor 1962:89). k pew-viz.— . 4.. I..- ....,.. ‘1 mt ..., n I ‘ -I‘...’ n ' _ " . 5a a. -M..-,.- ‘ v ' l 5"..-“ 7 fit-q 9.0:." ’/ .;.J . V Jilliiifjgs 1:- ...r. a '1‘ ‘ Ila: 6‘7”], ‘7- 23 Campbell summarized his extensive excavations at Anaktuvuk Pass in the Brooks Range of Alaska where his Natvakruak complex represented ". .an inland Denbigh Flint complex manifestation" (l962a:44) agreeing with the earlier interpretations of Solecki and Hackman from limited samples of the same sites. Campbell believed the alternation of several disparate complexes through time in the Anaktuvuk Pass region to ". . .represent separate hunting societies, often having quite separate cultural origins, who gradually expanded their territorial boundaries into the region. . .and who, in each instance, were content to settle in the higher reaches of the range and to exploit its resources, quite probably for genera- tion 3" (Campbell l962a:54)- Harp described the results of a survey of the Baker Lake and Thelon River areas of Keewatin where his Phase 2 was a ". . .Pre-Dorset Eskimo culture, derived from the central Arctic" (Harp 1962:72) which Harp suggested dated a 1000 B.C.. by cross-dating of artifact similarities with Igloolik. There was a notable absence of burins which he suggested resulted from the difficulty in working the predominant material-~coarse-grained quartzite. Harp aE1180 suggested that . .we cannot yet rule out the possibility that Archaic Indian culture may have contri- buted something to the Dorset Eskimos through this area. Such diffusion may have developed through the medium of Pre-Dorset culture there, and then have been transmitted to Dorset peOple who apparently adhered more ’ '1... ~ ~J '0‘) a- l 1:15 .17 ‘5“; 1'11 ’- '0 fit . 9 I 31v~./j/. _.‘ ' I l O A .Fnfi.fl C“. an: . Ub‘L “M. ' v I? I .- «age 3.1.. , 69:25) Bums y; 59:95 a n ‘ fr 3.4335 24 closely to the coast of Hudson Bay (Harp 1962:75). Irving compared what was then known of Alaskan and Asian stone industries dealing ". . .for the most part, with highly specialized types and modes (Rouse, 1953; 19 60), industries, complexes of types, and traditions. Alt tllea: present stage of work in this area it is seldom profit- able to give much attention to whole cultures" (Irving 19 62:55). He provided a list of traits characteristic of trite .Arctic Small Tool tradition. Large numbers of microblades struck from conical cores; IBurins with extensive retouch on one or both faces and prepared for hafting ("tanged burins") of several types; IBurin spalls retouched for use as minute en- graving tools; Many, very small, bifacially retouched, inset side blades, less than 4 cm. long, with dis- tinctive crescentic (not rectangular) shapes; Many, very small, biface points without stems or notches, but of specialized forms; Medium size (4-10 cm. long) biface points and knife blades, without stems or notches; Scarcity or absence of implements made by grinding or polishing, and of large imple- ments; .At most sites, absence of pottery; .A unique style and technique of tine workman- ship, which at most sites appear on most of the implements (Irving 1962:56). whi 1e this list differs slightly from that of MacNeish (See above), there is no doubt that they are referring to t:11€3 same manifestations and to the same series of sites. Campbell, in this same volume, mentioned the presence of distinctive societies with distinct cultural traditions a. t Anaktuvuk Pass, thus attempting to speak of groups of ¥ 25 Iaeaople with sets of behaviors other than technological. IEJrving, however, continued the traditionalist approach of sticking with technological studies only rarely consider- 11153 the implications of these concepts for groups of 19eecyple. A tradition, as the term will be used here, is an aggregate of type complexes which, by virtue of their sharing distinctive artifact types and other distinctive features such as styles of decoration and geographic distri- butions, give the appearance of having been derived from a common predecessor. Persis- tence and historical continuity over long periods of time are implied. A tradition is t:ual system designed for the study of peOple or behavior ()1: (DE culture. Irving's is a paradigm for the study of W. The reliance on a set of characteristic traits (guide 038113) for the description and categorization of Arctic ¥ ‘ I P t|AvVII ’5‘. \ ‘ INN. ‘v‘b A 5 v V. ' '.!I .‘R\ Ovapn.“ .4 JP. ‘--.‘.“ ‘ - T!;q._v fl‘“n: v h‘.~“.‘l “Vs-b‘u ‘uc “)- “‘ ‘ A ‘: .‘“ u " 5| ~.s‘ 9 ‘3': -. In I 4 fl? 1. .uvnu L’Cu,: Q H ‘F I. . "A‘ .Iabd.“, :13- | ‘A. s .u .N .5“ be 26 Small Tool tradition sites led to the establishment of a fairly homogeneous series of sites, especially since un- usual artifacts from sites believed on typological grounds to be within the Arctic Small Tool tradition were fre- quently considered intrusive from other cultures if they did not conform to the trait list (Irving 1964:143-148,294; Tv~‘1-)rlor 1962:88; Larsen and Meldgaard 1958:40; Campbell 1962az44; Harp 19582227). While most Arctic archaeologists recognized that not all sites which they would include within the Arctic Small T001 tradition would fit exactly the trait lists of Machleish or Irving, they were reluctant to include within the ir sample any artifacts that occurred in small numbers Ground stone arti- and were different from the trait list. facts in Pre-Dorset, microblades in Sarqaq, adzes in Denbigh at Iyatayet, polished burins and more crudely made end blades at Punyik Point, and large crude quartzite bi- faQ es in Dismal-2 are all examples of problematic interpre- tat ions that were once typologically eliminated from the co":ll‘plexes under discussion but which now seem as if they c011 1d be included as easily as a priori excluded--P€rhaPS mo]: e easily. The tendency to compare the distributions of specific a. - rtlfacts is nowhere more obvious than in Hadleigh-West's s 3rulposium volume in the Anthropological Papers of the U 1‘13.“7ersity of Alaska. While the Arctic Small Tool tradi- t i on should not be expected to rate much coverage in a , a... 4,...“ leede BC!» . .‘AH ,-' A:C:et , SC a hug ‘ o y-‘. Chats DJ. -. R Q -. 1 n b~dli 5001 SETHZLOZZS tention (1:- ‘ "A. ' ‘ .0..:ec an: 5.31:1:13 3‘; ULCCL: 3 care 5;, r3r3:n 27 volume devoted to "Early Man in the Western American Arctic", several papers discuss it in passing. Bandi dis- cusses burins in the Eskimo area, using data from Arctic Small Tool tradition sites across the Arctic, but his ob- servations have proved of little use and merit little at- tention (Bandi 1963). His definitions of ordinary, re- touched and pseudo burins are virtually useless in under- Standing the role of burins in Arctic technologies. Gidding's paper on Arctic spear points continues to c=C>Iupare specific artifacts of similar attributes from wide 1‘<'-=-l.‘l:1ging and probably otherwise unconnected cultural com- Plexes, in particular with those of the Great Plains (1963: "Yet if we 1]— ) . He justifies this approach by stating: are going to compare sites of the Arctic with those of di stant and warmer parts of the world, we shall have to do on the basis of a few wide-ranging styles, rather than 30 who le complexes of culture, for the Eskimos of the tundras Harp's were never the Sioux of the Plains" (l963:l). w“altitrning regarding the use of attributes and typologies de- .the analysis of ancient complexes far to \re loPed for ". the south" (1958:242) in the Arctic was tmfortunately not 1leaded. Lowther filled a major gap in our knowledge of the distribution of Pre-Dorset sites with his excavations at cape Sparbo on Devon Island (1962). The majority of the aartifacts were "typically" Pre-Dorset with several possibly ltl‘usive from a later Dorset occupation in the area, and ' ' a :18 622138 C. . . V I ‘ "q”. Raj: 3", :Le‘. w...” 5": ~21 uni he. ‘6'... ‘fl'lafi v,‘.. JVADC.’ . .~,. \ . . .3032: czara large H Li '9 . '1:'- a, any 4‘ an s I It O I 1 D nu. . ”a . a, “:1 ”c ‘v-.’: ' F ‘3 1 ...‘.v‘ :1 “va. 3.5: e a“ v.“ ' .‘ U‘. ‘ - §‘," '\ 28 the entire collection was very different from that exca- vated by Knuth in northern Greenland. Lowther concluded: "Thus the material from. .Cape Sparbo is principally pre- Dorset, Arctic Small Tool, with some that may be of a Dorset character" (1962:14) . Maxwell surveyed several areas southeast of Lake Harbour on southern Baffin Island, finding eight sites all Of which he categorized as part of the Arctic Small Tool tradition (1962:36). The sites appeared to ". . .demon- Strate an unbroken cultural continuum from a Cape Denbigh- like pre-Dorset period to the beginning of Dorset culture" (1962:39), linking Dorset culture to the Arctic Small Tool tradition. Rousseliére reported the results of a brief reconnais- Sance in the vicinity of the Pelly Bay Mission in 1964 where his Kugarjuk IV and St. Mary's Hill sites were con- s:LCIQred to be traces of a Paleo-Eskimo occupation older .generally within the Arctic than Dorset, and fitting ". While he ob- Sula 11 Tool tradition" (Rousseliére 1964:181). SerVed two different types of houses at these sites he was unable to relate these differences to other data. Subse- quent work at St. Mary's Hill revealed a significant dis- 1:T:j’~'bution of artifacts in relation to the house plan, but this enigmatic end note has never been expanded in print. Irving completed a major analysis of Arctic Small Tool t3: é"dition complexes at Punyik Point near Howard Pass in the B): (>ka Range in 1964, where he elaborated on the conception ' - ‘ I .‘A VA a. .-e n-Lu-¥ s ... C H >qu an fi.3.—3-¢'e :: tie: a: a L¢¢"A . ;~nn_ :qr ~‘.“|' M‘ssb Lbs Inn‘:| A" 'l-‘Afi p‘t-‘les' my: b 1- 0 3‘1 5' Awt-Qq .V. y‘" “‘bh‘d *hav 9 ,. .IO:.e a.: .. l nu Yr' ~...S H~CC I ”"Afiwn— n DUCL“.’3_. O. ‘ ‘n 9.. A, ‘ec b_.e bk 1 "“Yr~v >:-:.a; \ ‘ . . av-‘ku. “’ due "_ ‘ I ‘1 ‘ nave F u. ' +b~ a . . ‘-‘E r'. '-.',e 36:; C599 Def :ne p11? o‘. q “b91791 . «€712; : ‘jr #5 alSO S._ 2:3. 29 ()f the Arctic Small Tool tradition (1964:290). Given his trormative approach and the lack of distinctive stratifica- txion at Punyik, it is not surprising that he saw much evi- dence for mixing of artifacts produced by several com- plexes, most of which were related to the Arctic Small 123(31 tradition. There are also many cultures yet to be found in this area, indicated now by only an occasional specimen typologically different from anything in the complexes described above. In particular, several varieties of microblade and large-blade— and-burin-industry are definitely present, but have not yet been isolated in sites (1964:22). . .the Punyik complex as a whole is enough like the Denbigh Flint Complex at Cape Krusenstern and Cape Denbigh so that we may assume that they are closely comparable in age, that is, both date from around 3000 B.C. But a few rare types in the Punyik complex, such as ground burins and small, triangular points, compare closely with implements characteristic of the 2nd and lst mil- lennia B.C. in the Central Area and Greenland. There probably is a simple explanation for this, but as indicated in Chapter VII it is not now readily apparent (1964:320). Irving's suggestion that many of the finely worked arti- faQ ts are the result of the work of a group of craftsmen STE)‘E=<:ializing in stoneworking (1964:325) is one of the fin": st attempts in the study of this tradition to infer cul- tural characteristics other than technological, and his f"~1tt‘tzher suggestion of extensive trade networks involving both raw materials and finished artifacts ". . .provides a I>a“I‘tial explanation of the high degree of similarity fhroughout Alaskan sites of the Arctic Small Tool tradition; :L t: aElso suggests that change within the tradition of stone plement manufacture may well have happened somewhat : u ' C :q‘;‘\ “14"- O 5;"... aunt-ob ‘ I u q a tuna! .— * *h-1 on O but i but :rt:::'Je C rsarl Q med on tne - artisans" (19 UAOVAO'A‘. .wl‘Cv‘C. ‘ I P"’nq 0-,. 1: S\¢.C“ bA‘e C. n '“q..‘ . o...,:-'~ "w“ . " ‘LnLc 5A ,; uLI "‘e st'dd.: .I no. on; ”.3. Here A“ - t. Ergo . “(S RlfiJE COX 4. ‘ «as a ‘I.g. ‘ 3.x": “L0n ’ ‘9. F. 5‘ ‘Q \L‘ COM“ {a 30 independent of change in other aspects of culture, and fur- ther that there may have been regional cultures whose dis- tinctive character is partially obscured by uniformity im- posed on the only preserved relics by a sub-culture of artisans" (1964:327). However, it is also possible that the methods of analysis employed by Irving and others may also have ob- Scured the distinctive character of regional cultures by imp osing uniformity that resulted from a normative approach to the study of this tradition. It is indeed unfortunate tElan: more students of Arctic cultures have not followed Irving's lead to test hypotheses regarding other behaviors of people characterized by the Arctic Small Tool tradition. It is also unfortunate that the development of these ideas ha 8 never been published in a more accessible form. Dumond conducted an extremely extensive series of exca- Va~t:i.ons across the base of the Alaska Peninsula, where his BI‘°<>ks River Gravels Phase dates fl. 1900-1000 B.C. and contains a small biface and burin-microblade industry with 'I .an obvious affiliation with the Arctic Small Tool - Q tradition, the nearest exemplar of which is the Denbigh Fl int complex" (Dumond 1971:40) . Ground stone burin-like t:c"<>1s and adzes are found, and while . . .the B.R. Gravels assemblage includes many more snub-nosed end scrapers and many fewer burins and microblades than the type collection of Cape Denbigh, the resemblance between the collections is unmistakable. These people may be considered the first speakers of Eskimoan to enter the region (see Dumond 1965, 19693 with additional references) (Dumond 1971:40). ¥ — Ease s;:es ; I...:. _.11 nth-Lb 4.. ~ | 2:31:31 r. r I Q” ‘1"‘ J I. bd..‘.q U B.R. Gravels “" “J ‘ JL‘JQLU . l I Q'Ifl ' a to fi 4 W H~.a.-3“ 0‘ 1 4... U:..~e 53:: r A c Q “'5 .V'no-o',‘ f“ on» 1“».‘b D aAv 5. ~. 1 333 a rat 31 These sites mark the southwestern-most extension of the Arctic Small Tool tradition, which apparently laid the cultural horizon for the heritage of subsequent "Eskimo" cultural development, as Dumond saw continuities between B- R. Gravels and subsequent Norton-like phases (1971:40-41). Dumond has suggested quite convincingly that the com- bination of ethnographic, linguistic and archaeologic evi- dence supports the hypothesis that the people who produced the Arctic Small Tool tradition spoke Eskimoan dialects (DI—Imond 1965) and were thus the first Eskimo culture across a large expanse of the American Arctic. If this hyPothesis is appropriate, and it has received general ac- ce13tance and support, then it lends support to the use of ethnographic data on contemporary and historic Eskimo 1'll-l‘ll'lters as analogues for the development of models and hyPotheses regarding the behaviors of the people of the Arc tic Small Tool tradition. The similarities between his Brooks River Gravels as- semblage and that from the Closure site are truly remark- ab 1e, considering the geographic distance which separates them (nearly three thousand miles as the crow flies, but Inch more as the Eskimo walks, paddles, or whatever). Alexander excavated three chipping stations and a (lamp site in the Atigun Valley 70 miles west of Anaktuvuk P «388 in the Brooks Range of Alaska, lumping them into the I tiVlik Phase (following MacNeish 1959:14) of the Arctic s mall Tool tradition (Alexander 1969:51). While the ¥ c fit ‘ q .I" On». ‘I-ofi‘ -.‘ .u: inlay . s ” ...,,: . A .V.. {1‘ 3"...“ :“ ha .. ‘ “ ‘P. a d“1 - v». “u .. ' ' 1 . ": ”Win ”a. u... Vu..~“b ‘ I'NA . V l‘ A “"‘ib Rafi": - I 1 o., a. ‘ mi - I 0“ L..t ,. . ‘0- "Y V ‘3‘ . 51y ~p“‘ y' (1') n) ’l! O J ,»—\ "C (3 ,_.. ;_J A \(I) (I) t.‘ f ‘-l L...) P—J CL LL’ ' ‘1‘ 32 Egreatest similarities were with four chipping stations in tine Shubelik Mountain region northwest of the Atigun \Lalley (excavated by Solecki; Solecki et a1. 1973) because <31? the lack of side grinding on burins, he estimated the age of the Itivlik Phase as coeval with the Onion Portage lDeeribigh Flint Complex occupation of EE- 2200 B.C. (see be- lcruv; Alexander 1969:51-52). His sample size of 26 arti- ifélczizs and 187 flakes underscores the tentative nature of 11i-53 conclusions. The Utukok-Denbigh Complex was excavated by Humphrey ‘Irl ”the upper reaches of the Utukok River north of the Brooks Range in Alaska where two sites (56 artifacts) are tEVTzazically Arctic Small Tool, except that they are found fully on the tundra and they include no microblades (Iiilzumphrey 1970:120). Humphrey points out that the location of the sites almost .assuredly reflects an inland big-game hunting subsistence for Arctic Small Tool peOples in this area, for neither fishing nor gathering would provide sufficient food. Perhaps this different ecological situation could provide a functional explanation for the lack of microblades in the assemblage (1970:120). Giddings has probably excavated or collected from more 8 ites of the Arctic Small Tool tradition than any other ‘AL]:.<=1:ic archaeologist, beginning with the Iyatayet site at Cape Denbigh (the "type" site; see above) and continuing at <2: Ei19€3 Prince of Wales and Cape ESpenberg (unpublished; 1966) a ‘r1(1 Cape Krusenstern (1961, 1966) where his beach ridge E . .. .jt'.‘ "3 l 0 ' .J ' I A ' :95 to": F . ‘ .QA‘I' ‘ Cedeieu on a W O A 5A. “3. $2412 n!‘“ .A“ .."_5 It. A n q .Lv ‘fi‘ a v”. Jud ‘0 01' . . ’ ’z-ob: ‘ ~“-L.s3.i p e e0 'ed 3* A .81" s‘.i - ‘u‘v. » | New» . . WI ~A 1‘ 3 r»\ T s A: .N. h 5 KM. 33 chronology was generally corroborated by his subsequent and highly succeszul excavations at Onion Portage, on the Kobuk River (1962, 1965, 1966, 1967). Gidding's 1964 report on the Iyatayet excavations is the most complete and lavishly reproduced report we have received on an Arctic archaeological site, summarizing the data for several occupations including the Denbigh Flint conlplex. However, it is a disappointment in that it does not contain a single reference to the Arctic Small Tool It tr adition or to Paleo-Eskimo as generalizing concepts. is apparent that these idiosyncracies on the part of Giddings reflect his attitudes and maybe even his disap- po intment in finding that the Denbigh Flint Complex was not accepted as resulting from "Early Man." While his data on Chipped stone artifacts are baselines for comparison as are his excellent illustrations (both line drawings and photo- gT3E3-phs), his failure to consider adequately the extent of gr inding on burins and adzes has reduced the utility of his sTr-l-il-(iies. Again, his use of a normative approach to his data has resulted in the exclusion of infrequent traits at Iya tayet from the careful analysis that characterizes the ‘Ir‘eflnuuainder of his research. The Denbigh Flint complex may be summarized speculatively, then, as representing pe0ple who 'visited Iyatayet only seasonally for sealing and caribou hunting, bringing with them the raw flints needed for temporary manufactures, fore- going the heavy work that one might expect around Inore permanent camps, and moving into the forest for the winter season. Skilled enough at boating to procure seal in quantity and to live along a U ‘ Y. ’y, 'I. 44- IE5! .‘uc ' Alaska, 5. cellen: 8‘ dry, viii and the: a princi; fie" ‘~ not. A: 1 A 1" COO) " t 15 SCEI‘CE used '3" E 25$ 0: " fin 1 '- fC3;.e a_ I .- 33312131117 'Y'Aflc 15;: mm c" -._ g . - N . on “8:: m4 a“; :Vfi . 34 very wide stretch of the seacoast of western Alaska, they may also have possessed the ex- cellent snowshoes without which life in the dry, wind-free forests would be unthinkable, and they probably concentrated on caribou as a principal food source throughout most of the year. As to whether or not they were Eskimos, we can answer only by defining "Eskimo" more closely than we are accustomed to do. There is scarcely a Denbigh object in the same form used by Eskimos of Thule or later cultures, yet some continuities into the Ipiutak period of about 2000 years ago are quite direct. Regard- less of how we designate them, these Denbigh people appear to be in a direct line of cultural continuity with Eskimos. Technically, the Denbigh Flint complex is more closely related to the European Paleo- lithic (in burin techniques and variety, certain scrapers, and miscellaneous rare forms), the ‘Mesolithic of Europe, and the early "Neolithic" of Siberia (in microblade industry and inset side-blading) than to early temperate America; yet fluted and diagonally flaked points and ex- pert bifacing of small flints are old in America. The complex was unique in the world, however, in the meticulous skill employed in flint flaking, and probably in the origination of the burin spall artifact, the most minuscule of the widely used flint implements (Giddings 1964:242-243). G:Latlvziing's treatment of the Eastern Arctic is indicative of tle-EB- unwillingness to see Denbigh as part of any larger cul- tural system. Thus, moving from the Bering Strait region around the coast of Alaska and Canada to Green- land, one finds clear-cut evidence of a conti- uity of the Denbigh Flint complex--in a "small tool" horizon--the entire distance, though with a time lag of 1000 or 2000 years between sites of Denbigh Flint complex of western Alaska and Independence I in Greenland. Whether or not there proves to be a still earlier Greenland manifestation, the derivation of Independence I from a Denbigh-like base is clear. As I inter- pret the data, a Sarqaq-like cast of culture in the east resembles the Choris and pre-Choris casts in the west enough to suggest a second and following continuity across the Arctic (Giddings 1964:261). Va~ “Van!" 7“ O‘Dedhbontb’ it ‘,~‘ 3“. 4 >0 -t'....5»€u 5 U z i ' ' uq- ' p”, Oohch a} 0|...C '3' . a ‘ 2" o . .o-oy_‘ J H ~ ‘ANV " a. ( J o u I . o~yvu. \./ . ‘ “0":- N \ A u to.““ A... ‘ l;'~ : , ‘.~ 0"“n . ”I" «we. .- ‘ rm. 3;; “qu , ~‘IV "‘\. EVA. A ‘ U 41 This sequence is corroborated by the mixed samples at Engigstciak, Irving's Punyik Point, possibly the Trail Creek Caves (Larsen 1968:78), and Onion Portage. Excavations by Dennis Stanford at Point Barrow have revealed a similar late stage of Denbigh with pottery, which Stanford has called Walakpa Denbigh. According to McGhee (1971:491) "The Walakpa Denbigh inventory includes cordnmarked pottery, ground slate, burins, microblades and other typical Denbigh forms" and is dated by two dates: 3400i520 years (93. 1450 B.C.--Gak-2299) and 2260*300 years (93; 310 B.C.--Gak-2300) of which Stanford prefers the earlier (McGhee 197laz49l). Note that the difference be- tween these two dates would occur by chance from.the same population less than once in twenty samples (p=g§. 0.05 by chance), thus either the occupation was for a long duration, the samples are in poor association, or there is possible contamination of one or both dated samples. Other sites were found in the same locale, including several Classic Denbigh sites. These Denbigh sites on the northern coast should provide important evidence for the adaptive diver- sity of Denbigh, as they are rare along the Alaskan coast. Presumably, their coastal location is suggestive of a rather late date in the continuum of the Denbigh Flint com- Plex. Hall conducted an extensive series of excavations and SLlirveys at Tukuto Lake, approximately 140 miles west of ‘AILakruvuk Pass in the central Brooks Range. His 92: 200 .l ' .vo-ozq-s 9v“ 1..b--Abu “\ uuvnc‘flvbfll ‘ finotvvc. u . I .'l W O! a? Oooe‘ .M J. I :Vva. . ' W‘ FA ov~ .9 b-‘evfl':'cb .- S‘A“ a” N 6“ «b_ C u ~.‘ :Q-* .4 ‘ ' .4 n, L,‘v-- L‘s-.0 :o '3‘; ‘u ) 0* § A \: “1P“, 1‘“ \u_. 1.7% l ‘ OP \J . "‘~.1 *4 an x “ ' 42 artifacts from Arctic Small Tool tradition sites are as yet unreported except for his preliminary report (1970). Hope- fully the precision with which his report indicates that he excavated will be reflected in the final report on these materials. McGhee's report on Arctic research also indi- cates that some of these artifacts were associated with "thin, hard-fired pottery" (McGhee 197laz491), thus we may expect additional information on the later phases or the Denbigh Flint complex from these researches. In 1961, Ralph Solecki conducted an archaeological survey of two areas North of the Brooks Range in the far northeast corner of Alaska. Here, four sites contained materials relating to the Arctic Small Tool tradition, but their extremely small sample size forces reliance on "index fossils" such as chipped burins, burin Spalls, and the use of the Arctic Small Tool technique of parallel flaking (Irving 1964:325) (note that Solecki relates the diagonal parallel flaking to Angostura techniques, citing a personal confirmation by MacNeish--Solecki et a1. 1973:88). None of the burins is mentioned as having grinding facets, although the entire sample from four sites is 9 burins and 9 burin spalls. Solecki infers that this sample of sites ".,. .is simply the representative record of a seasonal hunting ground, or a part culture" (Solecki et a1. 1973:88). Since its inception in the early 1950's the Arctic SInall Tool tradition and its component sites have not been 8lejected to a major synthesis, although Irving's 1964 nae? vi» 59;.“9: 3“~y a-~¢o¢.. D‘AAA "A V .”:= ‘tbea.\ 90" . ’ 5 a-“ *0 rt rub-y' . o u. -' p 91.6.3103 01 43 paper is an approximation. Almost everyone doing archaeological survey work in Alaska has found some Arctic Small Tool tradition sites, but they have not received pri- ority attention as they did not suit the problems that these researchers were pursuing. Even as this thesis is an attempt to review and revise our conceptions on the eastern extension of the Arctic Small Tool tradition, there is a need for such an intensive study of the Alaskan representa- tives. There is a further need for those with small Arctic Small Tool collections to publish them, or at least make them available to other researchers, as there are really a lot of sites located, tested and even excavated, but the amount of information in print is woefully inadequate for anything but a cursory analysis. The data are there for analyses going beyond the study of technology to the study of the human behaviors of these tool-makers, but this will require a major effort by some researcher and great coopera- tion from Arctic archaeologists. Turning to the Eastern Arctic, Taylor reported on his excavations of the Roberts and Arnapik sites on Mansel Island in northeastern Hudson Bay where his nearly 1900 specimens (1791 artifacts) make Arnapik one of the largest Pre-Dorset samples reported in print (Taylor 1968a:15) (see Taylor 1968b for a photograph of some Roberts artifacts). 1The site comprised some 120 find spots predominantly Characterized by a scatter of somewhat larger rocks than the usual surface scatter with artifacts found on the sur- 44 face and within the rock matrix. These rock and arti- fact clusters extended over approximately 1200 yards along a headland at Si- 90 feet above present sea level (Taylor 1968a:12-14). Samples were recorded and kept separate from each cluster, but analysis revealed no consistent differen- tiation by area. Because of its elevation and relations with other radiocarbon dated Eastern Arctic sites (com- paring frequency of polish on burins, popularity of burins and burin spalls, and specific artifact forms), Taylor es- timates the occupation as dating between 1500 and 1000 B.C. (1968az43). While specific comparisons with the Closure site and other Pre-Dorset sites in the Lake Harbour region will be made in other portions of this thesis, it is inter- esting to note at this time that no soapstone was found at the site and that the proportion of burin spalls is among the smallest from any burin producing site in the Arctic. This anomaly may have resulted from a combination of natural and technical factors, as the freeze-thaw cycle in the Arctic can lead to differential sorting of lithics by size, with smaller sizes descending and larger ones rising (see Corte 1963:499), thus imbedding the tiny burin spalls deeper in the surface gravels under investigation here. Additionally, Taylor explains that "The crew was small and generally unaware of the nature of archaeological field work. Nor were we blessed with good weather and an abundance of time. Consequently I restricted our activi- ties on the site to surface collecting and recording" l f 5 a". .an/be 0 showed - ”00‘ I I L .1111 1 : .1 1.111 11 l 1:. l l 11 ..L ..c l l. . _ n-I. rt 5 5 C 1 a L.” .C .7} S .. d 73 33.1. :5 .3 1 a 5 .3 .C 3: O w. an a“ .an” o-lx. "nu No: i: :D..c ~.D .43 w.“ v. “a. a5 0 ... J. a» a» —.. L... .u .C ... ... .1 2.. :e v. 2.1 is .an 2» .C A. J. «T... “J ..e a. v... as. J. as an . a I. .7». . . . a“ on g“ .u .. vu . 3. .~ ._u ... 2.. . e I... n.- .u a: We 3, w. .4 w. n a o . ..e nM n. 2 BU a.» .\u ~\v . o n v e . H1. . a \ J A... ....« a; 1.. a: an Table 1 Summary of Publication Author Dates Giddings l950,51,64 Giddings Unpublished Giddings Unpublished Giddings 1960,61,66 Solecki 1951 Solecki 1951 and Hackman Irving 1951,53,55 Campbell 1962a Irving 1951 MacNeish l956,59,62 Irving 1962,64 Dumond l964,65,69,71 Alexander 1969 Humphrey 1970 Giddings 1965,66 Anderson 1968a,70 Stanford Unpublished Hall Unpublished Solecki 1973 et a1. Solecki 1973 et a1. Anderson 1972 Map Key \0 (I) \lO‘U‘I U1 U1 U‘IJ-‘UONH F‘ ta ta H‘ no to :4 c> 13 14 15 16 17 45 ASTt Field Research: Western Arctic Location or Sites Excavated Cape Cape Denbigh, Alaska Iyatayet E3penberg, Alaska Cape Prince of Wales, Alaska Cape Krusenstern, Alaska Anaktuvuk Pass, Alaska Natvakruak Anaktuvuk Pass, Alaska Natvakruak Anaktuvuk Pass, Alaska Imaigenik Anaktuvuk Pass, Alaska Susitna Valley, Alaska Firth River, Yukon Engigstciak Itivlik Lake, Alaska Punyik Point Alaska Peninsula, Naknek River. Brooks River Gravels Atigun Valley, Alaska Itivlik Complex Utukok River, Alaska Utukok Denbigh Complex Onion Portage, Kobuk River Denbigh Flint Complex Onion Portage, Kobuk River Denbigh Flint Complex Proto-Denbigh Point Barrow, Alaska Walakpa Denbigh Tukuto Lake, Alaska Arctic Small Tool tradition Sadlerochit-Shubelik Mtns. Denbigh Flint Complex Franklin Bluffs Denbigh Flint Complex Noatak River, Alaska Denbigh Flint Complex 46 1 ‘. 0.. c o, n .‘o .:.:.:- J ' .H .‘5 C " - '.. .14 .10 '. "' 0' os ' ‘- On 1 A57! Sites .6 Figure 1 ASTt Sites from Table l d- O' .U C q-ol’ .‘ . .4 u-rb-‘b E J *' ~" .C —-:‘“' IVA u 5“ .HR‘ gag. ~~¢ ' 3“,..«6 : N Mouth “‘ 17 V _ 3.... -... -;,"1 U . 5..., 0.." -'§ um sud 1-. pr 0 I n O 1‘. I "I4 .Q . n . l1 ‘- PI All. 47 (Taylor 1968a:15). The sampling techniques and the environ- ment may have combined to produce fewer burin spalls than we might expect from this site. Taylor's research here, and at the Tyara site reported in the same volume, virtually nailed the lid on the argu- ment that Dorset developed rather directly from the Pre— Dorset in an area which included the Hudson Straits region of the Eastern Arctic, and his summation of research through 1966 is extremely useful as a guide to the litera- ture. In addition, he seriated the Sarqaq and Dorset sites reported by Larsen and Meldgaard (1958) from Disko Bay, West Greenland, where he suggested there had been a period ". .of more rapid cultural change near 500 B.C." (Taylor 1968az9l) but that the ". . .Disko Bay Dorset de- veloped in large part from the preceding Pre-Dorset Sarqaq stage without any appreciable break in the occupation of the region at that time. Since the late Sarqaq of Disko Bay seems contemporary with the earliest recognized Dorset occupations in the Canadian Arctic one may suggest for con- sideration that, as the Sarqaq-Dorset transition occurred slightly later around Disko Bay, the stimulus or causes for the change derived in some measure from the Canadian Arctic" (Taylor 1968az93). This hypothesis is almost diametrically opposed to the previous interpretations by the Danish archaeologists of a migration of Dorset people into a virtually depopulated area (see above for their earlier discussions of migrations into Greenland). .T «a-‘R f ‘5 Own-Jr . I . -nq.A‘ t.‘ . 4:..-tu ‘. - fl‘fiuaqu R‘jfl.. 5:1,. V-»av'_. O n.. - s ': Fag ‘-.‘ yon» .J ~_I AvaA-‘ H ‘5‘ Id-b:.“‘d . by ‘ ' fl-v r . ““43- D i. 4 C l" 9 .1. \- .. q NV bvaub:0 '| .an ‘— .V., L..C e -~==e as ~ 5 ‘,AI‘A:‘ 1 {NDD‘DLV C U 1“ . :‘asfa _, o. . . ‘\ ‘ l was“- ‘b-.“::“ 2s . u» ago ". ’5 .q . .- «.1. 1c, . YA‘ ‘9ev'; I‘: ‘LJL‘ 48 Knuth summarized his interpretations of the results of his extensive series of excavations in northern Greenland and adjacent Ellesmere Island, where his Independence I and 11 occupations followed the "musk-ox way" (Knuth 1967). The general picture of these high Arctic peOples is fairly clear, with a reliance on musk-ox and seal hunting, supple- mented with other land mammals. Caribou and walrus were conspicuous by their absence. Their technology was most like the Denbigh Flint complex, than like other assemblages in Greenland and the Eastern Arctic, characterized in par- ticular by a lack of grinding on any sort of stone tool, larger burins, double tapered end blades, microblades, and no soapstone. While specific comparisons will be made be- low, the early radiocarbon dates have led some to regard these as the earliest migrants into the Eastern Arctic, possibly coeval with or earlier than Classic Denbigh in Alaska. Knuth approximates dates of 22° 2000 B.C. for Independence I with a stage transition to Independence 11 at about 1440 B.C. and Independence 11 at 93. 640 B.C. (Knuth 1967:26). Knuth emphasized that cultural changes involving Pre-Dorset and Dorset cultures in other areas may only have weakly reached these remote areas, where the peOple ". . .were forced to maintain certain basic princi- ples in their high—arctic economy” (Knuth 1967:40). However, there are also some aspects of Knuth's metho- dology which may have contributed to his considerations and conclusions. The image that Knuth has portrayed and 49 published is one of an Independence I house form involving central passages (the mid-passage hearth), but his report indicates that this feature was found in only 30% of the cases (47 ruins) of which 40 were excavated which is a much higher portion than was excavated of the ruins without mid- passages. The more extensive sample of mid-passage ruins may have biased the sample in favor of representing what- ever factors lead to the creation of the different house types, so that the picture which Knuth paints may be biased towards the unique features of the Independence culture. His conclusions are somewhat difficult to substantiate from the data presented, as the "big picture" is well documented, but the artifacts and their distributions are slighted, even to the extent that his single figure of stone arti- facts has appeared in print before, and so we are in the dark as to the diversity and characterization of his lithic artifacts--the ones which have the greatest affinity with those from.Alaska. However, his series of well-conducted and truly re- markable excavations under extreme environmental conditions are only to be lauded, as is his use of corroborative data from other scientific researches in this region: Whether his Independence cultures represent regional ecological variants of the Arctic Small Tool tradition, or evidence for the earliest of several migrations across the Eastern Arctic remains a question for later in this paper. Maxwell continued his researches near Frobisher Bay O . .O'A. “\n‘ u'v :.o:.boaéot‘n v Q bhva..~a a . v .onOV ‘n-j o p 3" 3'4 * ‘V' nu . :."‘ ,A-A ‘ a 441.94.; {1.4 '._V O: '~ ‘ _ 'a.-a., arc ‘ . a 1 a... v‘v' \ .fl 1 ' ~hl“. AA‘A‘ . vu' ‘-"Q‘ A“ Vt. “y“v‘ 50 and Lake Harbour on southern Baffin Island where he established a long sequence of continuous occupations stretching from the Pre-Dorset Closure site (93. 2000 B.C.) through a transitional period to Dorset culture (1100-700 B.C.) and fully developed Dorset lasting at least through 500-600 A.D., (Maxwell 1973az287-88,351; 1967). His exca- vations are among the most extensive (and intensive) in the Eastern Arctic and his combination of trait seriation, sea level, and radiocarbon chronologies of develOpment of Pre-Dorset--Dorset is the most completely documented and published in the Eastern Arctic (Maxwell 1973a). Maxwell's conclusions emphasized the technological continuity (with relatively slow technological evolution and stylistic variation) throughout the sequence of occupations, empha- sizing the evidence for the in_situ evolution of Dorset culture within an area of which the Lake Harbour sites were a part (1967). A second general point of interpretation to be derived from the Lake Harbour sequence (and augmented by information from the whole pre- Dorset-Dorset geographic range) is that this was essentially, if not completely, a closed cul- tural system. Once the Arctic Small Tool Tradi— tion arrived on the coast of the Eastern Arctic it became the tool inventory base for all subse- quent develOpments. At least in the Lake Harbour region, there is no develOpment in material cul- ture (and as yet no discernible development in non-material culture) throughout the continumm that cannot be accounted for in the indigenous system. Each specific tool can be traced through a developmental sequence to its prototype in the earliest sites (Maxwell 1973az343). Maxwell has long been impressed by the implications of this H- 'h’U (D U) I" (D (J () I“? I [I 1: r1 (D D’ If ‘1) \‘i'l H- r? m o 51 almost three-thousand year continuum of steady-state de- ve10pment (personal communications) as it is manifest of an unusually conservative technological system. The interpretation this evidence leads me to is that from the very beginning the average tool kit of the average man functioned suffi- ciently well to allow him to maintain an effi- cient position of equilibrium within his eco- system. Over time there were minor attempts to modify the forms of certain tools, presumably in the direction of greater efficiency. But there does not seem to be any compulsion toward change, and in fact the very changelessness of the tool kit suggests ideational sanctions against major change. Such minor stylistic drift as does occur, more often than not re- turns to the original starting point, reminding one of Kroeber's famous study of oscillating hemlines on ladies' dresses. In ecological terms, in this particular continuum, the cul- tural, or extra-corporeal adjunctive aspects of man's adaptive equipment became a constant-- comparable to the fangs and claws of the polar bear -- so that he remained at a climax state in a condition of dynamic equilibrium as the top carnivore in a virtually unchanging ecosys- tem. In other words, seeing culture as an adaptive mechanism, the adaptation once evolved did not have to be modified, and there appear to have been selective pressures to maintain a changeless technology. It seems to me that this provides a subtle but significant difference in the way that culture is often viewed. Perhaps it is because culture history has provided few examples of a steady state maintained over several millenia, or because, from a platform of ethnocentric progression, we see the adaptive as- pects of culture as providing increasingly ef- ficient ways of processing resources. But cer- tainly in the technological aspects of culture we are more apt to think in terms of progression or regression rather than in terms of steady states. The parallel here, in evolutionary terms, is the specialized rather than general- ized animal, although this comparison lacks apt- ness from the very fact that this average tool kit could be used to exploit a finite but varied complex of land and sea animals. It is of this type of steady-state adaptation that Charles Cleland (Cleland:l967 [1966]) is speaking when 52 he refers to adaptation characterizing societies with focal, as distinct from diffuse, economies. This ecological equilibrium.is even more im- pressive when we consider it in the matrix of climatic change. The three millenia continuum we are considering here bridges a number of climatic shifts which have had major importance elsewhere in the world. To date, insufficient research has been carried on in the Arctic to suggest the im- pact of these changes in northern latitudes, but there is little reason to believe that climate has been at a constant state in this region through this period. . . .Presumably, however these shifts in temperature were not of such mag- nitude to cause dramatic shifts in the ecosystem (Maxwell 1973az341-343). These conclusions come from data which include the presence of ground slate adzes and burin-like tools in the earliest Pre-Dorset and true-burins in the latest Dorset (to name only two). An alternative explanation for the extent of cultural homogeneity through both space and time is that there was considerable mixing of component samples by the re-occupation of sites and by the use of cut-sod from ad- jacent sites (both older and younger) for house construc- tion which effectively mixed components from vastly differ- ent time periods. However, few would argue that this would account for mixing of the very earliest and the very latest components, and mixing would mostlikely occur in adjacent sites, which Maxwell's radiocarbon dating program suggests are usually also the closest in time (see sites in the Tanfield Valley). Additionally, this continuity may be the result of one of the most complete diachronous samples from a single locale in the Eastern Arctic. Discontinuity else- where may be because the archaeological sample is . ;:.AO."- n “-31., .,_‘.. "Av. ny- £13-30... . . P- , e ”IF..." .‘ V"'rL ‘ o "l ." .n« "‘"W "*1. b...u‘v ‘ :~- ~— 0 I! ,.n “‘3 D‘A~.: . ‘Cv Q um C' ,n._‘ ‘ l C :a ‘fi I a“ l“ ‘c C: . I5 “'e S‘.‘ r: -“I\‘~ 0 tr ~\J\la1 ‘ I ‘$ \"\A|‘o - t’v _'\ a l ‘5: \ $2 ‘I 53 discontinuous--the opposite of the situation near Lake Harbour. While Maxwell's analytic approach is essentially typo- logical (see 1973a: Chapter II) with regard to artifacts, he was largely unable to discover any set of horizons in artifact style, material, or function. To a great degree, the completeness of the artifactual sample through time (note that this does Egg suggest that any complete cultural system has been sampled, but only that a piece of geography through time has had its human behaviors sampled) made the usual normative approach (which Maxwell had used effec- tively before, and with which he began this series of exca- vations) of less use after the inferred culture history had been established, and Maxwell embarked on a discussion of cultural processes. This discussion could only come after his previous analyses had resulted in a regional chronology and culture history. In so doing, he was breaking new ground in discussions of the Arctic Small Tool tradition, again largely because his data were so complete. In attempting to go beyond the study of artifacts to the study of the relations of other systems (cultural, social, technological, etc.) with their environment, Maxwell found that his interpretations, if formulated as hypotheses as they could easily be, were not capable of easy testing in other locales in the Eastern Arctic because the data were either not available in print or not complete or not collected in a manner as to make them comparable. 54 While he did not articulate these problems with his inter- pretations, these complications with the data available for the testing of sophisticated hypotheses are endemic to Arctic archaeology and result, in part, from the Operations of what passes for a paradigm among archaeologists (see above for a discussion of paradigms). The techniques and methods were adequate for the examination of the Spatio- temporal distributions of artifacts, sets of artifacts, and attributes of artifacts, and as long as this was seen as the major goal of research, then our discipline was ade- quate for the task. However, after this had been accom- plished in a locale (such as the Lake Harbour region, or even southern Baffin Island?), Maxwell was unable to pro- ceed with certainty, as previous methods and techniques had not provided adequate data for the consideration of his hypotheses. This situation has contributed to my present concern for archaeological techniques and methods and data from.the Eastern Arctic (see Dekin l973a:4l-42, l974--in preparation and in press). Maxwell reiterated his inclusion of these assemblages within the Arctic Small Tool tradition agreeing with Irving that the tradition may contain ". . .distinctive regional variations, but the commonality of specific complexes of types is more impressive than the variants" (Maxwell 1973a: 346). He went on to make the following suggestions ad modifications of Irving's core of characteristics of the Arctic Small Tool tradition. .: .7 ”a .c «L Q. 2. .5. .C .2 T. O .: ... a. a: .G A... u. 9-. I as s. Le... a. C ... . 1.3.1... 3: .; ... a. a. T. 3 E a. e . v. ”0 .u a; 2.. .c .4.... .3...- a» T. .d... an...“ a.» ..- v. 3; .3 a. e 3 a. .....u... 3 _.. v. 0 .e r... a r.e 3 2a.... 3 at O a C e e. e.. a e. a no .. 5 .L ee r at... C 7L at C a; e h; 2. .an e .3 .2 s _ he E s. e vs L .. a: a) e .3 «5 FL ab nu“ «V s an a. an .e —d at ‘5 v. \ w a . a s. «5 .\4 H! .\J is .5 .3 v--. 55 This list includes large numbers of microblades struck from conical cores. I would change this to read "prepared" cores, as a more general term to include a wide variety of core shapes through- out the distribution of the complex. Continuing, he adds burins of several types, and burin spalls retouched for use as minute engraving tools. This last phrase should be modified since it is now apparent that burin spalls were used in many activities in addition to engraving. The descrip- tion of side blades should be less specific, since, in the Eastern Arctic particularly, there appears to be a wide variety of side blades, but his de- scription of small biface points, or end blades, I would make more specific to refer to two dis- tinctive forms -- the small triangular form and the constricting, tapered based form. To this list I would add a distinctive complex of scraping tools which include edge-expanded end scrapers, and side scrapers with oblique and concave edges. Grinding and polishing of stone tools now appears to be more common in the tradition, and the unique style and technique of fine workmanship which Irving refers to appears to be much more diagnos- tic of the western than of the eastern part of the range. Presumably the well-spring from.which this tradition departs for the east is the Denbigh Flint Complex (Giddings:l964). While this complex contains features distinctive from early traits at Lake Harbour, there seems to be little doubt that the assemblages from the two regions represent a common technology for adapting to common ecosys- tems (Maxwell l973a:346-347). These revisions were suggested apparently to make the trait list of the Arctic Small Tool tradition more descriptive of the Eastern extension of this tradition as the original characterization was based largely on Western data. Much of the substantive data on which this thesis is based was collected under the aegis of Maxwell, and rele- xJant portions of the data presented in his 1973 report will be re-examined in conjunction with the previously unpub- lished data from subsequent field seasons and from the asure S? F‘ L- O“ a...“ g.:"\‘ 3"qu - d a 1 ”f .A\ "E g a '\ .fidot Centrai o 071 me 33' canzai A H “a. .‘C' PI,‘ QN.‘ 56 Closure site. Many of the ideas presented here have their stimulation from Maxwell's work and thinking, both pub- lished and unpublished, as my selections from his work may indicate. Taylor has reported on his 1963 reconnaissance of the Central Arctic coast between Cape Parry and Cambridge Bay (1964, 1972) and on his continuation of this work in 1965 (1967) on Banks and Victoria Islands. The Buchanan site on the south bank of the Ekalluk River on Victoria Island contained the only Pre-Dorset component encountered in 1963, where its assemblage included burins, side blades, microblades, end blades, and other small chert tools (Plate II) along with a considerable number of large coarsely chipped biface blades or chOppers and worked quartzite flakes (Plate III). Of the very few nonlithic artifacts, five of antler and ivory are surely arrowheads. They have rounded cross sections, conical or scarfed tangs, and, in one case, a barb. Although seal and fish remains were recorded in the bone refuse from the three test trenches in the Pre-Dorset midden, caribou bones were again predominant. The age of this component leads to indulgent speculation; my speculation that it was occupied about the period 2500 B.C. to 3000 B.C. may well be a very bad guess (Taylor 1964:54-55). This site was significant because of the association of large crude quartzite tools in situ with the smaller Pre- Dorset chert tools all within a closed stratum--a buried soil horizon with sod separation from a higher surface scatter of Dorset artifacts. However, subsequent work (in 1965) cast doubt on the association of microblades with the Pre-Dorset artifacts (1972:64) but confirmed the Ad -v H .. SO 13 o . 0L no v1 an ‘ Au‘b‘ a 031C 3 v ‘ A . ‘&‘CC ?.~’ 1 q . a “ ‘ \u. 33 "1' av “‘B. 57 association of the chert and quartzite artifacts found al- so in Harp's Dismal-2 and Kamut Lake assemblages ca. 300 miles to the southwest (see Harp 1958, and above). Both the location and the predominance of caribou bones (both raw and made into artifacts) indicate the heavy dependence on caribou with virtually no sea mammal bones. Taylor's 1965 work on Banks and Victoria Islands con- firmed his previous work at Buchanan (with the exception of microblades-—see above) and resulted in the discovery of several additional Pre-Dorset sites. A small copper frag- ment was also found in the Pre-Dorset component of Buchanan and the association of larger quartzite tools continued in the Wellington Bay site, Which also had a more maritime orientation deduced from the majority of seal bones from the refuse (Taylor 1967:225). The Menez site also was a Pre-Dorset site with artifacts similar to those described above from Buchanan. Several radiocarbon dates surprised Taylor by their lateness, with dates from Buchanan, Welling- ton Bay and Menez ranging 200 years each side of 3000 B.P. (ca. 1000 B.C.) (see below for date list) indicating an oc- cupation late in the Pre-Dorset period (Taylor 1967:229). On Banks Island, the Umingmak site and the Shoran Lake site echoed the general picture of Pre-Dorset that emerged from the Victoria Island excavations, but with a reliance on musk ox for subsistence (Taylor 1967:227). The artifact sample from.Umingmak was somewhat larger (N=400) and it in- cluded rare examples of polish on burins and microblades ..' m— ‘. [I'J . A“ 'Fq‘r 3 -QA- C 0 5‘4- y‘uv 5 '0‘3.’ J" u ‘fl-gbn .an—s DV‘.." “a: UV A x ) V n \ v’rs h.“ ’ .35‘5..: '.‘ '1: 58 with somewhat lower frequencies of large quartzite bifaces. Taylor suggested ". . .that the large, ovate or semi-lunate quartzite bifaces represent a low frequency Arctic trait that, by stimulus diffusion from Archaic complexes to the south, was proliferated in the Pre-Dorset of Victoria, and to a lesser degree Banks, Island" (Taylor 1967:228). Taylor's age assessment of Umingmak (as earlier than the Ekalluk River sites on Victoria Island) was borne out by radiocarbon dates of approximately 1500 B.C. (Wilmeth 1971: 80). McGhee continued the National Museum of Man's interest in Victoria Island in 1966, where he found two Arctic Small Tool tradition components with small quantities of arti- facts (OhPo-4 N=4l including 19 retouched and utilized flakes; Oqu-4 N=15 including 6 retouched and utilized flakes) (McGhee l97lb:l60-163). The large-tool quartzite industry from the Victoria and Banks Islands sites exca- vated by Taylor is not prominent, although several flakes and an end blade were found. The use of basalt and schist at Oqu-4 was similar to Harp's Dismal-2 site (Harp 1958), and McGhee saw the affinities of these sites as being with southern Arctic Small Tool sites rather than with the northern ones on Banks and Victoria Islands, with their ages as early (OhPo-S) and relatively late (Oqu-A) in Pre- Dorset time (McGhee l97lb:162-3). Noteworthy is the pre- sense of a contracting stem end blade of quartzite and a heavily ground burin (0qu-4). It seems reasonable to Q :36 5333‘. kl“ "‘ '05 ‘5: amen: o ‘. ., 5 ‘--3 rt. 5. ‘3 ‘¥ y 59 hypothesize that these sites pre-date the acquisition of the large quartzite bifaces in this region, although sea- sonal behaviors provide an alternative explanation, given the scanty nature of the comparable data available. McGhee excavated an Arctic Small Tool tradition com- ponent of the Bloody Falls Site in 1968, which is now one of the best described and dated collections from north- western Canada (McGhee 1970a). The 250 artifacts (and 2955 detritus flakes) contained 30 burins (3 with surface polish), 84 burin spalls, 15 quartz crystal microblades, end and side scrapers, end blades (several with ground facial facets), and several rare artifacts of possible diagnostic significance -- two ground slate end blades, a chipped and ground adze, and three "pins" of native c0pper (1970:54-59). Wood charcoal from a hearth was dated 23- 3300 t 90 B.P. (8-463) which McGhee views as consistent with the artifact assemblage. It is perhaps surprising that the large biface industry from Victoria and Banks Islands is not represented in this collection, which dates several centuries more recent in time, and McGhee suggests that such an industry had not diffused this far in this direction (1970:59). Noble conducted an extensive series of site surveys in the Central District of Mackenzie from 1966 through 1969. Noble used the concepts of tradition and complex to define representatives of three archaeological traditions and nineteen complexes (phases). In so doing, he followed s;ec;: ~ .1 ‘v’t.0 ~"i.‘1 ‘- 5.; ‘s 0”... ' N -5 a- £3 6O MacNeish and Wright in their use of the concept of tradi- tion: a tradition is a distinct way of life reflected in the diagnostic material culture of a series of generically related complexes, which persist through appreciable time and across space. In specific cases, the tradition may conceivably de- velop from a fusion of several traditions, and it may also give rise to a number of closely related historic groups (Noble 1971:104). Noble's Canadian Tundra Tradition contained ". . .four late small tool complexes which may be considered part of a regional tradition within an overall generalized Arctic Small Tool co-tradition. Seriation and radiocarbon dating suggests a 1000-year period of develOpment for the four complexes, between 1200-200 B.C." (Noble 1971:107). The finding of these complexes extending .up to 150 miles within the present tree line" (107) suggested ". . .a cau- tious and slow movement inland of taiga-tundra adapted peoples" (Noble 1971:107). The earliest complex, Rocknest Lake dated 23- 1200-900 B.C., contained triangular end blades, microblades and rare burins (including a polished burin) and polished adzes, but the most distinctive aspect was the presence of oval bifaces of quartzite which tie this complex rather directly to Taylor's southern Victoria Island sites (see above). "Rocknest Lake, therefore, is considered to be a late contemporary or a derivation of the more northerly Ekalluk River sites mentioned above" (Noble 1971:108). Subsequent complexes in this tradition continue and in— crease the frequency of polished burins while the quartzite a aqua ""-'\ b'vuu -C u ‘ TV‘V: A N &. i -..L V I nfl?’-;~ ab. c...-.1 ' o 5-‘9 “‘1‘. -.¢: ...L- (r; 5 u; 'o o N. ,. v "9““ A ‘ “"na a: an» 3. b.‘ . . A I‘. ‘igcb d .. \fi - .‘01gL . . A .4 -‘ s//§.“C> .: .‘VA.S . 61 bifaces, microblades (some of quartz crystal), and small concave-base points continue and bear resemblances to Irving's Keewatin sites and Nash's Twin Lakes assemblage in northern Manitoba (see below, Nash 1969). Noble discussed the entire tradition, concluding that it probably arose from the more northerly Buchanan complex of Taylor and pushed ". . .eastward into southern Keewatin and probably gave rise to Nash's (1969) local expression of Twin Lakes" (Noble 1971:110). Noble found the appellation "Pre-Dorset" ambiguous when applied to these complexes, especially since it would be difficult to see many of the "southern" traits as ancestral to Dorset. Noble saw the Canadian Tundra tra- dition as resulting from Arctic Small Tool tradition cul- tures but not develOping into Dorset, yet the final phases, with their inland orientation and forest adaptation (Noble 1971:110) have no clear continuity into subsequent tradi- tions. With their ethnic identity uncertain, the Canadian Tundra peoples remain an enigma which is difficult to solve, given the nature of the data with which we are forced to deal and of the paradigm of assumptions, concepts and methods we have applied. Gordon discovered 61 sites (remarkably, seven were stratified) on the Upper Thelon River where extensive exca- vations were conducted at Kij-6 and 7, including several levels of Arctic Small Tool occupation (Gordon 1972). Arctic Small Tool Traditio Variant Level, dated at 390-§5 (T-5g75 an -95 (I-59735 B.C., bears close resemblances to Arctic Small Tool Tradition .— H prese n Prty .- ,‘1.~ Lo: LO» 62 (ASTT) materials from the Buchanan, Menez and Uminga [sic] on Victoria and Banks Islands. Fewer affinities exist between the Thelon sites and the ASTT sites of Noble and Harp in the western barrenlands (Gordon l972:3). The presence of copper ulus and the large quartzite indus- try document the widespread extent of this regional variant of the Arctic Small Tool tradition at about 1000 B.C. It is noteworthy that the botanical analysis is indicative of at least partial forestation during this period (Gordon l972:3), although the specific dates are not yet published. It is apparent that the tundra-boreal forest ecotone has played a major role in the cultural dynamics of the Districts of Mackenzie and Keewatin and that we must be cautious before concluding anything about the relations among the varying Arctic Small Tool assemblages in this region. Irving conducted additional surveys of northern Manitoba and southern Keewatin in 1960, 1963 and 1964 finding additional sites of the Arctic Small Tool tradition (Irving 1968:36,46), although a full report has not been published. Irving did discover the Twin Lakes site near Churchill (see below), concluding that both relatively early and relatively late Pre-Dorset occupations had oc- curred: "The forms of very distinctive implements, in- cluding burins, microblades, and very small, bifacially chipped, side-shafted knife blades and weapon points, com- pare closely with Pre-Dorset (Carlsberg) cultures of the central Arctic, which probably date between 1,500 and 2,500 n h*v-O..VQ A b— v 55..-..- . ._, VVV &. ‘ .“ “«A.“ 50. -34- . :‘U‘ea ..' 5.x» ‘ U be‘ *Y. 1 ‘ “'L".~ 'N ‘ w A 0 5., .. .g‘e ’f‘ «-3 ‘zfi,“c..‘ A“ o u - _ “gigag‘ J:-: “. q 3" V V I~.. ’ ..:va‘ 6‘. 63 B.C. . !'(Irving 1968:46). "Thus, here, as elsewhere in the Canadian north, the Arctic Small Tool tradition spread to its farthest geographic limits (Churchill and the tree- line) very quickly and remained in possession for several centuries at least" (1968:46). Irving, recognizing the implications of the use of the concept Pre-Dorset in these areas, also used Taylor's term Carlsberg culture to refer to these Canadian Arctic Small Tool tradition peoples who may have only indirectly influ- enced the evolution of Dorset culture. In this article, Irving points out clearly that these early Eskimos were not the first people in the barren grounds and that subsequent populations in the interior were not derived from this tra- dition (note that this confirms Harp‘s general conclusions, 1961). Irving's work at the Twin Lake site was followed up by Mayer-Oakes in 1964 and by Nash in 1965. Nash conducted extensive research into the Pre-Dorset occupation of northern Manitoba, reinvestigating Giddings's Thyazzi site, and working at Seahorse Gully, near Churchill. Nash sum- marized his work in 1969 relating the Twin Lakes assem- blage to the late Pre-Dorset of eastern Canada, in particu- lar to the Davidee site which Maxwell has categorized as possibly having mixed components (Maxwell 1973az317-318). The predominance of a notched and ground burin industry strengthens Nash's conclusions. Seahorse Gully was seen as somewhat younger than Twin Lakes, and its large tools of \ f‘ L 3: Ya. ‘ 0 u ‘ ~ n , ‘ - F e“: p- y :13 Joe. 33:3 38 3 ‘0 61,- s; .OK C . i ’1‘ 1 .L n‘ ‘ HI... ‘4‘ M P .‘J e n ‘ k . V K 41L . {A :3 w. u .7! 1 L L . a,» v x a c O u. a n a . . h.\U A nu a v s s ¥ s M» \ MN“ MAIN u ~‘k‘ “ IMN fin \ wmdh .u‘ N t‘ (\. . a a 7.1. "I .J. 1. u S 5 1L /\ u Pk hm .. a! v. c. 64 chert, granite and pyroclastic stone (Nash 1969:143) were thought to reflect the possible availability of wood or contact with Indians in the adjacent forests (1969:143). It is significant that these large stone tools were not seen by Nash as being either directly or indirectly related to the large stone industry of Noble (Canadian Tundra tra- dition) or Taylor (Banks and Victoria Island Pre-Dorset) (see above). Subsequent analysis by Meyer (1970) of the Seahorse Gully Pre-Dorset component has confirmed the general conclusions of Nash, with the additional evidence from the analysis of internal divisions of the site and of the radiocarbon date on seal bones of 23- 3000 years B.P. (3-251, 2900t100 B.P., 93. 945 B.C.) which has pushed it back further in time (Meyer 1970:167). The presence of the large tool inventory makes these assemblages as distinctive within the Arctic Small Tool tradition as any other and provide more data on diversity within continuity. Nash's 1969 volume also contains a relatively complete depiction of the then state of our knowledge of the Arctic Small Tool tradition complexes and affinities and his statement of problems includes the role of the tree line in limiting the southern extension of these early Eskimos (Nash 1969:155). Recent survey has revealed additional Arctic Small Tool sites further eastward along the shore of Hudson Bay within the present limit of trees, so we must await further refinement of paleo-botanical sequences in this area (as along the Thelon, see Gordon above) to find , . wvcfl‘.‘ ,_ .- EIova‘ (.W . .re- .1... e... -..t. H O“‘ h’ . A *7) 5.. 'fi‘“. J “b I ’l (D I) 65 evidence to test these ideas. 0n the East Coast of Hudson Bay, Plumet discovered a Pre-Dorset site near Great Whale River radiocarbon dated to EE- 1300 B.C. (personal communication) which seems appro- priate for the burins, burin spalls, microblades, and end scrapers that I have seen through his courtesy. Structural remnants are rectangular (Gosselin et a1. 1974) and there are no large tools from his present sample. We must await further research and publication in this area by Plumet and Harp. Fitzhugh has conducted an extensive series of excava- tions around Hamilton Inlet in Labrador and further north in the vicinity of Nain, where a few Pre-Dorset sites were found, but none were found further south. The artifacts were notable for the usual dominance of burins and burin spalls, the inclusion of a ground celt and an ulu fragment, a paucity of microblades and a variety of end blades, in- cluding contracting stemmed bifaces (Fitzhugh 1970; 1973). This Thalia Point site was radiocarbon dated to l710t140 B. C. and seemingly represents the earliest Pre-Dorset in this region. Fitzhugh has also reported a significant find in a Maritime Archaic context from Hamilton Inlet. The Rattlers Bight site contained a single Pre-Dorset burin made on Ramah Chert which is the only evidence for such contact in the area and is far to the south of the known distribution of Pre-Dorset along the coast. Because the Rattlers Bight site has dates of EE- 1900 B.C. and 2600 B.C. (Fitzhugh WKJ ad] «.I. U A," U.” -0. C 1 H .h. ‘4“ ad v. O .- Aflu C. D. .d .C a3 1|. .1. n fi‘vn" CLU3‘- fi .38 36 c Q~L 0v.- “- A—e..3 8 vi a) 4L 66 1973) and contains no burin industry, this is evidence for both contact of some nature between these diverse peoples and for the early presence of Pre-Dorset within (or at least in contact with) the range of Maritime Archaic. This is partially substantiated by the presence of several flakes of Ramah Chert (apparently found only in northern Labrador-Jsee Fitzhugh 1972) in the debitage from the Closure site (see below) (I am indebted to Fitzhugh for the personal identification of these flakes). As Nash concluded from his analysis of the Pre-Dorset extension southwards along the west coast of Hudson Bay it is remarkable that these peOple carrying the Arctic Small Tool tradition spread along the tundra coast to its limits, either as defined by the treeline (or other biological limits) or by the presence of indigenous peoples. Tuck excavated at Saglek Bay, northern Labrador be- tween 1969 and 1971, and has reported Site Q, dated 1880t115 B.C. (Tuck 1973). The Pre-Dorset site K contains a small tool assemblage dominated by burins, burin spalls, and microblades (totalling 67% of the assemblage, after Tuck 1973) with grinding present on a single burin and lacking steatite lamps or bowls (the sample is sma11--N=150). There is a distinctive complex of end blades and knives, characterized by fine workmanship, edge serration, and thin cross-sections. The small, thin contracting stem end blades are similar to those found by McGhee in the high Arctic (see below) and rarely at other Pre-Dorset sites in 67 the Eastern Arctic, while other artifacts (burin-like tools and triangular end blades, for example) were typically Pre- Dorset. However, Tuck saw Pre-Dorset affinities as less than those with McGhee's northern Independence I complex, which possibly evolved into his Site K complex (Tuck 1973) (see McGhee below). Tuck saw no continuity between his Site K Pre-Dorset and later occupations, thus seeing two migrations into northern Labrador. In short, the link that Tuck and McGhee see between Independence I and the northern Labrador Pre-Dorset is predominantly in end-blade form, and we will examine this proposition in some detail later in this thesis. Rousseliére continued his archaeological research in the vicinity of Pont Inlet, northeast Baffin Island, where his Oqalik site was thought to be early in the Pre-Dorset continuum (based on the lack of polish on 17 burins-- Rousseliére l968:40,43). His Mittimatalik site (PeFr-Z), containing a similar assemblage, was radiocarbon dated as one of the earliest dates in the Eastern Arctic (8-589, 4385t155 B.P., 2a. 2435 B.C.) but the seal bones dated may cause the sample to be somewhat older than the actual date. The Pre-Dorset material, while present, is scanty and its preliminary nature does not allow conclusions (Rousseliére 1973) other than the fact that an early Pre-Dorset occupa- tion comparable to those from Igloolik and southern Baffin Island was present in this region. McGhee conducted an extremely important archaeological 68 reconnaissance of the south coast of Grinnell Peninsula of Devon Island and nearby Dundas Island in the summer of 1972 and his Independence I and Pre-Dorset assemblages are impor- tant data for the development of this thesis. McGhee's Independence I sites had apparently both summer and winter dwellings and were categorized by the presence of stone out- lines and slab mid-passages (see Knuth 1967 and above). The artifacts included unpolished burins, frequent micro- blades, many end-blades (including thin triangular and bi- pointed forms reminiscent of those from Tuck's Saglek Bay site K, see above) with often deep edge serration, and a non-toggling form of harpoon. McGhee infers coastal sealing as the major subsistence, from the prevalence of seal bones, while caribou and musk ox also occur. The moderate size of the sample (N=267) and lack of complete excavation suggest that negative evidence may again be of relatively little value (McGhee 1973). McGhee dates this occupation as coeval with that excavated by Knuth from Ellesmere Island and northern Greenland, probably 93- 2600 B.C. The settlement pattern comprises strings of adjacent houses along raised beaches. McGhee's Pre-Dorset assemblages are somewhat larger (N=48l) and their house forms less-descript without evidence for internal structures or clearly marked walls. As ex- pected, burins, burin spalls, and microblades comprise almost two-thirds of the assemblage and only two burins have polishing. The raw materials are different from the 69 Independence I artifacts, and the Pre-Dorset appear some- what cruder, lacking edge serration and lacking the thin triangular and bi-pointed end blades. Their subsistence patterns appear to be similar to the Independence occupa- tion, but their settlement pattern is not, being house clusters rather than alignments. The somewhat lower average elevation of the Pre-Dorset is taken as an indica- tion of somewhat later age, and McGhee suggests possibly 1700-1500 B.C. (McGhee 1973). The finding of these two distinct complexes in situa- tions where their settlements overlap on the same beaches was not expected in this area, as our previous thinking suggested a clinal relationship between regional variants of the Arctic Small Tool tradition in the Eastern Arctic. McGhee, emphasizing the distinctive characteristics of each assemblage, suggests that they represent two distinct mi- grations from the Western Arctic, with the Independence I being the earliest occupants of the far North, followed by migrants of Pre-Dorset people from the core area (McGhee 1973, 1974). McGhee also suggests that this earlier Inde- pendence I migration was not limited to the high Arctic littoral, but may have spread over a large area of the Eastern Canadian Arctic, where their sites have gone undiscovered because of their linear settlement pattern (McGee 1973, 1974). Within the core area, these Indepen- dence I peoples were displaced or acculturated by a second migration, this time of Pre—Dorset peeple. Jae n 1" a“? 5);. LA‘ [2 O ’5‘ r» 4 :L r e s .t a v . 2 e c a.» t O .i a n0 .f y a 4 a .5 . a 1 1 ~ , E {L E .0 Au f. 1 C S S S 3.11. 7O McGhee applied concepts derived from the study of Historic Eskimos to attempt an understanding of the social dynamics that might make such technological and social changes comprehensible. He grouped bands of 22- 50-100 people into larger "maximal breeding units" of perhaps 500- 1000 people with distinctive cultural systems setting them apart from similar groups in adjoining areas. He called these larger groups "tribes" and believed them to categor- ize the diverse regional archaeological assemblages that we have observed in the prehistoric record, following an analogy from the study of the relations between regional groups during the Historic period (McGhee 1973). Aside from the importance of McGhee's archaeological data from the Port Refuge area of Devon Island, his 1973 discussions with regard to the population dynamics of the people who made these artifacts mark an important shift in the nature of our discussions of the Arctic Small Tool tra- dition and of Eastern Canadian Arctic archaeology. The fur- ther discussion of these matters will be deferred until later in this thesis, but McGhee's ideas make a fitting close to this preliminary history of the study of the Arctic Small Tool tradition. As a result of the above assessment of our past re- searches in this area, I see Pre-Dorset as an Eastern exten- sion of the Arctic Small Tool tradition, in which the dis- persal of the population marked by the Arctic Small Tool horizon resulted in a diverging and elaborating horizon and § 3:: 3‘8? 022‘ theor A y 0 A- n". Y ' ~13 r .. L T. . A ~ ., r . . Au Wu Av e 1..“ MW 6Q 5... L; .C “u no a. at «u at .an .a« yr cc AC . a .1. I E D. E «I O a S h. J t. ed .cn «.an «U 0 3. may (x a 71 tradition (Wauchope 1956:43), with complex internal dy- namics and interrelations among its constituent populations. These dynamics were influenced by the diversity of en— vironments in which the horizon extended and by changes in these environments through the time of the tradition. The understanding of changes in Pre-Dorset culture which emerges from the survey of the literature, is one of great complexity. We should not, therefore, expect to reach such understandings through the use of a small number of simple models. Because I anticipate the necessity for a number of complementary and complex models, the methodology and theoretical framework in which such studies must occur de- mand further attention. The last several years have seen an increasing concern among Arctic archaeologists for archaeological method and theory (see Dekin 1973az4l-42). In part, this is because Arctic archaeology has been a classic example of "The more we know, the less we know" because as we accumulated more knowledge of the diversity of evidence for behaviors, we became less sure of what we had known before, re-opening old questions with new data (McGhee 1973 is a case in point). Several of us have tried to make our methodologies more Openly explicit, so that others could examine them (see Nash 1973, Fitzhugh 1973, Dekin l972a, l972b, 1973b), and the recent Seminar on Pre-Dorset -- Dorset Problems at the School of American Research was notable for the concern ‘ Vv.‘ A .‘.¢. L" 72 Table 2 Summary of ASTt Field Research: Eastern Arctic Author Solberg Thostrup Meldgaard Larsen and Meldgaard Mathiassen Knuth Giddings Meldgaard Harp Harp Lowther Taylor Rousseliere Knuth Maxwell Taylor Irving Nash Fitzhugh Plumet McGhee McGhee Noble Meyer Gordon Tuck McGhee Rousseliére Publication Dates 1907 1911 1952 1958 1958 1958 1956 1960,62 1958 1959,6l,62 1962 1962,65,68a 1964 1967 l962,67,73a 1964,67,72 1968 1969 1970 unpublished 1970a 1971b 1971 1970 1972 1973 1973 1968,73 Map Key CO CD \lO‘ U'l bu WWNH H l2 14 15 17 18 19 15 21 22 23 24 Location or Sites Excavated West Greenland Northeast Greenland Disko Bay, West Greenland Disko Bay, West Greenland Sermermiut, Sarqaq Disko Bay, West Greenland Independence Fiord, Northeast Greenland Northern Manitoba Thyazzi Igloolik, N.W.T. Coronation Gulf, N.W.T. Dismal-2 Thelon River, N.W.T. Complex A/Phase 2 Cape Sparbo, N.W.T. West Hudson Straits, N.W.T. Ivugivik, Arnapik Pelly Bay, N.W.T. St. Mary's Hill Ellesmere Island and Northeast Greenland Independence I and II Frobisher Bay, N.W.T. Lake Harbour, N.W.T. Pre-Dorset Banks and Victoria Islands N.W.T. Pre-Dorset Southern Keewatin Northern Manitoba, Twin Lakes, Seahorse Gully Labrador Coast, Thalia Point Great Whale River, Quebec Bloody Falls, N.W.T. Victoria Island, N.W.T. Central District of MacKenzie Canadian Tundra Tradition Northern Manitoba Seahorse Gully Thelon River, N.W.T. Labrador Coast, Saglek Bay Site K Devon Island, N.W.T. Port Refuge Independence I & II Pond Inlet, Baffin Island, N.W.T. Oqalik, Mittimatalik 73 ASTt Sitos (.20 u. L . = . I —_ ____ Figure 2 Sites from Table 2 express« we were through snagged At: methods the unaz Sistence 74 expressed and discussed for method and theory. Seemingly, we were striving to go beyond the study of technology through time and Space, and we were more than occasionally snagged by inadequate or unavailable data. Attempts to utilize perspectives, techniques, and methods of ecological systems research were frustrated by the unavailability of good quantifiable data on the sub- sistence of Arctic peoples, although a number of suggestions were made on a somewhat more general level of abstraction (see Fitzhugh 1973, Dekin 1973b). General systems theory has also been used in an attempt to generate hypotheses for testing with regard to prehistoric cultural systems (Dekin 1972a) and as a framework for the consideration and display of data (Nash 1973 used a more inductive approach linking his generalizations with the data, while Dekin's hypotheses could be tested on the archaeological data as they were de- veloped independently). The utility of these approaches remains to be seen, as does the level of their acceptance by other Arctic archaeologists. Recent studies have also demonstrated the interest of Arctic archaeologists in utilizing the implications of functionalist anthrOpology for archaeology (see Leach 1973: 762; Binford 1973) where Maxwell has attempted to assess the activities at the Morrison site (1973b) and Dekin has attempted to define patterns of structural design and in- ternal division into activity areas at the Closure site (Dekin 1973b and this thesis). However, these attempts at analv: in: groups ah and are ti 75 analyzing the internal behavioral dynamics of social groups which produce sites have been the rare exception, and are themselves still in the developmental stage. The discipline remains dominated by a culture- historical traditionalist paradigm (if one exists; see above) in which the unit of analysis is the site-assemblage of artifacts. There is, however, increasing concern with ecological relationships, with sampling techniques, with representativeness of samples, and with data preserved, pre- sented and published. This thesis is a further step in this direction of methodological change, in that a major concern is in the behaviors of individuals within archaeologically-known cultures--an area of interest in which some have argued we would make no contribution, and thus collected andpublished their data in a manner which occasionally precludes and usually hampers such study. Recapitulation The state of the art of producing generalizing concepts is unclear, as we have gone for some time accepting and using concepts differentially in our analysis and it is often clear that the meanings are not shared by all of those working in the field. Paleo-Eskimo has been used as a catch-all to include almost all Arctic cultures which were seen as developmental from a Denbigh Flint Complex-like base and not yet de- veloped into Thule (See Mathiassen l958z3). As a chipped- E11185 1:31; better 3; 76 stone complex, this also is too general to be of much use and it is difficult to see what utility this concept has, in that it does not define human behaviors accurately (or precisely) enough for operationalizing research. In the Eastern Arctic, it may still possibly be useful to refer generally to pre-Thule Eskimo cultures, but only on the most general level. In the Western Arctic, the diverse and dynamic cul- tures which would be included in this category would be better served by greater specificity, and it is interest- ing to note that the concept has not received much use or attention in the West when compared to its rather inter- mittent use in the East. In short, its recent use in the Eastern Arctic does not seem apprOpriate, and I will sug- gest some alternatives later in this thesis. The Arctic Small Tool tradition has undergone subtle changes in its conception and definition since its first development. It now appears that the homogeneity inherent in the concept of tradition may not apply to as long as time depth as initially perceived. As I will suggest below, the data are supportive of an Arctic Small Tool horizon, spreading rapidly across the entire North American Arctic between 2100 and 1900 B.C., which should subsequently be considered possibly as both a diverging and an elaborating tradition (see Wauchope 1956:43) with particular attention to the development of diversity within the tradition, which, toward the upper limit of its existence, has lost all implicatio It is most usefu Dorset co: core area will devel With Pre-j 77 implication of spatial homogeneity. It is apparent that the concept of Pre-Dorset would be most useful when applied to a restricted area in which Dorset could be seen to develop from it, in this case the core area of Foxe Basin and Hudson Straits. To do this, I will develop categories to delineate sub—traditions co-eval with Pre-Dorset, but which were not directly ancestral to Dorset culture. Noble's Canadian Tundra tradition may qualify for such a tradition. Taylor's Carlsberg culture may provide a useful dis- tinction between the Western deve10pments from the Denbigh Flint Complex and the Eastern developments, but it requires subdivision to categorize the cultural diversity present in Greenland, the high Arctic, the core area, and southern and western peripheries, and we will look carefully at such schema to develop concepts which will be inductively sound and operationally useful. It is apparent from the above discussion that while the concepts used to generalize about Eastern Arctic pre- history have been used with relative explicitness, they lack precision in both their definition and use. As the corpus of data to which these concepts refer has grown, the con- cepts have not been re-defined but largely reified, and we are faced with a number of different concepts none of which seems to fit the data well. There is a need for greater explicitness and precision in the use of the general con- ceptual framework in which we wish to work towards a 78 better understanding of prehistoric cultural dynamics in the Eastern Arctic. Nowhere is this need for a precise conceptual frame- work more obvious than in the consideration of the temporal and environmental contexts in which the human occupations of the Eastern Arctic occurred. The deve10pment of explana- tory models of human behavior must be based on a precise and accepted time frame, especially when we are dealing with complex sets of ecological relations (Dekin 1973az4l). It is essential that our treatment of temporal and environ- mental data be precise and explicit, since differences in this basic framework are a considerable source of variation in interpretation. Part II will develop a precise and ex- plicit depiction of the chronological framework of archaeological data and of the changing environmental setting between 4000 and 2500 B.P. as a basis for the de- velopment of more complex models of human ecological rela- tions (see Part III). .p‘ quay- Cusafibt- h “a C3133: 91' PART II THE SPACE/TIME FRAMEWORK Chapter II. Dating the Arctic Small Tool Tradition Chapter III. The Changing Environmental Setting 9: . ,‘ b-0n Y‘aa Chapter 2 Dating the Arctic Small Tool Tradition The first formulation of the Arctic Small Tool tradi- tion was based largely on the assumption of time depth, even though there were then no absolute dates for any mem- bers of this tradition and the only relative dating avail- able was that for the type site of the Denbigh Flint com- plex at Iyatayet where Denbigh was stratigraphically separable and overlain by Norton and Nukleet strata. However, the thinking on the deve10pment of this tra- dition has evolved hand in hand with the increased knowl- edge of radiocarbon dating, thus most of our interpretations have developed with radiocarbon dates available for many of the assemblages in the tradition. Wherever possible, our chronology for the tradition has been based largely on radiocarbon dates. In specific locations, these have been supplemented with other techniques of relative dating, most of which in- volve cross-dating from chronological sequences derived from other sciences. The geology of the Arctic has provided a powerful tool for cross dating because of the isostatic rebound of the land rising after the ice sheets had melted. 79 I VAF‘A ' ‘-\.L-\. C present in the a general} (1962 1‘ ‘ e Pie * ‘9 Ea:- C .- 19V?) 1: 79 This possibility was first used in the Eastern Arctic by Mathiassen (1925:206) when he attributed considerable age to house ruins found at higher elevations and also further distant from the shore. On a rather simplistic basis, Arctic archaeologists have often used relative height above present sea level as a measure of relative age, especially in the absence of other relevant criteria. This practice is ubiquitous, in spite of the cautions voiced by Collins (1962:128-129) who considered the method useful ". . .and generally valid when no complicating factors are involved” (1962:128) (see Taylor l968a:98-99). Sequences of beach ridges raised in elevation above the present sea level have been used effectively in both the Eastern and Western Arctic (Meldgaard 1962; Giddings 1960, 1961, 1966) but this is useful in those relatively rare occasions where distinct beaches with definable occu— pations were preserved. It is also obvious that these data are relevant only in those areas in which the relative sea level has varied and only along the coast. Strictly speaking, age estima- tions of this sort can only be minimum ages of the occupa- tions of the raised area, as has been effectively pointed out by Andrews et a1. (1971) in their compilation of up- lift isobases and relations with elevations of archaeologi- cal sites. Formal attempts at the seriation of archaeological data with the goal of establishing relative sequences of 8O assemblages are virtually non-existent. In part, this may reflect the general decline in the use of formal seriation methods in North American archaeology that occurred after the advent of radiocarbon dating and before the increased utilization of computer technology to speed the data pro- cessing. Taylor's 1968 seriation of the West Greenland Sarqaq and Dorset sites reported in Larsen and Meldgaard, 1958, is perhaps the most formal attempt and even he acknowledged the apparent bias in favor of cultural con- tinuity inherent in the method (Taylor 1968az9l). Informal seriations of data have involved: 1. the establishment of synchronous horizons of arti- fact style or type (attribute or attribute set); and 2. the establishment of trends in artifact frequency through time. Thus, Dorset was seen as a horizon of the following set of artifacts: burin-like tools; microblades; steatite lamps and pots; isosceles triangular end blades; rectangular houses; ground-slate projectile points and knife-forms; side-notched flint end blades; etc. \1 mm-wat-I Yet, this horizon did not mark a clear cut separation from earlier artifact forms and sets, as Meldgaard's data from Igloolik clearly indicate (Meldgaard 1962:P1ate 5). Harp's categorization of traits of Dorset culture was a statement of primary traits that might appear in some locales as horizons, but which appeared in his Newfoundland sites as a fully- was revo part on the anal mess of the arch “a .1: Arctic S 1. 2. £\ 81 a fully—developed complex (Harp 1964:94). The issue of whether the development of Dorset culture was revolutionary or evolutionary apparently depended in part on the nature of the microscope (level of abstraction) the analyst was using, as the relative speed and pervasive- ness of cultural changes were subject to the judgment of the archaeologist. The following trends were perceived in the Eastern Arctic Small Tool tradition: decline in the frequency of true burins; increase in the frequency of side notching on artifacts, including end blads and burins; increase in the frequency of grinding and polishing as construction techniques; increase in the use of quartz crystal as a raw material; increase in the use of steatite for lamps and pots; etc. U‘J—‘WNH It is interesting to note that those archaeologists who saw the major cultural changes in the Eastern Arctic as re- sulting principally from migrations were prone to utilize a guide-fossil approach (Dorset as a horizon) (see Mathiassen 1985; Larsen and Meldgaard 1958; Meldgaard 1952; Knuth 1967) while those archaeologists who saw a sequence of development from Pre-Dorset through Dorset were more concerned with "trends" rather than type fossils (see Taylor 1962; Maxwell 1962; Taylor 1968a). In a real sense, the techniques of serial dating utilized in the Eastern Arctic reflected the theoretical biases of the investigators, in most cases only little affected by what limited radiocarbon dating had been done at the time. .—a VA. velop: trends -‘Je Of (COTDar 82 However, rather little effort was expended on the de- ve10pment of formal statements of horizon markers or trends, in large measure because there were few locations where extensive series of sites demonstrated great variance through either space or time. Most Arctic Small Tool tra- dition sites listed in Tables 1 and 2 are small isolated finds not directly related to a regional developmental se- quence. The notable exceptions in the Eastern Arctic are the researches of Knuth in northern Greenland and Ellesmere Island, Meldgaard in the Igloolik area and Maxwell in the Lake Harbour region. Of these, Maxwell has provided the best published examples of a regional sequence, while Meldgaard has provided a preliminary report and some con- clusions and Knuth has treated his material as representa- tive of relatively static stages of cultural development (compare Maxwell 1973a; Meldgaard 1960, 1962; and Knuth 1967). Meldgaard stressed the breaks in his sequence of occupations while Knuth stressed their homogeneity utilizing a general model of cultural replacement by migration and Maxwell utilized a model of general cultural evolution in situ. Thus, Meldgaard used a trait list of characteristics, Knuth stressed those characteristics that differentiated his regional occupations, and Maxwell discussed more-or-less cyclical variations about modes of stylistic variation. Maxwell has come the closest to providing data for the establishment of horizon markers, yet his emphasis on cul- tural continuity stresses the lack of clearly defined 83 horizon markers of cultural change and he further empha- sizes that there are few instances where several items of change seem to occur coevally to mark a period of greater change and a horizon. The study of Arctic Small Tool tradition assemblages in the Western Arctic has been a history of "splitters" (Dumond l974b). While some have seen cultural continuity, with changes, through the Denbigh Flint complex, Choris, Norton and Ipiutak and on to proto-Eskimo and even Eskimo traditions, none has presented a synthesis emphasizing the cultural continuity and demonstrating the existence of horizon markers or any evidence for similar changes across a large geographic expanse. Further, the greatest interest in Arctic Small Tool tradition sites in the Western Arctic has been expressed by Anderson with his developmental se- quence at Onion Portage from Proto-Denbigh to Denbigh. In the last ten years, very few Denbigh (or Arctic Small Tool) sites of significance have been excavated or fully reported, largely because other problems have dominated the research interests of Western Arctic archaeologists, and Arctic Small Tool sites have been by—products of other research programs and reports (see research by Stanford, Humphrey, and Alexander: McGhee 1971a; Humphrey 1970; and Alexander 1969). Thus, the Western Arctic has been dominated by distinctive cultures and changes were seen as large enough to indicate breaks in sequences of cultural development. Chronologies were developed from beach ridge sequences on the coast 84 (Giddings 1960, 1961, 1966) and, since 1963, from strati- fied sites in the interior (Giddings 1966; Anderson 1968a, 1970) relying on artifact similarities and radiocarbon dating for cross-dating. Perhaps the greatest change in the Arctic Small Tool tradition in the Western Arctic was the addition of pottery, presumably from Asia, contributing to Choris culture. The continuties between Denbigh and Choris were recognized, but the cultural assemblages and their respective cultures were seen as being different, and not simply as a continuation of Denbigh with pottery. The Western Arctic, being adjacent to four contiguous regions with distinctive cultural traditions (The Aleutians, Southwest Alaska, Interior Alaska, and the Asiatic Coast), was influenced by ideas stemming from.migrations and other movements of people apparently more frequently than was the Eastern Arctic, which may be seen as a relatively closed system during the Pre-Dorset -- Dorset period (Maxwell 1967, 1973a). The prehistory of the Western Arctic has been marked by a number of distinctive migrations with clear horizon markers, stratigraphic separation, and only general threads of cumulative sequential development. In such circumstances, radiocarbon dates have been the major source of absolute chronology, with cross-dating from regional sequences as the major comparative tool. This reliance on radiocarbon dating for the establish- ment of a regional chronology has resulted in rather spotty rEfErr; 85 chronological control, as the well-dated sequences are widely Spaced and can only be linked by cross-dating using a guide—fossil approach. The general approach to radiocarbon dating has in- volved the following analytical steps, after the receipt of the date from the laboratory. 1. Convert from conventional radiocarbon years before present (B.P., 1950) by subtracting 1950 years. 2. Accept or reject date based on previous finds and prior speculation or conclusions (a priori). 3. Draw rationale for acceptance or rejection of date (including bad association, contamination by more recent materials, fractionation by natural processes, differences in the radiocarbon reservoir, etc.). 4. Operationally, consider the measure of central tendency of the date as if it were representative of a year in real time, ignoring the standard deviations and the probabilities associated with them, even to the extent of referring to the date by its measure of central tendency alone. This approach has produced a chronology which is rela- tive and which is generally accepted as a fairly close ap- proximation to real time, even though it has been known for at least fifteen years that the ratio of available C14 to 012 had changed through time, and that there would have to be additional conversions to obtain dates in actual years (real time) from radiocarbon years. Suess has led in this throsg recon: . - ¢ (m n-r these CWO p 1' 86 attempt to date tree rings (of known age) to define the re- lationship between radiocarbon years and actual years 12 ratio (which is based on the fluctuations in the Cl4 / C through time) and his curves have been used in attempts to reconcile apparent problems in regional chronologies (Renfrew 1971). However, it is important to note that these problems in European prehistory came as a result of two presumably absolute systems of time measurement: written history and radiocarbon dating. At certain times, these two systems did not mesh well and only after Renfrew made extensive conversions based on the Suess curve did some semblance of order result from the previous disagree- ments. However, it is important that we recognize that this situation was a special case resulting from the avail— ability of historic records in the area. The relative se- quence was largely unchanged and merely shifted somewhat. We have no reason to expect great changes in our regional chronologies based on radiocarbon sequences as a result of the application of a Suess chronology to convert to ap- parent real time, as we shall see below. There continue to be other sources of discrepancies in radiocarbon datings, resulting from.differences in the use of half-lives (5568 vs. 5730), in the application of asssess- ments of fractionation (C14, C13 , 012), in the use of 1950 as an artificial "present", as well as the inherent varia- tion in sample collection, pre-treatment, and the material itself. It is difficult to control all of these variables, 87 especially since the users of these data (archaeologists) are not conscientious in their reporting of what has been done to their dates (half-life, fractionation assays, etc.). Recently, McGhee and Tuck (1973) have attempted to de- velOp a regional chronology for the Eastern Arctic by con- verting dates in conventional radiocarbon years to years AD/BC using a Suess curve from Stuiver and Suess (1966) and Renfrew (1971). They attempted to reduce systematic variation by applying a fractionation correction of -400 years to sea mammal dates, but concluded by eliminating dates on any materials other than indigenous charcoal from the establishment of a chronology. In part, this was necessitated by the spotty availability of information on fractionation corrections that have already been made be- fore release of the dates, and by our lack of precise in- formation on the nature of variations in materials from different life zones. The major impact on a series of dates from conversion to "real time" using a Suess conversion curve is the change in clusters of dates through time, as certain periods of time are "stretched" and others are "shrunk" vis-a-vis dates in conventional radiocarbon years from these periods. Thus, in a radiocarbon chronology, time passes more slowly at certain times than others, clustering dates from.more widely in time. In general, conversion has little effect on relative positions in time, especially if the numbers of dates are relatively small and thus relatively dispersed throu conve‘ dates HEW Vi 88 through time. Perhaps the greatest contribution of such conversion is in the increased alternatives to certain dates which might otherwise appear anomalous. However, this new variance is often less than the original variance of the date itself (standard deviations). Tables 3 and 4 are summaries of dates associated with cultural materials from the Arctic Small Tool tradition in both Eastern and Western Arctic as of January 1974. The data presented include the original date expressed in con- ventional radiocarbon years, other characteristics of the date and its material, a conversion to "AD/BC" by sub- tracting 1950 years, and a conversion to Suess-curve dates and the alternatives. With regard to interpreting these data, I would empha- size that the relative order is only rarely threatened by alternatives and that the Eaggg of alternatives is by and large less than the two standard deviations range of the original date (of the 85 dates in this sample, the ranges of 76 are less than two standard deviations, the ranges of 8 are between two and four standard deviations, and the range of only 1 is greater than four standard deviations) which supports the following conclusion: the ranges of the alternatives created by conversion to the Suess curve demonstrate no more variability and dispersion than the variability and dispersion of the original date determina- tion itself. Conversion thus does 325 change the variabili- ty (and therefore imprecision) inherent in the dating 89 process. This point is significant in light of the diffi- culty which some archaeologists have with regard to inter- preting such imprecision, as they perceive it. One drawback to the use of the Suess curve, is the reluctance of archaeologists to reduce voluntarily the apparent precision of their dates by recognizing that there may be several al- ternative choices. In most cases, I believe that this is indicative of a general failure of archaeologists to deal adequately with the dispersion of the original date in radiocarbon years. We have not generally c0ped with the stochastic processes inherent in radiocarbon dating (Polach and Golson 1966). Spaulding pointed out in 1958 that statistical tech- niques were available to allow the testing of the proba- bility that two radiocarbon dates dated the same event (were coeval) or dated different events. Because this testing is not usually applied and represents an important contribution to our understanding and assessment of radio- carbon dates, I will reproduce this method here in some detail. The essence of the technique is the assessment of whether or not the two dates under consideration could be samples drawn from the same population. The statistic "t" is the expression of the difference between the two mea- sures of central tendency (means of the "samples") as the number of standard errors of the difference between these two dates. The standard error is computed in the H] C) H 90 following manner: i Date 1: ml sd1 m is the mean t Date 2: m2 sd2 sd is the standard deviation Standard error of difference: /(sd1)2 + (sd2)2 t = Iml ' m2 Std. Error The probability of such a value of t occuring by chance in two samples drawn from the same population is obtained by reference to the curve reproduced in Figure 4 (Hodgman 1958). This procedure allows us to test whether two dates are coeval, and to assess this probability. As an example, Dumond assesses the evidence for the be- ginnings of his Brooks River Gravels phase, rejecting a pre- viously held valid date of 2022 B.C. t 440 (Y-930) on the basis of association, using the date of 1950 B.C. 1.130 (I- 1629) as the earliest date of the Brooks River Gravels Phase, which he assesses as 1900 B.C. t 100 (Dumond l97lz9). As part of his evaluative process, it would have been appro- priate to assess whether or not the two dates above could have been coeval. fl The standard error is: ‘/(440)2 + (130)2 = 458.8 t = h022 - 1950i= _1g__ 458.8 458.8 = 0.1569 From Figure 4, we can see that such a value of t would oc- cur ga. 87 times out of 100 by chance in two samples from the same population and the probability that the dates are coeval is p = 0.87. c.~.m< Quack/x :LeuILZIInSQQQ COQLCCOIXQ m. mNcafi. 91 hmm.mmz III. I lull III! mwmanammm< cowwmuwo wwww.q uem< owuou< cumummsllmmumn conumoowumm m manna e “m3 mm mummy !/ uFm< Uqu0h< Chflutfifllltflufin Zonhflunquian v QHSQL. 92 omaaaaouad .5338 33 ufim< u Hush—4 Ewunwullumuan aoauauoavam e 333. . .an; an menu» AV. «:3an Q .J~.—:L. 93 2:368 Q Show 94 § 0 O O C 'U .'-———-—d-—-—-—- 1000 BC (aha; Oluon 197°31'II'I Figure 3 Portions of the Suess Curve for Tree Ring Calibrations in th radior C0291! 0f ch. 95 The radiocarbon dates from Pre-Dorset sites analyzed in this thesis are listed in Table 5, as conventional radiocarbon years. However, their evaluation is a rather complex phenomenon. These dates are all on what are believed to be cinders of charred animal fat and sand produced by the burning of oil lamps, which are usually (historically in this region and today) thought to burn sea mammal fat (probably seal; the species is uncertain and is probably unknowable in the forseeable future). Dating of this material has proved somewhat different from other materials, and several sources of variability are possible. 1. Reservoir exchange--the marine ecosystem.may con- tain carbon somewhat older than the terrestrial ecosystem because of a time lag in the exchange of C 14 produced in the atmosphere with the dissolved carbon in the sea. With- in limits, this difference is probably almost a constant, although some differences could be caused by changes in circulation of atmosphere, storm tracks, open water, etc. The adjustment of these dates 400 years more recent in time as recommended by I. Olsson could be seen as an at- tempt to compensate for this delay in reservoir exchange (McGhee and Tuck 1973). 2. Fractionation--certain metabolic processes cause the pr0portions of the various isotopes of carbon to be ac- cumulated differentially and to be deposited in living material in a proportion different from.that present in the b: shoal: to CD4 nvir: Sever. H01d ‘ Ocean EXChar Ullderr, addit: Sea -... L . 96 the biosphere (atmosphere or water). This difference should be prOportional to the amount of carbon taken in and to the different proportions of the isotopes present in the organism's environment. This is not a time-delay phenome- non, but a constant ratio of the carbon composition of the environment to the carbon composition of the organism. Several figures have been suggested varying from 10% to 20% reductions in conventional radiocarbon years to com- pensate for this source of variance. 3. Variegation in the Marine Reservoir-~this is the "old water" phenomenon, in which it is recognized that deep 14 ocean currents receive C produced in the atmosphere in exchange with surface waters, and then go on about their underwater business keeping well beneath any subsequent 14 addition of new C until they upwell on a distant shore or sea rise possibly as much as several centuries later. Ob- 14 content of that old upwelling water will viously, the C have been reduced by the decay processes and any animals or plants in the food chain produced by that water and its re- sources will contain a diminished proportion of C14, thus dating somewhat older than their death. Migratory seals are especially prone to this sort of variation, and while they are not as numerous in the Lake Harbour region as other marine mammals, this possibility cannot be ignored. Harp's adjustment of his Port au Choix radiocarbon dates 200 years more recent in time to compensate for this ef- fect was based on changes recommended by this difference from v Robert source amount areas Reach: ate hi shell tion x in th. COaSt,‘ adjus hate ClVel 0f th Cfeas SPan attee to time C918~ - 97 from wood charcoal dates from the same houses as advised by Robert Stuckenrath (Harp 1973). 4. Variegation in the Marine Reservoir--an additional source of variation is in the presence of substantial amounts of old carbonate in coastal waters, especially in areas where carbonate rocks form the shore line. The weathering process may produce amounts of this old carbon- ate which may enter the ecosystem in the form of marine shell or other forms. It seems unlikely that such varia- tion would influence the above dates, but they are a factor in the geological dating of raised strand lines and other coastal phenomena. Table 6 is another listing of these dates and possible adjustments to them, based on the above discussion. Please note that the chronological order of the dates remains rela- tively unchanged. Perhaps the most important implication of the Suess adjustment of the dates is the ga. 33% in- crease in the time Span covered by the dates. While this span may be relatively unimportant in the present study, any attempt at analyzing the rate of cultural change through time would have to take these changes in presumed time into consideration. Table 7 is a table of values of t for the possible con- temporaneity of all of the dates from Tables 5 and 6. While the probability that Kqu 11-6 and Kqu 11-8 are coeval is .55 and thus almost an even chance, the probability that Shaymark (Kan 2) and Closure (Kqu 11) are coeval rises to 98 .80 .70 P is ”Ia probability of having ' rhis ._ _ largo or larger by chance. .60 .20__ 0 .s 1.0 t 1.5 2.0 (A3!!! Hodgnan ”suns-217) Figure 4 Probabilities for Values of "T" Com-Ye: Radio; Yea Y‘1 4590 3 4450 1 4140 i 4380 i 3314 I 3577 i 3480 t 3390 : 99 Table 5 Radiocarbon Dates from Pre-Dorset Conventional Radiocarbon Years BP 4690 4460 4140 4080 3814 3577 3480 3390 1+ l+ |+ I+ 1+ I+ |+ H- 380 100 130 O‘C‘Ch 000‘ 200 210 BP BP BP BP BP BP BP BP Lab. # GSC-1382 Gak-128l GSC-849 P-707 P-708 P-710 M-1531 GSC-1051 Site Kqu 11-8 Kle 11—6 Kan 2 Kqu 11 KeDr l Kqu 10 Kqu 13 Kqu 23 Sites Analyzed in this Thesis Material Dated Charred Charred Charred Charred Charred Charred Charred Charred fat fat fat fat fat fat fat fat cinder cinder cinder cinder cinder cinder cinder cinder U) ('0 ('3) (1’) U) (1’) ('3) (1’) P-YOI P-708 P~7lC *1-153 GSC-: R0:e 100 Table 6 South Baffin Island Pre-Dorset Radiocarbon Dates, With Adjustments Lab.# GSC-1382 Gak-1281 GSC-849 P-701 P-708 P-710 M-153l GSC-1051 Note these time spans: Years BP 4690T380 4460i100 4140T130 4080T66 3814T69 3577T69 3480T200 339of210 -1950 AD/BC 2740T380 2510T100 2190T130 2130T66 1864T69 l627t69 15301200 1440T200 Suess AD/BC 3420/ 3510 BC 3380 BC 2610/ 2750/ 2770/ 2925 BC 2520/ 2680/ 2740/ 2790/ 2830/ 2900 BC 2175/ 2300/ 2340 BC 2075 BC 1715/ 1780/ 2020 BC 1680 BC 1300 yrs. -400 Yrs. BP 4290T380 4060T100 3740T130 3680T66 3414t69 3177T69 3080T200 2990T210 1740-1830 yrs. -10% Yrs. BP 4221T380 4014T100 3726T130 3672t66 3433T69 3219T69 3132T200 3051T210 m 0,1. r L.3 7.34:3 COLLIUC21$~ 2.4.4376: «Ya-.IJLQLHHQ 3. ,J--,.J~-~u._...._uu. L: :3~....-~£:£:LL :2: 3...»: L: T...u.~.:...~. K LHCPJE oowwomqm unnnnn Hmma-z me. "a 101 ma warm Hm. nu .Hm>moo mum mmumw mohfimmm IIIIII. mnu umau huflaflnmnoum mgu aw a cannm mm. an «e. um 0H wave am. up we. up mowqawm wONlm no. "Q HH. "a HO. "Q H Hamm No.Huu mm.Hnu N¢.~uu cowowoq noe-m Ho. "a Ho. um Ho. an ac. "a HH UQUM mH.muu mw.Nuu om.muu wn.~uu omawoqaq mew-umw Ho. an ac. an ac. "a #0. "a No. "a N SQMM qo.muu om.muu Nw.muu HN.Nnu He. He ooawooeq HmNH-xmo Ho. "a Ho. "a Ho. "a Ho. an ac. "a no. "a ouaa coax oo.qnu mm.quu nm.nnu mm.muu na.muu mm.auu owmwomoq Nmma-omo Ho. "a Ho. "a Ho. "a No. "a HH. "a NH. "a -wmq-mm mnHH Unvm mm.Nuu Nw.Nuu mm.muu mN.Nnu wm.auu mm.Huu mm. nu oamwommm oowwome mowuamm moweamm mowowoq omawoqaq ooawooqq Hmoa-umo Hmma-z oaa-m woN-m “on-m mqm-omo HmNH-xmu mm wave ma wave OH cave H Mame HH case N coax o-HH case n mHAMH mo mmuma coaumoowvmm cmwBumm moamnmmmwn “smegmaawam mo mmauaflanmnoum ecu :u: we mmanma a magma 102 .67, making the odds two to one; the probability that Kqu 10 (Loon) and Kqu 13 are coeval is .64, again almost two to one odds; and the probability that Kqu 13 and Kqu 23 are coeval has risen to .75, which is three to one odds. This analysis may be interpreted as suggesting that several of the dates are statistically indistinguishable from each other, in spite of the up to 90 year difference in their measures of central tendency, which should be a caution to those prone to accept radiocarbon dates un- critically. I have not averaged this series of dates from the Closure site because I have no evidence that we are dealing with a single occupation of contemporaneous structures. Averaging would not reduce the initial variability of the dates unless we assume all of the dates to date the same event (a discrete event in time) for which there should be only one year in real time. This would be like trying to establish an average date for World War Two (for example) which would yield a false sense of increased precision and contribute another source of bias in our assessment of the dating of these occupations. Averaging of radiocarbon dates is not always appropriate and should be used only in those circumstances where a case may be made for the pre- cise contemporaneity of the samples dated. Otherwise, we are reducing variability based on a priori assumptions which may not be valid. If averaging is used, then the case for contemporaneity should be made explicit and should tempo: I Arctic with ( appro: the Se 10% my radio Paren 103 be based on data which are distinct from the radiocarbon dates themselves. In the absence of data indicative of con- temporaneous occupations, averaging is not justified. Figure 5 is a summation of the chronology for the Arctic Small Tool tradition, plotting all radiocarbon dates with one standard deviation and placing undated sites in approximate temporal positions. Figure 6 is a recasting of the same data as Figure 5, with a selective adjustment of 10% more recent in time (reduction of 10% in conventional radiocarbon years) to compensate partially for the ap- parent greater age of those samples based on sea mammals. Those dates adjusted in this manner, have had an arrowpoint added to the more recent end of their deviation plotted in the figure. In addition, dates on driftwood, which may al- so date several centuries earlier in time than samples from indigenous growth, have been labeled with a "D" and dates obtained on antler, which is also thought to yield anoma- lously early dates, have been marked by an "A". It should be obvious that the manipulation of these series of dates has almost infinite possibilities, with really very little data to validate some of the manipula- tive possibilities. I have attempted to cover some of the more relevant sources of variance, but it is apparent that one may cook these data until the consistency provides a good fit with whatever predelictions one has. We have failed to deal deliberately and appropriately with the variability inherent in the dating processes which we have '00-- 2'36 ‘2059 .1336 .12“; .145: .lise 9-033 l .1?“ .Hsg .19“ "555 104 llilfll [II I! l ‘ll |III I I: | MI |I I 11.1": I: Ililll. ON C — “I """"""""""""""""""""""""""""" M--. r--- -- .I c h M n O u u u w m m 0 '0 .0"--- '1' nu I 0 . n ma J n u M 4 m o IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII .- c O . n I o p. u - II" n u . _ M P u ' C ‘ . I o o o o o o o o o a“ u u M N S S 5 5 5 H H c. 11 I 9 o I I I 4 S 6 I 3 fl 3 1 I I 3 1. 3 1. 3 3 o o o . . o4|50 .4150 . 0330 . 0650 Figure 5 A Provisional Chronology of the Arctic Small Tool Tradition 105 Alaskan Coon All“. '00 "on . 2130 [aim-v... .n..-'---'--.1‘ .......................... . o C m u o c . 0 I .1 0 u m u N Ill :0. . o M u o .1 (is ' _ I‘ll-.. . In? 9 . r ....................... v... ......... .- C u p . ~ g o m Q. w m Q ---"-"'. - .1! ~ Q . v Q . . . 1 . J M C Q n p u l O p H 4 .w 0 Q 1 c o . 1 I - n I . 1 h r u . n 5|! O O u 11 O a i O h I N o N I O J n- -"'--III'“--.'-----".. ......... a-.- --...--" "-"""-!-.~‘ 0 o o O 0 0 o 0 0 m u S W 5 3 S 5 S S S C 5 I Q 0 0 3 8 O 3 Q 7 C ” o .1 I. 3 3 J 1. 3 3 1. 3 3 3 4 0 O O C o O O oOISO .4350 . C‘SO Arctic Small Tool Tradition Figure 6 An Adjusted Chronology of the 106 used, and such failures have contributed to the variation in archaeological interpretations. It is essential that we learn to understand the dating processes and deal with the statistical probabilities which they involve. We must make explicit the adjustments which we make in the data for dating, whether this be pre-treatment of the sample for radiocarbon dating or adjustment of the statistical descrip- tors of the dates which we receive. We can no longer ig- nore the statistical variations in our dates, and must face the statistical imprecisions as part of a more precise strategy of chronology building. I justified the 10% reduction in radiocarbon years for dates on charred fat (seal?) by noting the logical consis- tency and noting that such a reduction removes a number of anomalies in the dating (the great antiquity of the Baffin Island Pre-Dorset, when compared to presumed earlier com- plexes in the Western Arctic for example) while causing no apparent anomalies of its own. While this reduction of 10% may not be a precise estimate of what such an adjustment should be, we would be fooling ourselves to believe that we can be more precise, given the present data. I have not used (and will not use) the Suess Tree Ring Calibrations as a basis for comparison nor as a framework for the depiction of data, especially in light of the growing doubt that the "kinks" of his curve will be sub- stantiated by additional data and that these kinks are ap- plicable for all samples. Wendland and Donley have prepared 107 a smoothed curve of the relationship between radiocarbon years and calendar years (1971), but this does little more than adjust the "spread" of distributions already estab- lished by the use of radiocarbon years alone, or converted to "AD/BC" by the subtraction of 1950 years. Perhaps we would benefit by responding to a plea for greater explicitness and precision in our attempts at establishing a chronological framework, which parallels a recent plea aimed at students of Early Man in the Americas: I wish to end this critical review with a friendly plea to my fellow specialists for rationality in our search for Early Man: (1) we should be more cautious in the use of dating techniques, geomorphological context, and presumed associations and correlations; (2) we should rely less upon preconceptions about technological and typological progres- sion; (3) we should broaden our sense of his- torical perspective, being wary of former mis- takes with quarry and surface sites; and (4) we should stOp arguing so confidently where the evidence is weak. Perhaps, if we stop looking so hard for Early Man, we shall someday find him (Lynch 1974:375). If we are to understand the sources of variation in Pre- Dorset behaviors, we must make our chronological reconstruc- tions as explicit and precise as possible so that variations in our understandings do not result from imprecise chrono— logies. Any attempts at modeling behavioral change are de— pendent upon a chronological framework, and the models pre- sented in Part III follow from this discussion. Chapter 3 The Changing Environmental Setting Introduction Any attempt at paleoenvironmental reconstruction is based on the postulation of a uniformitarian principle—- that the present environmental processes were operating in the past through geological time (Barry and Perry 1973:349). The validity and reliability of this postulate are supported by recent studies which have considered historical data and the extension of trajectories of change both forward and backward in time from periods in which historical data are available (Lamb 1964, 1973 etc.; Le Roy Ladurie 1971; Johnsen et a1. 1970; Dzerdzeevskii and Sergin 1972). This use of the present and recent past as explicit and implicit models for depiction of past environments is wide3pread, and forms the basis for this and other studies. The general structure of climatic and environmental changes in the North American Arctic is known, and there was apparently relatively little change after the tundra biota had been established following the withdrawal of the continental ice sheets (Bryson et a1. 1969). During the periods of occupation by the Arctic Small Tool tradition, 108 109 there was little general change in the Arctic ecosystem, and what changes there were were subtle when compared to the major successional changes in southern North America, and are thus somewhat more difficult to discover and to in- terpret. Major changes in the Arctic ecosystem are docu- mented by ecotonal changes, such as the location of the tree line, the variance in peat communities, the changing distribution of marine fauna, etc. (see, for example, Nichols 1967a, b, c; Noble 1974; Fredskild 1973; Sorenson and Knox 1974; Matthews 1967a). However, as Bryson et al. have pointed out, ". . .there must be many climatic-biotic core areas with very little change" (1970:72). Thus, we may be able to specify the nature of the general climatic changes which affect the entire Arctic eco- system as part of synchronous global changes (see Bryson et al. 1970; Dekin l972a, 1972b), with greater reliability in ecotonal areas. The negative evidence for changes in some areas is somewhat more difficult to interpret. Where limits to distributions of biota are not clearly defined and are subject to great local variation (Polunin 1948z3), we have not been able to specify the precise nature of environmental changes on these distributions, and nega- tive evidence favors the interpretation of environmental complacency. The ecological relationships among environ- mental variables, while generally known, are not precisely defined, especially with regard to time-delay factors associated with related changes (attempts have been made to 110 relate these changes and to model the delay in responses-- Bryson et al. 1970: Fig. 14; Nichols 1967a: Figs. 4,5; McGhee l972az54; Miller 1973:574-5). Barry and Perry have suggested that vegetation reSponses to climatic change may take from 10 to 100 years to respond, while major ice caps and ice sheets may take thousands of years (1973:350). Miller has demonstrated the general relationships of glacial response to climatic change, with particular atten- tion to the problem of dating these relationships. Whereas glacier activity is determined by the magnitude and duration of climatic change, the ap— parent age of a moraine from which we infer glacier activity is a function of the response of the glacier to climate and the dating technique used. Arctic glaciers are generally more sensitive to small climatic shifts than are more southerly glaciers, but have longer response times, which in turn effects the apparent age of the moraine. A climatic deterioration of greater magnitude is required to affect southerly glaciers than for Arctic glaciers, but the response of low-latitude glaciers is more rapid due to the high mass turn— over involved (Miller 1973:574-575). The timing of this delay is difficult to ascertain, be— cause the errors associated with radiocarbon dates (see above) are larger than the apparent time-delay factor for vegetation, and we cannot really eliminate this imprecision. The technique used by me in previous papers (1969, 1970, l972b) was to observe plots of dates of logically and eco- logically related environmental events, drawing a "best-fit" distinction between sets of dates on either side of the ob- served boundary. A more sophisticated and less judgemental technique was used by Bryson et al. to establish a global 111 sequence of climatic changes, in which sample deviations were minimized and regionally significant changes were sup- pressed in favor of more widely significant changes (1970). However, the Eastern Arctic and North Atlantic regions have been subject to greater climatic fluctuations in historic time than have other areas of the northern hemisphere (Lamb l966:58,94,l7l,201) and may be expected to demonstrate en- vironmental changes which were not synchronous with other parts of the world in light of this greater sensitivity. We would expect the environmental episodes of the Eastern Arctic to be synchronous with major changes of global ex- tent and to contain additional variations of more local significance, thus presenting a complex sequence of en-' vironmental changes. While the general character and distribution of Arctic biota are known, specific population densities, distribu- tions, and fluctuations are rather poorly known. The avail- able data vary from the rather extensive study of Vibe (1967) of the utilization of Greenlandic fauna (albeit in a managed economy) to Boas's brief description of the avail- ability of walrus and caribou (1964:53-54). Studies by the Fisheries Research Board of Canada have described the his- toric distribution of pinnipeds (Mansfield 1959, 1964) and other sea mammals, while recent studies of the Canadian Wildlife Service have considered muskoxen and caribou (Tener 1965; Kelsall 1968), although neither is complete for any period of history and a precise paleo-geography is no: 1 dica data 112 not yet possible. Correspondence with Arthur Mansfield in- dicates that we must generate the prehistoric distributional data ourselves, and cannot expect independent data from their studies. Within the limits of reliability apparent from these scattered studies, we still obtain a generally reliable pic- ture of the distribution of Arctic biota which is useful as a general model of the biotic environment to which prehis- toric Arctic human populations adapted. These data will be used below in an attempt to specify the spatial variations in relevant biotic distributions and to look at possible changes in these distributions and characteristics through time. Space--Variations on an Arctic Theme The general picture that emerges from the study of the last five thousand years of Arctic occupations is one of human populations with subsistence adaptations flexible enough to adapt to almost any regional or temporal varia- tion of the Arctic tundra and coast. Attempts to dichoto- mize adaptations into coastal and inland adaptations have proved far too simple even as heuristic devices because the totality of knowledge and adaptive experience of any of these cultures under study was adequate for the maintenance of an efficient subsistence base (see Taylor 1966). No matter how we categorize the regions of Arctic Small Tool Occupation, it is clear that these peOple had a successful 113 adaptation which allowed their movement to the fullest ex- tent of this Arctic coastal-tundra environment. In general, the Arctic Small Tool tradition was con- fined to the Eskimoan Biotic Province (Dice 1943) with sites located on the tundra (exceptions include river valleys in Alaska where sites may have intruded into the boreal forest) (see Figure 7a) and on the coasts. Coastal sites appear to have been limited in their distribution to those areas where winter sea ice was prevalent (see Figure 7b) (see Dumond 1969, etc.). The coincidence of cultural tradition, Arctic climate, and biotic province is striking and we should not lose sight of these generalities when focusing on specific variances. The following discussion will focus on what is known of the most important resources available in this region, while slighting much of the lower trophic levels on which the food web is based. In part this is necessitated by the lack of specific data on many biota from this region and by the rather obvious point that some species were more important to historic and prehistoric Eskimos than were others. These twin constraints of data availability and relevance focus our attention on land mammals (caribou, musk ox), sea mammals (seals of various Species and walrus), and fish. While most of the species in the above categories are distributed across the North Aunerican Arctic, there are significant differences in their iirequency and availability, and these will form the basis ft>r the following discussion. L 114 (Irving 197236) Figure 7 North American Arctic-- Tundra and Sea Ice 115 The researches of Vibe (1967, 1970) have demonstrated the drastic limitations that weather can force on the dis- tributions of Arctic fauna. The long term patterns of weather which we call climate also have long term limits. The circulation pattern in the Eastern Arctic is dominated by the southward flow of air from.the Arctic Ocean and High Arctic Islands (Bryson 1966), with occasional depression tracks penetrating the region from the south and.west. Southern Greenland is more affected by such storm systems, especially since it juts southward into the North Atlantic, penetrating the extensive low pressure system near Iceland. Figure 8 is the mean contours of the 500-mb surface (with elevations in km) for January and July (from Hare 1968) which demonstrates the similarity in tropospheric condi- tions over the Eastern Arctic in both summer and winter. Of particular interest is the prominent trough extending from Ellesmere Island to Foxe Basin. The surface pressure means in this region document the prevalence of low pressure systems crossing the general triangle bounded by Davis Strait, northern Labrador, and Southampton Island (see Barry and Chorley 1971:140-141) in the summer and their relative complacency in winter, then dominated by more or less permanent highs. The prevalent ‘trOpospheric trough over Foxe Basin is significant in in- JEluencing long-term meteorological continuity in this locale. Frequencies of storms, wind directions, form of 116 ‘9 JANUARY JULY (Ravel...) AD Figure 8 Mean 500-mb Contours (in Km) 117 precipitation and temperatures all have a direct impact on the environment, particularly on surface phenomena such as snow or ice cover, wave characteristics, water temperature and current direction and force (Foote and Greer-Wootten 1966; Breverton and Lee 1965; Nelson 1969:34-53; Vibe 1967). The marine environment is directly influenced by the above variations in meteorology, and there is considerable evi- dence for the variation in current strengths off West Greenland in changing the marine conditions resulting in the historic increase in the availability in cod (and the decline in seals and arctic fauna) (Vibe 1967; Jensen 1939; Dunbar 1951). Variations in the relative strength of Arctic currents have been suggested as explanations for changes in marine fauna and in ocean bottom sediments (Alverson and Wilimovsky 19662855; Vilks 1970:108-109), yet the general pattern seems to have been one of relatively little change during the period under consideration here (Matthews 1967azl92). Perhaps the greatest potential for change existed in West Greenland, where the strength of the Irminger and West Greenland currents may have been variable and have had direct influences on coastal fauna. These have been the most variable of Eastern Arctic currents (Vibe 1967; Dunbar 1951; Breverton and Lee 1965), and would 'have had their greatest impact on the environment of West Greenland. At the opposite extreme is the area of Foxe Basin, Hudson Bay and straits, where the currents have probably not been subject to major changes (Dunbar 1968:49), 118 with water flowing from the Arctic Ocean dominating a south- ward flow through Foxe Basin (see Figure 9; the area marked "CORE AREA" is between 70 and 90 degrees West Longitude and 60 and 70 degrees North Latitude, which approximates an area of Pre-Dorset and Dorset cultural continuity and efflore- scence), mixing the freshwater from the southern rivers in Hudson Bay, and flowing out into the Atlantic through Hudson straits. The major zone of mixing of Atlantic (Sub- Arctic) and Arctic waters occurs in Hudson Straits where a westward moving current along the South coast of Baffin Island curves southward into the Straits off Big Island mixing with the eastward moving waters from.Hudson Bay and Foxe Basin. The strength of this westward current may have varied in intensity, causing increased Atlantic water fur- ther West in Hudson Straits, but the coastal configuration near Big Island has meant that considerable mixing in this area was a consistent oceanographic feature, even though additional mixing may have occurred further to the West along this coast. This mixing of water produced increased amounts of marine growth at all trophic levels (see Dunbar 1968). Thus, the Foxe Basin--Hudson Straits region has probably not been subject to drastic marine changes during the time period under study here, and the marine waters near Lake Harbour have supported a somewhat richer fauna than other portions of the Hudson Straits (Dunbar 1951; Soper 1928). The coincidence of relative complacency in the meteorological and oceanographic systems over Foxe 119 '(Du‘nbor 1951:“; I3) . Figure 9 Surface Currents of the Eastern Arctic -iw\“ ' O {RaL ani: 120 Foxe Basin and Hudson Straits will be of great importance in the definition of an ecological core area in the Eastern Arctic and in the relation of such an environmental core to regional culture history (see below). In recent times, the economy of the Eastern Arctic has rested mainly on the seal hunt, and it is for this reason that the distribution of seals is best known among Arctic animals. Very little is known about the specific timing and local availability of birds, which we will have to as- sume as being of rather minimal influence on the distribu- tion of Arctic peoples. This follows from their position in the subsistence pattern as being of supplementary use when available (Usher 1970:80; Bissett 1970:102; Villiers 1970:70; Higgins 1968:172). Unfortunately, we will have to relegate fish to the same position, as most of the larger rivers in the Eastern Arctic had runs of Arctic char and most coastal lakes of any size had char in some numbers (Scott and Crossman 1973: 203). However, there is no evidence that the distribution of either these rivers or lakes with their fish resources is not coincident with the Eskimoan biotic province dis- cussed above (Scott and Crossman 1973:203). While there are a number of species of fish that may have achieved significance in isolated locations and on rare occasions when other resources may have failed, there were only a half dozen or so which could have been reliable subsistence bases for any number of pe0ple for any length 0:11: land 121 of time. These include the Arctic Char, the Lake Trout, the Arctic Cisco, the Lake Whitefish, the Arctic Grayling, and the Longnose Sucker, all of which have been taken by Arctic residents in historic times (Scott and Crossman 1973). Other resources may not have been used because of technological limitations, ignorance, forbidden behaviors (taboos), or the availability of more desirable (efficient?) alternatives. Of those mentioned above, the Arctic Char has been the mbst significant in recent years across the Canadian Arctic, while commercial fisheries of char and Whitefish have de- veloped and grayling and lake trout have been of interest to sport fishermen. The Longnose Sucker seems to have been ‘widely used as dog food (Scott and Crossman 1973:535) and only slightly as a commercial fishery. Figure 10a is the distribution of Arctic Char (Salvelinus alpinus) at present, which covers all of the northern coastal regions, except that .arctic char does not usually range far inland except in the larger rivers" (Scott and Crossman 1973:203). Most char make their way between inland lakes and the sea several times in their lifetime, but some populations are landlocked. The limitation to coastal rivers and lakes is apparently because "Char cannot leap like Atlantic salmon and depend on moving in with the tide to surmount obstacles" (Scott and Crossman 1973:204). The isostasy of Arctic land forms has meant that landlocked char papulations are quite frequent. Li in can wa: 122 The Lake Trout (Salvelinus namaycush) is a land locked "char" which is extremely intolerant of salinity, and is generally confined to lakes (see Figure 10b) (Scott and Crossman 1973:223), where it is most frequently caught through the ice or netted (see Villiers 1970:55). The Arctic Cisco (Covegonus autumnalis) is an anadro- mous species spending much of its time in marine coastal waters, but running up Arctic coastal rivers in some num- bers to spawn, where it has been caught by native peoples (Scott and Crossman 1973:245-246) particularly in the Mackenzie River. While its distribution in North America is limited (see Figure 11a), it is of potential interest along the Western Arctic coast. The Lake Whitefish (Covegonus clupeaformis) is widely distributed in the lower Arctic (Figure 11b) where it has become one of the most valuable commercial freshwater fish in Canada. It has been netted in historic times and can be caught by hook and line, but their preference for deeper water (at least in southern lakes) suggests that they may not have been readily accessible to Arctic fishermen. The Arctic Grayling (Thymallus araticus) (Figure 12a) has been used by both Indians and Eskimos in recent times, mmst frequently for dog food when trout or Whitefish were scarce (Scott and Crossman 1973:304). They are found in lakes, large rivers, and rocky streams. As surface feeders, they may have been more easily accessible to fishermen with a limited technological ability, and thus a significant alte usua nati ever whit ‘ area Car 123 alternate resource when more efficient ones were scarce. The Longnose Sucker (Catostomus catostomus) while not usually mentioned as a significant resource, was taken by native peoples in varying amounts, and it is .used everywhere as food for dogs, but even they prefer lake Whitefish" (Scott and Crossman 1973:535). In southern areas, it is more prevalent than trout. It is pertinent to note that I know of no instance where the lack of fish resources in any particular locale has been noted as an adaptive problem to either historic or prehistoric inhabitants of the Eastern Arctic (for example: Rostlund 1952; Birket-Smith 1928; Boas 1964; Soper 1928: 116; Graburn 1969:22; Balikci 1970:28; Knuth 1967:31; Graburn and Strong 1973:147). The relative insignificance of fishing to the historic and modern Eskimo economies (Usher 1970:81; Bissett 1968:96; Villiers 1970:70; Higgins 1968:172) has resulted in a dearth of specific information regarding the spatial distributions and frequencies of fish resources. Land animals of significance include the musk ox and caribou, supplemented with rabbits and small carnivores. The musk ox (Ovibos moschatus) distribution is not con- tinuous within the Eskimoan Biotic province, being re- stricted to the Arctic Islands and adjacent mainland (see Figure 13; MacPherson 1965:Fig. 11). While the data on both historic and prehistoric were never present in significant numbers on Baffin and Southampton Islands, nor were present 124 / . ’ r fl 5% fl ucnc can: 0. "U" lAKE IIOUT Distribuflond (Seen I. (wanna I973) 40. Figure 10 Distribution of Arctic Char and Lake Trout 125 AICIIC CISCO lAKE WHHEFISH Distribution Q (Scott ‘ Crossman l973) 1.0. Figure 11 Distribution of Arctic Cisco and Lake Whitefish 126 ARCI'IC OIAYUNO '~ [ONGNOSE SUCK!!! Distribution a (5:13" 8. Crossman 1973) A.D. Figure 12 Distribution of Arctic Grayling and Longnose Sucker 127 I fix .' J '3- MUSK ox A), J’- Figure 13 Distribution of Musk Ox 128 in any numbers far from the Arctic coast (these conclu- sions and the data for Figure 13 were compiled from: Bee and Hall 1956:253-54; Boas 1964:7,42; Flerow 1967:278; Irving 1972:84; Harington 1970:6; Manning and MacPherson 1958:67; Pruitt 1966:527; Soper 1928:28; Tener 1965:16; and Vibe 1967:181-192). Vibe and others have demonstrated that climatic changes and hunting have contributed to local ex- tinctions of populations within the historic period, but there is no evidence to indicate that these changes have caused any drastic alteration of the range of this species, nor that such changes could explain the absence from.Baffin Island, for example. The ease with which musk ox herds were deplentished by hunting and climatic phenomena (Vibe 1967), suggests that while they may have been of supplemental value to human subsistence patterns, they were not of sufficient reliabili- ty to support an economic pattern focused on them. Even Knuth's "Musk Ox Way" (1967) did not demonstrate such a long-term focus, as the bone remains suggest a balanced use of fish, land fauna, and shore birds (1967:30-32). The most important land mammal was the caribou (Rangifer tarandus), whose distribution filled the tundra (MacPherson 1965:Fig. 9), providing one of the most im- portant sources of food and raw materials for both prehis- toric and historic Eskimos. While the distribution of these animals on the tundra is uneven, irregular and unpre- dictable, they nevertheless provide an essential resource 129 on which any successful hunting adaptation must depend (Taylor 1966; Soper 1928:63). While the habits of these animals vary significantly from locale to locale, all pOpu- lations seem to have seasonal movements whose specifics are influenced by local environmental conditions and geography. Thus, the extensive migrations of the Central and Western Arctic mainland (Kelsall 1968; Irving 1972:85-87; Manning and MacPherson 1958:65; Manning 1960:7-10) are also re- flected in the seasonal movements on the islands of the Eastern Arctic, with the exception that the herds were not as large nor were their movements as "coordinated” (Soper 1928:63-72; Manning and MacPherson 1958:65; Vibe 1967: 174-178). Several areas of the Eastern Arctic have re- ceived specific mention as containing unusually large num- bers of caribou, especially during the summer, and these include Central District of Mackenzie (Kelsall l968:46,47, Maps 11-24), northern Keewatin (Boas 1964:54), and the western plains of Baffin Island (Boas 1964:54; Soper 1928: 63-72). In historic times, the movements of caribou were frequently erratic and undependable, leading to several well known periods of hardship among Eskimos and local ex— tinctions of caribou p0pulations (Irving 1972:85; Vibe 1967: 163-180; Manning 1960:9; Kelsall 1968:17-18). However, throughout their range, Caribou provided one essential link in human subsistence and it is doubtful that any human oc- cupation could exist well for any length of time without them. While the stochastics of their availability are not 130 known, subsistence strategies that took this uncertainty into account were apparently successful. The element that best captures the stereotype of what is "Eskimo" is perhaps its maritime economy based on the hunting of sea mammals. Even though Taylor has drawn atten- tion to the omnivorous character of Eskimo subsistence (1966:119), it is still its perspective on the sea and its fauna that is a part of almost everyone's definition of Eskimo. Among marine mammals, the ringed seal (Phgga hispida) stands out as "the one great and unfailing stand- by of the Eskimos the year round" (Soper 1928:63), being found throughout the range of historic Eskimos wherever there is fast ice for breeding (Mansfield 1964:17) (see Figure 14). Variations in the distribution of ringed seal stem.from.variegations in the complexity of coastline and in the stability of sea ice, with highly convoluted coasts with stable sea ice being the most productive of ringed seal populations (McLaren 1961; Smith 1973:50). These highly productive areas have also functioned as sources of animals spreading to other regions where hunting pressures or general lower productivity had caused a depression in population bGIOW’that which the food chain could support (Smith 1973:50). The bearded seal (Erignathus barbatus) also shares the widespread distribution of the ringed seal (see Figure 15), but is much larger and does not compete in the same niche of the wood web (Mansfield 1964:19,23). As a comparison of 131 f1 ltd 0" A” 4 . - \ ’ L at Pomona! . ’ . ‘ o . - I o co 1 .1. .7 O . . Ringed Soul , . (Mansfield ”64:21) Figure 14 Distribution of.Ringed Seal ” IIIWM. ..I.' I 132 ‘ ,1 Lillie! POVIOIOI'W 1 P o 9 ’0 (I ', . 1- a. ' ‘ O o. . ‘ ‘: Q , 3...», Islam. abundance ( ‘l at lid. __ AA Bearded Seal _ r (Monsiiold 11904:“) .40 Figure 15 Distribution of Bearded Seal 133 Figures 14 and 15 may suggest, the bearded seal is not as prevalent in the Eastern Arctic as the ringed seal, but the toughness of its pelt has made it highly desirable for special purposes such as boot soles and line. The harbour seal (Phoca vitulina) is a fish eater widely scattered throughout the Eastern Arctic (see Figure 16) occurring in numbers in areas where local conditions create small areas of open water throughout the year (Mansfield 1964:4). Again, its numbers are much less than the previously mentioned seals, but its distribution seems to approach theirs. The harp seal (Phoca groenlandica) is a summer resident of the Arctic, migrating north from'breeding areas near Newfoundland (see Figure 17). Their breeding pattern and food habits serve to differentiate their habitats and loca— tions from other seals, as they seem to swim in offshore schools, and have been most easily taken by nets (Mansfield 1964:11-13). While they may be of economic interest in a very few locations in the Arctic, their importance is over- shadowed by the year-round seals and by the fact that summers are times when other resources are also available to Arctic hunters. The general pattern of Arctic seal distribution seems to be one of widespread distribution of fairly even popu- lations with occasional dense populations in areas particu- larly attractive to the habits of the particular species. It is perhaps not inapprOpriate to compare the distribution 134 .(Momiiold 196436) AID ‘ Figure 16 Distribution of Harbour Seal 135 / I . 9., \ \ \ \ .\\ \‘\ 0. \ CORE ARE , i o 0 \ i It can and * a 11.1."... No. ‘.” “I 31:25:; loan. I . ’- Hatp Soul _I q ,’ (Monsiiold\ii964=12) Figure 17 Distribution of Harp Seal 136 of seals to that of caribou, even though the caribou demon- strate high seasonal variation in availability, because both seal and caribou are ubiquitous throughout the Arctic. One sea mammal which has been of significant economic importance to Eskimos is a notable exception to the above generally even distribution of sea mammals. The walrus (Odobenus rosmarus) is found occasionally in a rather wide range which approaches that of the ringed seal (see Figure 18), but it is highly gregarious with large populations in the relatively few areas of the Eastern Arctic where the shallow inshore areas are free of fast ice in winter (Mansfield 1964:25). When compared with the distribution of seals, walrus concentrations are restricted to the Foxe Basin--Hudson Bay area and northwest and west Greenland (Soper 1928:48-49) (Figure 18). Walrus are rare in the Central Arctic and in the northern and western Canadian Arctic archipelago and do not extend north beyond Kane Basin (Mansfield 1959:Fig. 1). Their distribution was once more extensive to the south, where eighteenth century hunting extinguished their populations (Mansfield 1964:28; 1966:89). Thus, the walrus has the most varied availabili- ty of the Arctic sea mammals which have been of major sig- nificance to historic and prehistoric Eskimo economies. In spite of the limitations of the available data caused by lack of knowledge or by uncertainty regarding the impact of historic hunting on faunal ranges and local ex- tinctions, a general pattern emerges. Musk ox and walrus 137 ri. ma, 5‘ fl“ - I g — 0 a ‘1 . - Wuhan-ml “A - India. «tried.- 1 1' WALRUS IMoMH-M 194:2?) 1L0 Figure 18 Distribution of Walrus 138 have variable distributions and are scarce or absent in large areas. Caribou, ringed seal and bearded seal are found throughout the Arctic, and provide ecological and adaptive continuity through space. The mainland migratory caribou and harp seal have significant seasonal movements which alter their regional availability, such changes being most significant in the southern portions of the Arctic areas. It is also remarkable that those areas of walrus concentration do not coincide with areas of musk ox avail- ability, thus their distributions are complementary (com- pare Figures 13 and 18). In summation, the biogeography of the Eastern Arctic consists of variations on a theme, the theme being the cold desert of the tundra -- a land of low productivity with severely disruptive environmental processes and variations which have precluded the evolution of stable climax com- munities (Johnson et a1. 1966:279; Dunbar 1968:73,74; Margalef 1968:33; Irving 1972:15). Perhaps the most striking point, first expressed by Polunin (1948z3), is that within this general continuity there is extreme varia- bility from place to place in which combinations of geologi- cal, topographical, meteorological and vegetational pro- cesses have produced micro-anomalies of environment and biota. Fortunately for our purposes, the biota of direct interest to us and to Arctic peOples have had sufficient mobility to adapt to locales or regions and are not re- stricted to microenvironments of limited distribution 139 (Flannery 1968:67). In the marine environment, ecological continuity is achieved through the influences of extensive currents, with confluences marked by an increased produc- tivity, and of the generally more stable marine environment. However, the variations on the above theme are of sig- nificance to Arctic hunters, as some areas have access to more frequent and reliable resources than do others. In general, the region surrounding Foxe Basin has access to an unusually diverse and reliable set of resources, both in the sea (walrus, seals, beluga, etc.) and on land (caribou, musk ox, etc.). The opportunity for utilization of a more complex (and hence adaptively stable) food chain in this region is unusual when compared to the remainder of the Eastern Arctic. At the other end of the complexity and re- liability spectrum, both northern and southwestern Green- land have fewer alternative resources and demonstrably less reliability in their subsistence possibilities. For our purposes, we may find an Operational level be- tween the macro—similarity and the micro-complexity, but the specification of such a level of analysis awaits the consideration of the problems with which we wish to deal. Tflme--Continuity and Change in Eastern Arctic Environment From the previous discussion of chronology, we are in- terested in the period from approximately 4150 radiocarbon years to 2750 radiocarbon years (see above and Figure 6) which is approximately the period of Pre-Dorset occupation 140 of the Eastern Arctic. From my previous work on these matters (Dekin 1968, 1969, 1970, l972a, 1972b) it is ap— parent that the Eastern Arctic has evidence of somewhat greater climatic variability than most parts of the world and that this period under discussion here was not cli- matically nor environmentally complacent. Unfortunately, there are a number of problems with the data and with their interpretation that cannot be readily resolved either by me for present purposes or by those more intimately involved with the study of biotic evidence for climatic change. One of the major problems that continues to resist solution is that of time. The data for chronolo- gy and timing of climatic events are not subject to refine- ment beyond the limitations inherent in the radiocarbon dating process already discussed above. The period of re- sponse of some floral conditions to environmental changes exceeds the standard deviations of the dates for the cli- matic change, thus the response of various elements of the environment cannot be calibrated more precisely than plus or minus a century or two, and are virtually impossible to sequence inductively at the present time. It might be pos- sible to interrelate the various variables in an Arctic ecosystem as I attempted in 1972 (l972a:Part II, Figure l) and to devise an extensive series of studies to evaluate the responses of various components of the ecosystem to a well-dated and accepted climatic change, but it seems more appropriate in the long run for us to learn to deal with 141 the built-in imprecision resulting from the dating pro- cesses (see above) than to put off consideration of precise relationships until a hOped-for chronology is fulfilled. The present series of data are inadequate for the timing of events or processes of a duration of less than at least a century, unless the impact of these processes lasted that long, or longer. An unusual ice storm in mid- winter that prevented herbivores from access to plants for a month or more might lead to their extinction from a par- ticular area, while the vegetation itself may not have suf- fered any permanent damage. Several summers in a row having weather patterns producing more severe conditions than usual may prevent the production of spores in Sphagnum or other marshy plants, but unless this pattern continues for a long time, such a short term change may be unreflec- ted in the depositional record of micro or macro flora in that locale. It is apparent that we are forced to deal not with events BEE se, but with patterns of events through time. Even though seasons and storms are discrete entities at a precise level of analysis, we can only study patterns in these data through time, thus the approach which we utilize may itself structure the data and influence the patterns which we derive from such studies. Even the terms used in such studies express the predelictions which we take to them: cycles; oscillations; stades; fluctuations; stages; periods; etc. 142 As an example, Vibe's analysis of Greenlandic climatic change suggested a three stade sequence in 1967 which he subsequently modified to a sequence of alternations between two stades (compare Vibe 1967 and 1970), but he is still focused on the alternation between two periods of stable climate interspersed with periods of relatively rapid cli- matic change in two directions, similarities of which he virtually ignores because of the thrust of his research. His model utilizing stades is not designed to deal with the process of change and he virtually ignores data which sug- gest that the times of transition also have periods of similar environmental conditions, perhaps of shorter dura- tion than his stades. Rather than dwell unduly on problems of the data and their analysis, I will describe some of the relevant data for paleoenvironmental reconstruction, recognizing that micro-environmental changes may not reflect changes of the entire system.and that there probably exist core areas of the system which were relatively complacent, even in the face of major systemic changes which drastically altered the environmental conditions of ecotones and transitional areas (see Dekin l972b:13; l972azPart II;8). It is also apparent that within any particular system of interest (global, continental, hemispheric, etc.) a cli- matic change may have produced vastly different environ- mental changes in any two locales, so that changes in the movements of air masses, for example, may have resulted in 143 increased precipitation and temperature in one locale within a region and just the Opposite in another locale, thus an apparent paradox in which synchronous changes may produce very different results within the same ecosystem (Dekin l972b:13). Bryson et al. have used these data to their ad- vantage in an attempt to determine synchronous global changes even though they recognized that their sequences re- flected only wideSpread changes while dampening those of regional or local significance (Bryson et a1. 1970). Thus, between 4150 and 2750 radiocarbon years they suggest a single major change at 2890t510 marking the transition be- tween their Sub-Boreal and Sub-Atlantic stages with tenta- tive sub-episodes of the Sub-Boreal at SE: 3970 and 3480 radiocarbon years (Bryson et a1. l970:56,63). While it is probable that their methods have categorized those periods of transition between stable stages of world-wide occurrence, 1 have elsewhere summarized evidence to suggest that the North American Arctic has been subjected to changes that did not occur elsewhere (with such frequence or magnitude or dur- ation) and that we should not deduce Arctic changes from schemes derived elsewhere but need extensive inductive categorizations of the paleoclimatic changes for which we can find evidence (Dekin l972b:11). I have previously categorized this evidence as follows: 1) geomorphological--uplift phenomena, fossil strand lines, stratigraphic sequences, eustatic changes in sea level, glacial features, frozen- ground phenomena, etc.; 144 2) botanical--faunal distributions, floral dis- tributions, biotic stratigraphic sequences, (bogs), pollen profiles, etc.; 3) geophysical--isot0pic studies of variations of isotopes of oxygen or carbon in the Greenland ice sheet; and 4) oceanographic--sedimentation, ice formation and distribution, timing and sequencing of sea-ice deve10pment (distribution and break-up), temperature and salinity varia- tions, driftwood distributions, etc. (Dekin 1972a:Part 11:2). Perhaps the most widely available data on relevant geo- morphological processes result from the study of the dyna- mics of glacial fronts, yet there is not universal agree- ment on time and sequencing of glacial events from Ellesmere Island (Lyons and Mielke 1973), Baffin Island (Andrews and Ives 1972), and Greenland (Weidick 1968; Malaurie et a1. 1972). MOreover, the relationship between glacial events and climatic changes is by no means precise (Miller 1973:Fig. 9). Malaurie has commented on this state of the art, suggesting that the relationship between glacial characteristics, movements and environmental pheno- mena requires further study. It appears that, here [Thule] as elsewhere in Greenland, taking into consideration the dimen- sions of the glacial mass, for the well known causes of glacial dynamics and inertia, ad- vances and retreats of glacial tongues have not been synchronous with general climatic phenomena (Malaurie et a1. 1972:112). However, Weidick suggests that "The subsequent period since 6,000 years B.P. reflects, in the glacier variations, a period of relative stability when the glaciers must be pre— sumed to have closely followed the variations of climate, 145 locally as well as in general trends" (Weidick 1972:197). The problem is perhaps one of the dating, when trying to reconcile dates obtained from gytta, peat, lichenometry, shells, and radioisotopes of oxygen, plus having to attempt to deal with time-delay phenomena within the data (Bryson and Wendland 1967a,b). Thus, in spite of the fact that several periods of Greenlandic readvance dated by lichenometry coincide rather well with dates from.other phenomena believed to be related (Dekin 1972a:Part III:5,7; weidick 1972:196), it may be best to avoid the problem of selective selection of dates that fit from series of dates available, by holding the data in abeyance. The general picture of glacial events is clear and does correspond with interpretations of general climatic change inferred from other evidence, but the spe- cifics are not yet subject to a clear regional synthesis. During the time period.in which we are interested, Miller reports lichenometric dates averaging 3200t600 B.P. for the ending of a Neoglacial advance of presently unknown beginning and duration from.Cumberland Peninsula in Eastern Baffin Island (Miller 1973:577). This episode may be re- lated to a climatic deterioration dated in the same locale by plant material overlain by gravel stream deposits, but the evidence is inconclusive. However, the date appears to be of the appropriate magnitude and its context suggests that the relationship to climatic change is sound (3570tl40 B.P., GSC-1507; Miller 1973:577). 146 Oceanographic data are available from the Eastern high Arctic regarding the availability of driftwood in this region which is interpreted as relating to the amounts of open water on these coasts allowing driftwood to accumulate. Blake has accumulated evidence for the greater accumulation in the Queen Elizabeth Islands between 6500 and 4500 B.P. with an abrupt decrease after 4500 B.P. (Blake 1972:77) in- dicating the onset of more severe sea ice conditions. These data are in accordance with his survey of pumice dis- tributions (including pumice from the Closure site) which was also widely distributed before 4500 B.P. (Blake 1970). Fredskild pointed out that dates on driftwood from Peary Land are mostly between 4000 and 3600 B.P., suggesting that open water may have lasted longer here than in other areas (Fredskild 1973:221). These data are not indicative of major climatic shifts during the period under consideration (4150-2750 B. P.) with the possible exception of an ending to the pre- sence of more open coasts in North Greenland at about 3600 B.P. (Fredskild 1969:580; 1973:221). The apparent trend of climate from these data is one of increased cooling with increased shore ice of longer duration and greater extent than had been the case in this region be- fore 4500 B.P.. Matthews has interpreted distributional data on "warmth indicator" species of shells from.Hudson Strait and Frobisher Bay as indicating a hydroclimatic optimum of 147 warmer water (with associated warmth indicator fauna) dating 23: 3900 and 5200 B.P. (Sugluk Bay and Deception Bay in northern Ungava) and 6400 B.P. (at Frobisher Bay) (Matthews 1967a, 1967b). Matthews recognizes the problems of adequate samples of these locales and time periods, but believes the data to suggest a considerable period of warmer marine conditions, without evidence of marine en- vironmental fluctuations during this period. Andrews has summarized available data for variations in the growth rates of marine bivalves from raised beaches in the Eastern Arctic using these rates as an indicator of the marine conditions present during growth. From 8000 to 2500 B.P., several species of marine bivalve extended their range further north along eastern Baffin Island and along the entire mainland coast, but they retreated west and south following 2500 B.P. (Andrews 1972:157). Growth rates increased from 8500 to 3500 B.P. in Hudson Bay and eastern Baffin Island, but declined after 3500 B.P. (Andrews 1972: 157). Andrews concluded that warmer marine conditions pre- vailed between 8500 and 2500 B.P. with a marine optimum SE: 3500 B.P. across much of the Eastern Arctic (note that these data do relate to ecotonal changes but also to the bulk of the marine system as a whole). Of special interest to this thesis is the timing of a shift in marine conditions from cold to warm.in the Hudson Bay--Foxe Basin area at approximately 4000 B.P. (Andrews l972:Figure 8) with additional information 148 suggesting a climatic Optimum in this region and through Hudson Straits and in eastern Baffin Island at approxi- mately 4000-3000 B.P. or 92- 3500 B.P. (Andrews 1972:172). While the onset of this warmer period is relatively clearly marked (Andrews l972:Figure 8), the deterioration is less well-defined, possibly as a result of inadequate data or poor correlations among the various regions sampled, but they range from ca. 3000 to 2000 B.P.. By this time, marine conditions (and therefore productivity) had declined throughout the Eastern Arctic. Andrews under- scored his finding that the environmental changes in the marine systems, while similar to those from terrestrial systems, had lagged behind as much as 1500 years (1972:174). It is important to note that these changes are not limited to ecotonal situations nor do they result from shifts in circulation networks or currents. Andrews's data from widely scattered areas of the Eastern Arctic suggest that the entire system.was changing and that evidence for eco- tonal fluctuations must be superposed on these more general changes in the entire marine system. The coincidence of ca. 3500-3600 B.P. as a change in the growth rates of marine bivalves and as a change in the amounts of open water in the high Arctic (see above) should be underscored as an indicator of the basic agreement on the timing of major oceanographic changes in both the high and low Arctic. The geophysical data presently available consist of 149 variations in the ratio of isotopes of oxygen and carbon, available from ice cores of the Greenland Ice Sheet and from studies of tree rings from several parts of the world. The oxygen isotope studies have the advantage that they are directly indicative of climatic phenomena and are from the Eastern Arctic. Their disadvantage is that the chronology of the variations stems from the application of a theoreti- cal model for ice flow in Greenland. Variations in carbon isotopes, on the other hand, have excellent chronological control (being from annual growth rings sequenced from the present to beyond 5-6000 years), but their relation to cli- matic change is not as clear nor as convincing. 16 18 In glacial ice, the ratio of 0 to 0 is relative to the temperature of formation of the precipitation, with higher proportions of 018 at lower temperatures. The rela- tive proportions of the sample are believed indicative of the year's precipitation, and it is thus an annual sample. Long range trends in the averages of these data are directly indicative of climatic trends and changes (Langway 1970:43- 44). While it is possible to categorize the more recent layers into seasonal indicators (Langway 1970:47-51, Fig. 8), the thickness of seasonal layers decreases with time depth, and such precision is impossible at lower layers. Cross-checks of modern variations with historic climatic and temperature records indicate the validity of the ap- proach (compare Ahlmann 1953 with Johnsen et a1. 1970). The data published by Dansgaard and his colleagues 150 (Dansgaard et a1. 1969; Dansgaard et a1. 1971) are indica- tive of a climatic "optimum" from 8000 B.P. to 4100 B.P. (It is vital that we understand that this use of B.P. is apparently not subject to the kinds of fluctuations which have caused some to express radiocarbon dates as "radio- carbon years B.P.", and thus may be directly convertible to years B.C., but is indicative of actual years -- Dansgaard et a1. 1970:338-341) with colder periods 33. 3600 B.P. and 28-2900 B.P. bracketing a warmer period just before 3000 B.P. (from Dansgaard et a1. 1971:P1ate 3,344) (For an addi- tional discussion of these data, see Dekin 1972a). I must emphasize that these data categorize annual temperature variations, and thus may not relate directly to other phe- nomena more directly linked to the variations of a single season (Miller 1973:579). The major source of variation in these temperatures is thought to be fluctuations in solar radiation (Dansgaard et a1. 1969:378; Dansgaard et a1. 1971; Dekin 1970), which may also contribute tothe variations in other isotopes, particu- larly carbon (Dansgaard et a1. 1971:46; Dansgaard et a1. 1970:343; Suess l970:Fig. 2,599). If this is the case, then we may be able to infer variations in solar radiation from variations in the C14/C12 ratio. Explanations for the observed variations of the C level in atmospheric carbon dioxide may be sought in: (1) changes in the 014-production rate due to changes in the intensity of the geo- magnetic field, (2) changes in the production rate due to the modulation of the cosmic-ray flux by solar activity, and (3) changes in the 151 geochemical radiocarbon reservoirs and in the rates of carbon transfer between them (Suess 1970:595). While the correlations between climatic change and the ratio of C14 in the atmosphere are generally accepted and rather clearly delineated for recent times, there remains the question of the mechanism relating these phenomena. The data presently available seem to suggest that the sun has a 14 simultaneous influence both on climate and on C produc— tion, rather than the sun causing climatic changes which in turn influence the C14 reservoir (Suess 1970; Denton and Karlen 1973:201) (see Dekin 1970 for a somewhat more de— tailed discussion of the details of these relationships). The general curve of Cl4 deviations corresponds approx- imately to changes explicable by changes in the earth's geo- .magnetic field (Suess l970:Fig. 1), but there is an ap- parent systematic variation in the pattern of deviations from this curve, which we can reasonably infer to have been caused by solar activity (only limited data are applicable to the details of this problem, and these are compatible with the solar activity explanation). While these data are suggestive, it is presently impossible to state with any certainty that all of the deviations from.Suess's curve re- sult from solar activity. Within these limits of certainty, it is of interest to examine the curve assuming that fluctu- ations in CI4 deviations are linked to climatic changes, to ascertain if there are correlates with our other climatic data. 152 Suess's curve of Cl4 variations (l970:Fig. 2) demon- strates decreased C14 peaking approximately 2000 B.C. (note: tree ring B.C. date!) with an increase shifting to "normal" by about 1500 B.C. and continuing this marked in- crease to 1300 B.C., when a rebound to "normal" and just beyond can be seen, followed by a sharp increase in C14 a 1: about 800 B.C.. If we can be so bold as to translate these data into solar variation and thus into climate, those times of increasing C14 proportions are times of reduced solar radiation and thus of cold or cooling climate (see Dekin 1970; Suess 1970:602). Such periods occurred about 1700-1300 B.C. and 800 B.C. which these data suggest were colder periods (in the sense of having reduced solar radia- tion). Times of decreasing or decreased C14 production were either side of 2000 B.C. (which was one of the most sustained periods of increase in the 7000 years covered by these data) and 93. 1000 B.C., and these are suggested as warmer periods (increased solar radiation and decreased C14 production). It is difficult to reconcile these dates on tree-rings with the chronology available from radiocarbon dates (see previous discussion on chronology), as we are in grave danger of making the dating process circular. However, we can look at that portion of Suess's 1970 curve reproduced as my Figure 3 and see a different portrayal of the above data. On this curve, those portions of the curve which trend vertically towards the top margin away from the 153 diagonal are periods of increased Cl4 production and possibly colder periods (observe that portion of the curve after 1700 B.C.) while those portions of the curve which trend horizontally towards the diagonal (either directly or at an acute angle of intersection) are periods of decreased 14 C production and possibly warmer periods (observe that portion of the curve between 2000 B.C. and 1700 B.C.). It is this decreased C14 production which produces tree rings (and other organic materials in a readily mixed reservoir) this year with the same, or less, prOportion of radioactive carbon as last year's tree rings. The congruence of these data from the isotOpes of oxy- gen and carbon is striking, and lends credence to the in- ferred climatic changes presented previously by me (1969, 1970, 1972a, 1972b) and by others. The botanical evidence for climatic change is perhaps the most important, complete, and directly relevant data we have available. These data are of three major kinds: spatial--tree lines and soils change; vertical--changes in the growth pattern of Sphagnum bogs; and depositional-- changes in the composition of the rain of pollen through space and time. 4 The forest-tundra ecotone has varied in location during much of post glacial time (Bryson et a1. 1965; Nichols 1967a, 1967b, 1967c; Jungerius 1969; Sorenson et a1. 1971; Gordon 1972; Noble 1971; Ritchie 1972; Dekin 1972b) and its location is apparently related directly to other climatic 154 variables (Barry 1967; Bryson 1966; Bryson and Wendland 1967b; Bryson et a1. 1970; Larsen 1971; Hansell et al. 1971). While the locations of the tree line have been affected by phenomena not directly related to climatic change (for example fires: Nichols 1967a:188-l89, l972:324,339; Noble 1971:106; Noble 1974; Fredskild 1967:45), the chronology of changes in tree line location is sufficient to relate such changes to other evidence for climatic change. There is agreement that the tree line in the Central Arctic reached a northern maximum.sometime before 4000 B.P. and suffered a well-marked decline about 3600-3500 B.P. (Nichols 1967a:186-187; Sorenson et a1. 1971:471; Sorenson and Knox 1974; Nichols 1970:52; Ritchie and Hare 1971:337) after which the ecotone has not advanced to its previous limits in this area. Apparently, the minimum was reached 93. 3000 B.P. after which there was a slight readvance of the tree line North, with possible fluctuations, lasting until another major southward movement 22- 2600-2400 B.P. (Nichols 1970:54; Sorenson and Knox l974:Fig. 5; Sorenson et a1. 1971:471). The shift southwards at EE- 3500 B.P. is apparently indicative of a major shift in weather patterns and climate influencing the general circulations of the at- mosphere in the Eastern Arctic at that time (Bryson 1966; Bryson et a1. 1970:59; Nichols 1967a:185). Thus, it is not surprising to find that the evidence from lake and bog deposits in Arctic Canada confirms these general relations and the timing of these changes. While 155 there may be microenvironmental factors of local signifi- cance which do not accurately coincide with weather and cli- matic changes, in general the patterns of growth changes in peat bogs coincide with changes in climate. Peat growth at Sugluk on the South Coast of the Hudson Straits is indica- tive of a warmer period 93. 4000 B.P. and 2800 B.P. (Bartley and Matthews 1969:45) while Fredskild's data from West Greenland is indicative of marked Sphagnum growth changes Eé- 800 B.C. (ca. 2700 B.P. ) (Fredskild 1967:39, Plate 6b) with changes in charcoal and exotic pollen depo- sition 22° 2350 B.P.. At Ennadai Lake, Nichols reported marked increases in Sphagnum growth after 3650 B.P. (Nichols 1967a:187) with a corresponding change at Lynn Lake. Changes in bogs and Sphagnum.growth are frequent throughout the sub-Arctic regions in both North America and Europe, but the more common data from the Arctic are changes in the pollen de- posited in lacustrine muds, bogs, archaeological sites, and raised beaches. Fredskild has summarized the response to the question of when a change in the pollen components in- dicates a change in climate (1972:277), pointing out that in closed plant communities, deteriorations are easier to infer than are ameliorations, in part because of the differ- ences in response-times, and because climatic change cannot directly cause plants to migrate but it can cause their demise. Perhaps the best and most complete depiction of the 156 data from pollen analyses in Arctic Canada is found in the work of Nichols (especially 1967a and 1972), which has be- come a standard of reference in Arctic paleoclimatology. Pelly Lake profiles suggest a southerly movement of the Arctic front and decreasing arboreal pollen of pine and spruce EE- 3360 B.P. (Nichols 1970:49; 1972:316; 1967a:188). While the £33 3500 B.P. destruction of the advanced tree line at Ennadai Lake is well known and dated by macro- fossils (3430T110, 3550T120, 3650t100, and 3450T110-- Nichols 1967a:188), the pollen diagrams from Ennadai and Lynn Lakes suggest a period of environmental fluctuations beginning after 23- 3650 B.P. (Nichols 1967a:19l) and lasting until 23- 2670 B.P., when there was a marked cli— matic deterioration evidenced by a retreat of spruce forest, slow peat growth with oxidation, and spreading tundra around Ennadai Lake (Nichols 1967a:19l). Nichols points out the concurrence of these changes with other changes in Greenland and Europe. The general picture of climatic change which has emerged from.these studies is as follows. Using temperatures from present locations 200km south of Ennadai today to represent the temperatures at Ennadai when the treeline was 200km _ north of Ennadai, Nichols reconstructed the following se- quence relevant to our present investigations. The esti- mated changes in mean July temperatures from 4150 to 2750 B.P. were: 4150-3650 B.P. = +6 degrees F; 23: 3650 B.P. = drOp of 4 1/2 degrees F to +1 1/2 degrees F; 3650 to 157 SE- 3000 B.P. = fluctuations between +1 1/2 to +2 l/2 de- grees F; 22° 3000 B.P. to EE- 2500 B.P. = more stable tem- peratures at +2 1/2 degrees F above present; and ca. 2500 B.P. a marked abrupt drOp in temperatures to below the recent standard used in the research (Nichols 1967a:187, Figure 5). To summarize, there was a period of stable warmer climate followed by a marked decline and a period of temperature variation followed by another shorter stable period of moderate climate and another marked decline. The timing of these declines at EE- 3650 and 2500 B.P. is note- worthy. The general acceptance of these interpretations is sug- gested by their subsequent utility (Terasmae 1973:Figure 9; Miller 1973:Figure 11), and by their congruity with other data. The convergence of data on the nature and timing of environmental or climatic changes in the North American Arctic between 4150 and 2750 B.P. is remarkable, and should give us confidence in our reconstruction of paleoenviron- ments. These data continue to substantiate the nature and timing of the paleoclimatic sequence which I have presented in previous papers. Summary For our purposes, the period before and shortly after 4000 B.P. was characterized by generally warmer environ- mental conditions in both the marine land biospheres than was the case before or after that time. While there may 158 have been periods prior to that time when the land was some- what warmer, this time marked the coincidence of unusually optimal conditions in both the marine and terrestrial en- vironments. Both of these systems apparently underwent sig- nificant changes synchronously at SE: 3500-3600 B.P.. This contemporaneity (at least in so far as we can measure it) is suggested by a large number of dates ranging from SE- 3480 to 3650 B.P., each with a standard deviation of about a century. Even the date which Ritchie and Hare use to establish a southward movement of the tree line (which we might expect would be contemporaneous with other similar changes in the Central Arctic) is compatible with this dating, even though they use it to establish a date of 4000 B.P. (3630t140 B.P./GSC41338--Ritchie and Hare 1971:Figure 2). When we consider the diversity of materials and events dated, the clustering of dates is remarkable, and we are probably correct in recognizing this general time as one of significant global climatic change (Bryson et a1. 1970: Table 2,56). There was more Open water in the Eastern Arctic prior to this time (3a. 3600 B.P.), glaciers were generally re- ceding, summers were generally warmer (for a longer period), atmospheric circulation was zonal, some marine growth was more vigorous, less carbon fourteen was being produced as time went on, trees were advancing North, etc.. These environmental processes were apparently abruptly reversed at 3600 B.P.. There was a period of fluctuating but 159 Table 8 Summary of Eastern Arctic Environmental Changes, 4000-2500 B.P. COLDER C14 Increase Tree Line South 2750-2500 B.P. (93. 800 B.C.) ........................... Marine Ice Increase Marine Fauna Growth Less COOLER/FLUCTUATING Tree Line South C18 Decrease C14 Increase 2500-3600 B.P. (1650 B.C.). ......... Marine Optimum...... Open Coasts in North Greenland GENERALLY WARMER Warming Marine Conditions 4000 B.P. (93. 2050 B.C.) 160 moderate climate, and then a further decline at about 2750- 2500 B.P., again represented by a breadth and diversity of climatically significant events. The impact of these changes on Arctic flora and fauna is difficult to assess, especially since it is rare to find direct evidence of faunal change (Andrews 1972 is an excep- tion). Increased ice cover would cause a reduction in the availability of walrus in the high Arctic, as they are limited to areas of year-round Open water. However, ringed seal and bearded seal would have been minimally affected. If the historic period of weather patterns is a useful model, the increased strength of the Arctic Front would have been associated with meridional circulation, more northward storm tracks, a greater frequency of blocking high pressure cells in Greenland, with more frequent mid- winter storms with ice in northern Greenland and adjacent islands, possibly leading to decreased caribou in these regions (see Vibe 1967:169-172; Dekin 1970, l972a, l972b) but increased caribou in southwest Greenland. The decline in open water in the high Arctic may have reduced the lush- ness of the vegetation on which musk ox depend, leading to a decline in their numbers after ca. 3500 B.P.. Fish re- sources were probably largely unaffected. It is interesting to note that the evidence of major changes in Arctic environment comes from ecotonal situa- tions or from peripheral areas. These include changes in the location of the treeline in mainland Canada, in the 161 character and extent of inland glaciers, and in the amount of open water in the high Arctic. Major environmental changes did not seem to be evidenced from the central regions of the Eastern Arctic, in particular that area in the vicinity of Foxe Basin, marked as CORE AREA on most of the Figures of faunal distribution. This area was dis- cussed as a core of Arctic cultures during the Pre-Dorset and Dorset periods at a School of American Research Ad- vanced Seminar on Pre-Dorset -- Dorset Problems held in 1973. Interestingly, there has been little evidence for climatic or environmental change obtained from this region, with the exception of Matthews (1967a) and Andrews (1972). In addition, the distributiOnal data suggest that most animal species of limited distribution (or species with variations in availability) are accessible from the peri- phery of the Foxe Basin and are included in at least por- tions of the Core Area. Moreover, the weather patterns in the historic period are indicative of less variation (with generally lower rainfall and lower temperatures) in this area. Thus, the core of cultural continuity which was per- ceived by a group of investigators of culture history in this region may also be seen as an environmental core, in which there was ecological continuity through time, and in which human hunters of moderate mobility would have been assured of access to whatever resources the Eastern Arctic had to offer. In spite of this general picture of 162 continuity, climatic changes may have influenced the en- vironment and its human inhabitants, in part directly by altering the biotic conditions of even the core area (albeit perhaps only slightly) and in part indirectly by drasti- cally altering the conditions in the peripheral areas and increasing the movement of animals (including peOple) to and from these areas. Because of the importance of the re- lationship of changing environment to the behavioral re- sponses of hunters, we will elaborate on these questions later in this thesis, in an attempt to develop an explana- tory model for change in the behavior of Pre-Dorset people which relates environmental change to human behavioral change. It is essential that all_data relevant for the modeling of prehistoric environmental conditions be con- sidered and evaluated. In the past, archaeologists have sometimes chosen those data which fit with their predilec- tions and have ignored or minimized seemingly conflicting data and interpretations. Further, in our zeal to utilize data from other disciplines, we have often disregarded their cautionary or preliminary conclusions. I have at- tempted to cover all possibly relevant data presently available and to be as explicit as possible regarding my treatment of them and their shortcomings. In my opinion, the broad outlines as presented in Table 8 are generally reliable, while the specifics (of both time and space) will requireadditional work. 163 To summarize, at a general level the Eastern Arctic presents an environment for man which includes ringed seal, caribou and Arctic char in relative abundance, and which, with local exceptions, has probably done so throughout the time period under discussion (4150 to 2750 B.P.). There were variegations in these and other resources through space and time, some of which may be important variations for human adaptations. The restricted spatial distribution of walrus and musk ox and the environmental changes at 23. 3500-3600 B.P. and £3. 2700 B.P. may be of major signifi- cance to our understanding of human adaptations in this region. The historic period of hunting adaptations provides a model of general adaptation to genera and regions, rather than to a spatially restricted microenvironment (Flannery 1968:67), and thus man is like other animals in this sys- tem. It is this spatial flexibility which is worthy of mention, because it will be an important constraint in the following discussion of levels of analysis, methodology, and evaluations of explanatory models of Pre-Dorset behavioral change. This assessment of the changing environmental setting of the Pre-Dorset is important as a general background to any study, but specifically important to the development of models of Pre-Dorset behavior which attempt to relate changes in such behavior to the changing environment. These dynamics will receive further attention in Part III. Chapter IV. Chapter V. Chapter VI. Chapter VII. Chapter VIII. Chapter IX. PART III MODELING THE PRE-DORSET The Data -- Their Collection and Analysis Pre-Dorset Tent Structures and Internal Activity Areas From the Closure Site The Arctic Small Tool Horizon: A Behavioral Model of the Dispersal of Human Populations into an Unoccupied Niche ~ An Ecological Systems Model of Culture Growth, Atrophy, and Stability in the Pre-Dorset Discussion: The Conceptual Setting Recapitulation Chapter 4 The Data -- Their Collection and Analysis Artifacts, and in particular stone tools, are not pro- duced and deposited by random human behavior, but are the results of purposeful behaviors subject to a number of en- vironmental constraints. The artisan was influenced by the constraints of production and use, while the morpho- logical and Spatial characteristics of his artifacts were subjected to constraints of deposition and preservation. The archaeologist can exercise influence only with regard to data-gathering constraints (see Figure 19). Any student of human behavior attempting to utilize archaeological data in his analyses should be sensitized to the relations be- tween processes of artifact production, deposition, preser- vation, excavation, and analysis, because differences in data and analytical conclusions may stem from differences in these processes. It is desirable to study these pro- cesses in some detail before attempting to synthesize what in many cases are extremely disparate data from across the Arctic. As students of human behavior, we are initially limited by the fact that some ideas and behaviors do not 164 165 result in archaeologically available data (see Munton 1973: 686), however it would be wise to continue to evaluate this constraint as recent investigations suggest that it may be possible to investigate behaviors once thought impossible (Deetz 1965; Hill 1970; Longacre 1970). Some archaeologists have assumed that the artifacts found on a site represent activities carried on there, but while this is generally true, it may be difficult to speci- fy precisely what activities led to the deposition of par- ticular artifacts. In this sense, storage of artifacts may be defined as an "activity", but the use to which stored artifacts may be put may cause different interpretations of the "activities" carried on at that location (see Binford and Binford 1966, 1969; Isaac 1972:177). While this ap- proach may be sound and productive, there is grave danger of oversimplification. For example, the presence of bar- poons in a site does not by itself imply that the site was the base of harpooning activity, especially if the subsis- tence data do not support such an inference and if the storage and settlement pattern are compatible with a site inventory of tools curated for use in other locales and other seasons (Taylor 1967:223). However, this approach does focus on the adaptive re- lationship between technology and the behaviors occurring at a site, and is thus an important dimension of the analysis of technological systems and of the ecological relations of human behaviors and technological systems. 166 AnhoooIOQicol Dan: 0 Availablo Dace-Gathering Constraints a.“ Potential Archooologicol Data Avoiiablo Prourvucion Conflroinn ' . °°°°°°°° Products of Human ‘ : Mfi, |‘ ”r K. x k ' / Behaviov /p ,1 AD Figure 19 Constraints on Archaeological Data for the Study of Human Behavior 167 Students of artifact morphology have approached arti— facts as if artifact patterns represent some sort of mental set (mental template) of the peOple who made them (Deetz 1967, 1968). This approach has assumed the existence of a normative system.for artifact morphology, treating the data as if processes of artifact production are dominated by ad- herence to such a set of norms. Artifact form is seen largely as a result of processes which focus on the "repli- cation of uniformity" (Wallace 1961:26) and production is seen as a series of choices by the artisan (Deetz 1967, 1968). However, as Isaac has pointed out so clearly, this aspect of traditional patterning is but one constraint on the form of a particular artifact (Isaac 1972:177). Per- haps this approach is best suited to those circumstances in which other sources of variance (physical, geographical, economic, functional, etc. -- see Isaac l972:Fig. 4.1; Clarke 1968:Fig. 17) may be controlled (Cole and Kleindienst 1974:353-354). Recent efforts to study technological systems as sys- tems have suggested several useful categories of tool use which may influence their production and deposition. Tools designed and produced to be used for a specific task are apparently treated differently from those tools designed for several possible uses. While technological evolution has been defined in part by the increase in single-use tools, those instances in which ethnographic data have been analyzed from this point of view are extremely rare. 168 Maxwell has suggested that several artifact forms from the Eastern Arctic were functionally specific tools designed and used for a specific task (Maxwell l973a:321,344 for ex- ample) and he has gone on to suggest that the Pre-Dorset and Dorset technological systems were characterized by a high degree of functional specificity in virtually all of their lithic tools (Maxwell 1973a:345). By attributing a majority of variation to artifact function, Maxwell has em- phasized the similarities in artifact morphology across large areas of the Eastern Arctic and the extreme conserva- tism of the technological system (and other behavioral sys- tems) through the Pre-Dorset--Dorset continuum. Additional studies of tool use among contemporary peOples are needed in order to test the relevance of this categorization to technological systems in other areas and in other times (see Arutiunov and Sergeev 1973:4,5). The recent studies of Binford of contemporary Alaskan hunters have suggested the utility of a distinction between curated and expedient tools. Curated tools are those pro- duced and preserved for use at appropriate times and pos- sibly other locales, while expedient tools are those "tools of the moment" which are produced for imminent use and then put aside (see Binford 1972:133; 1973:242). Obviously, there is a more direct relationship between expedient tools found on a site and the activities carried on there than there is between curated tools found on a site and the ac- tivities carried on there. Curated tools may be lost in 169 transit on sites in which they were neither produced nor used, while expedient tools, by definition, were made and used on the site. Manufactured artifacts (and possibly some types of raw materials) which are traded are also curated tools, and may be found on sites where they were deposited "in transit" and therefore may not reflect sub- sistence or other activities carried on there. Because of their rarity, sentimental value, or uniqueness, some arti- facts may be curated (almost as in a museum) even though they have no relation to the customary behaviors of the people occupying the site (see Fitzhugh 1973 for a discus- sion of a Ramah chert Pre-Dorset burin from.the Rattler Bight site in a Maritime archaic site dated 22: 1900-2600 B.C.). Binford has suggested that the variation among arti- fact assemblages .should vary inversely with the de- gree that tools were multifunctional, and/or curated in an- ticipation Of future tasks" (1972:133), thus the homogeneity which Taylor and Maxwell have demonstrated would be in ac- cordance with a high prOportion of curated tools and/or a high degree of multifunctional tools. Inasmuch as Maxwell has made a case for a high degree of functional specifici- ty, we should re-assess this possibility in conjunction with an examination of the utility of the categorization of curated and expedient tools from the Pre-Dorset. From the above considerations, it is also apparent that all tools are not of equivalent diagnostic and 170 analytic significance. Broken artifacts are not the equi- valent of unbroken artifacts, and rare artifacts are not the equivalent of prevalent ones. We have come face to face with what is perhaps the real revolution in archaeolo- gy today. This revolution is the shift from the study of artifacts to the study of the relations among artifacts. It is this shift which has made systems theory so pOpular, because it makes explicit the study of not only the enti- ties within the system (artifacts) but of the relations among these entities. If we take an intellectual histori- cal perspective, such a shift has characterized a number of diverse sciences, from.anatomy (physiology) to ethnology (functionalism) and it is this parallel in the development of disciplines which has influenced contemporary discus- sions of scientific paradigms (Kuhn 1970; Fitting 1973; Leach 1973). In the Arctic, the most underrepresented aspect of archaeological studies is of the relationships among arti- facts, in spite of the habit we have fostered of talking of technological systems. It is when we make an explicit effort to study both the artifacts themselves and the re- lationships among them.(spatial, temporal, contextual, co- incidence, etc.) that previous analytical paradigms become limited in their utility, and we are forced to re-think our basic assumptions regarding archaeological data. Processes of artifact production have been summarized above and include dimensions of the physical environment 171 (physical geography and properties of materials), the learned normative tradition (mental templates), the ecolo- gical uses to which the artifact will be put, and the rela- tions among these dimensions (economy and subsistence patterns, functions of artifacts, etc.). It is important to emphasize that while processes of production are related to processes of use (especially with a feedback mode in- volving assessments of efficiency, etc.), the spatial rela- tions of production and of use are not necessarily similar, and we must be careful not to believe that these structured spatial relations are isomorphic. Processes of repair are perhaps a third set, whose spatial relations may be ana- lyzed in relation to those of production and Of use. While processes of tool production have been analyzed by Deetz and others, and processes of tool use have been analyzed by numerous "experimental archaeologists", rela- tively little attention has been paid to diversity in the processes of artifact deposition (Binford 1972, 1973; Schiffer 1972:156; Daniels 1972:203). ‘While tools may be subject to stochastically variable losses, it is doubtful whether these are of a random.nature (of equiprobability) in either space or time. In essence, tools may be pur- posely deposited, discarded, or lost. Artifacts may be discarded because they are no longer useful, broken, or otherwise not salvageable. Many broken fragments of arti- facts from archaeological sites result from this sort of discarding. Other artifacts may be still functional, but 172 are discarded because they are not worth retaining, or moving to a different location (those of us who have moved household goods know this only too well). These are Binford's expedient tools, which are "easier" to replace than to retain. In the Pre-Dorset, certain burins, burin spalls, and microblades may be of this type. We should also be aware that portions of broken arti- facts may be treated differently, depending on the nature of their characteristics (size, shape, etc.), and some may be converted to either curated or expedient tools, thus broken parts of the same artifact may be subject to dif— ferent depositional processes--some discarded, some re- hafted, some re-worked, and some retained. Lost artifacts are essentially those which would have been retained (curated) had they not been left behind or misplaced. While some artifacts may be lost or misplaced on a day-to-day basis, it seems likely that more would be lost when left behind during a movement of the household (a change in settlement location). Even here, this distinc- tion is not clear cut, because some artifacts may be cached (stored for later use) and not recovered. Among the set of curated artifacts, some are of greater value than others and are less likely to be lost. This value may be propor- tional to the replacement cost in terms of energy (polished stone tools), rarity of the raw material (stea- tite or slate), inability of replacement (traded artifact or tabooed raw material or work process), and aesthetic or 173 sentimental value to the user or to others. Pride of work- manship may also lead to a particular artifact being curated when similar artifacts are expedient. Artifacts may also be purposely deposited, perhaps most frequently in a ritual context, as in a burial, under house posts, etc. However, the unusual nature of these con- texts and their concentration in space facilitates the dis- covery of such purposeful deposits. The operations of these different processes of deposi- tion leads to a marked disparity in the frequency of cur- ated and expedient artifacts in archaeological collections. The rarity of extremely well made and still-functional ar- tifacts from certain Pre-Dorset sites (the Closure site, for example) and the application of a guide fossil approach to interpreting the data, base chronological development and comparisons on relatively rare artifacts, and possibly on those which we might expect to adhere most closely to some sort of mental template. In historical perspective, this conjunction seems fortuitous. Earlier attempts to focus on the presence or absence of expedient artifacts (such as Giddings's use of microblades and burins; 1956: 266, Fig. 81) have met with little acceptance, because of their general lack of specificity and because they tend to lump assemblages which are quite diverse in their other characteristics (especially among curated artifacts). Among those factors affecting the frequency of lost tools are: artifact size; ease of replacement (size of 174 labor investment, etc.); frequency of use (now and in the near future); importance to other behaviors (participation in a tool kit, for example); location of use (inside a structure, outside, on the sea ice, etc.); location of storage (pouched, on sleeping platform, along tent wall, cached, in entrance feature, etc.); identity of curator (owner, user, disinterested party, etc.); and number of possible uses (functionally specific or multifunctional). While it may be impossible to control all of these possi- bilities for comparative purposes, the above list may serve to sensitize us to alternative influences on the processes of tool deposition, and thus alternative sources of varia- tion in artifact assemblages (Schiffer 1972:163). At first glance, the Closure site artifacts are domi- nated by what may be expedient artifacts with more than eighty per cent of the artifact sample composed of worked chert and quartz, unifacially retouched flakes, microblades, burins and burin spalls. Unbroken artifacts of rare types are highly unusual in the sample. Illustrated collections from other areas seem to have a much higher proportion of unusual artifacts (unusual for the Closure site) and of still functional artifacts (large numbers of complete end blades, for example). Seemingly, some sites have a larger prOportion of curated artifacts deposited than others, and closer attention to such variation seems appropriate. What kinds of processes would cause the deposition of a larger number of curated artifacts (artifacts which by 175 definition would be abandoned only in unusual circumr stances)? Sites which were abandoned abruptly or in which the Occupants were killed or died would cause their entire technological inventory to be deposited, leaving a higher proportion of curated artifacts than would otherwise be the case. An abrupt subsistence shift of a more-or-less perma- nent nature would cause the abandonment of artifacts no longer useful (or less useful) in the projected tasks, thus curated artifacts became expedient ones with such a subsis- tence shift. A change in settlement pattern interrupting what had been planned for a seasonal round, might lead to the cacheing of artifacts for a future use which never came. The unplanned abandonment of a site due to changing ice conditions or weather may also result in increased deposi- tion of curated artifacts, as might any number of shifts in circumstances (taboos, social ostracism, murder) probably incapable of precise measurement in archaeological data. I would hypothesize that sites in marginal areas (of uncer- tain resources or environment) would include a higher fre- quency of curated tools (when compared to locales of more stable environment and resources) as a result of their adaptive instability. To be specific, sites within the "Core Area" (see above) would be expected to contain a somewhat higher prOportion of expedient tools and lower frequencies of curated artifacts, when compared to those sites in fringe areas. Unretrieved caches of potentially useful artifacts would be somewhat more frequent in fringe 176 areas. The general character of assemblages from locales in fringe areas would be distinctive, in part because of the greater possibility of finding otherwise curated arti- facts. Thus, depositional processes cannot be simply cate- gorized, and are much more complicated than simply the operation of activities or of learned norms. When we in- tend to investigate human behaviors which influence the production and deposition of artifacts, we are initiating a long and complicated study only dimly related to those of yesteryear in which we ignored or assumed the answers to many of these questions of diversity in site artifact assemblages. Once artifacts are deposited, we are faced with a second series of constraints Which influence the survival of the archaeological data. We can call these preservation constraints, largely related to organic materials. In the Lake Harbour region, it is not until after the colder period gg. 700 B.C. (Dekin 1972b:21) that Maxwell's Dorset sites contain significant organic preservation, and Maxwell has pointed out that early Pre-Dorset bone preservation is found only north of 59° North Latitude (Maxwell l973a:300). Extrapolating from subsequent Dorset sites and from our knowledge of Historic Eskimo technology, this preservation constraint on our knowledge severely restricts our informa- tion to a potentially tiny portion of the technological system -- lithic artifacts, and their distributions. 177 Once artifacts are deposited on the surface of a site (or within it), they are subject to a number of soil pro- cesses which may change their interrelationships (Ascher 1968). Since I have pointed out above that this dimension of technological systems (that of the interrelations among artifacts) has come to be of increasing significance in archaeology, it is significant that in the Arctic this di- mension may be subject to a number of changes, including movement in both horizontal and vertical directions. Alter- nate freezing and thawing may cause vertical sorting by artifact size, with larger artifacts moving upwards and smaller artifacts moving downwards (Corte 1963:499). If there is any slope to the deposit, solifluction may cause soil creep and the convolution of even clearly stratified layers (Iyatayet, Cape Denbigh and Engigstciak, Firth River for examples). On the Closure site, the disintegration of the convoluted bedrock outcrOps which bound areas con— taining artifacts has led to a more or less small but con- tinual depositing of small grains from this decomposed meta- morphic bedrock onto the surface of these sites. In some areas, this has led to sterile sandy layers just under the active sod layer. Both wind and water are the agents which move these particles downslope and to leeward of their origin, but these agents do not seem to have influenced the location of artifacts previously deposited, except for the possibility that the vertical extent of the site has been increased due to this deposition, thus increasing the 178 vertical distribution of freeze-sorted particles and arti- facts. In general, the Closure site has not been subjected to procesSes which have noticeably altered the distribution of the artifacts found, as no systematic distribution by size or elevation is evident. The major change in the data since the time of their deposition is apparently the loss of organic materials. Data gathering constraints are the last set to be con- sidered here, and they are the constraints most susceptible to our analysis and control. It is thus surprising to note that relatively little formal attention has been devoted to describing, explaining, or even justifying the technical and methodological choices which archaeologists have made regarding their field work and analysis, particularly in the Arctic. Archaeological research in the Arctic differs markedly from any number of norms of North American archaeology, in particular with regard to the necessity for self-sufficiency of the party and the largeamount of time and energy which must be devoted to "housekeeping" and logistics. MOst sur- vey operations have been restricted by transportation pro- blems and by the direct impact of weather and other en— vironmental influences (see Solecki et a1. 1973:11-18; Maxwell 1962:20-21). These restrictions have made such re- search expensive in terms of both time and money, limiting the amount that could be done in the time allocated. The relative brevity of the summer season and the frequent 179 necessity to wait for the ground to thaw in order to allow excavation are also unusual constraints. Further, field workers in the Arctic have been an unusual combination of untrained local workers (Eskimos) and a small number of trained archaeologists brought up from the south, leading to some variation in the abilities of field parties to con- duct precise controlled excavation (Taylor 1968:9). Equipment problems have made adequate surveying equip- ment an increasing rarity, yet the quality of published maps has been good. A number of substitutes for field equipment have been utilized, such as the substitution of aluminum "pegs for wooden stakes or of plastic bags for paper bags. Several archaeologists have used triangulation techniques for the horizontal control of the data to facili- tate the recording techniques (Gordon 1972; Hall 1970, 1973), especially in circumstances where an extensive grid system would have been inapprOpriate. Archaeological research and coverage of regions in the entire Arctic has been influenced by the existing transpor- tation networks and by recent economic deve10pment, especial- ly in the case of salvage or reconnaissance work (oil company reconnaissance and highway and pipeline salvage; location of existing facilities--airfields, radar stations, settlements, etc.). Our knowledge is by no means the result of any random sample, and perceived boundaries between archaeological regions may in part be the product of the structure of the sample taken. Our archaeological sample 180 of the Arctic is extremely spotty, but this is partially a result of the vastness of the region to be covered and by the relatively limited number of researchers who have been working in this area. There are few areas in which we may have a sample re- presentative of the complete human utilization of a locale through prehistoric time or of the complete set of activi- ties of any particular village or society at any particular point in time. The most complete samples within the Arctic Small Tool tradition are at Igloolik, Lake Harbour, North- east Greenland and Kotzebue Sound, but even these may sample the similar utilization of a locale through time, and not be representative of the complete yearly activities of any group of people at any point in time. To my knowl- edge, no one has yet the temerity to suggest that his sample included a yearly range of activities, thus seasonal variation in behavior is yet unstudied and uncontrolled for any locale or time period. Even the attention to the pro- blem of seasonal variation is as yet nascent among sites where faunal preservation holds hope of establishing the season of occupation, so we should not be surprised to find a lack of ability to control this variable where faunal materials are absent. The recent sensitization of archaeologists to the con- straints of sampling techniques has raised questions of sample bias and sample adequacy, for which we have only un- satisfactory answers. To my knowledge, no one has extracted 181 a truly random sample from any artifactual pOpulation in the Arctic, nor have they been careful to specify the nature of possible biases or rationales for many of the samples taken. Perhaps the major accepted sampling tech— nique has been to excavate test pits (of varying sizes) in "likely-looking" locations, expanding into adjacent areas when the results warranted further excavation. The two sources of judgment here are what is a likely looking loca— tion and what results warranted further excavation. Per- haps most frequently, artifact density is the major criter- ion for the second judgment, excavating areas of highest artifact yield. This bias may be against those areas where significant activities were carried on but which used a relatively small number of stone tools (albeit significant ones, or curated ones), thus a sample in which artifact yield was the major choice criterion for areas to be exca- vated contains a significant bias against certain behaviors being represented in the archaeological sample. The judgment as to what is a likely-looking location is also a source of bias, depending on the nature of the surface features used as criteria. On Cape Tanfield, we found that there was no discernible surface feature that would allow the prediction of the density of Pre-Dorset artifacts (from none to high), thus we tested every depres- sion in bedrock containing soil and sod, to make the recon- naissance as complete as possible and to minimize sample bias. While I too was prone to excavate initially those 182 areas where test pits had the highest artifact yield, we did excavate Kqu ll-B and Kqu 23 which had extremely low artifact yield, compared to other areas. While I cannot describe our sample as random, I can state my belief that it is representative of the population of Closure site arti- facts and that it suffers from.no systematic bias that I can determine. However, some of the samples which will be used for comparison with my data have been systematically biased by data collection techniques and strategies. The Arnapik site consisted of 120 find spots all of which were surface collected only (Taylor 1968a:15). McGhee's researches on Independence I and Pre-Dorset ruins on Grinnell Peninsula were largely surface collected with very little excavation (McGhee 1973; personal communication). Knuth's excavations of houses in Northeast Greenland were biased in that a sig- nificantly higher proportion of those houses with mid- passage features were excavated than those which did not have this feature (Knuth l967:28,47) which is compounded by the fact that the data are not presented in a manner that allows the comparison of house contents from these two types of houses. Other sources of bias would be the selec- tion of the best-preserved houses for excavation, the exca- vation of only those areas with surface indications, etc. Differences in field techniques are a great source of sample bias, especially in sites where the most frequent artifact may be minute (i.e., burin spalls, the largest 183 artifact category from the Closure site at 40%). Addi- tional excavations at sites previously sampled by Maxwell's field parties on earlier reconnaissances had their sample proportion of some artifacts (burin spalls in particular) significantly increased when greater precision and the sacrifice of efficiency for completeness characterized sub- sequent excavations (Maxwell, personal communication). We were fortunate to be able to devote considerable time to the excavation of the Closure site and the freedom from the constraint of time allowed the use of more precise and time consuming excavation and recording techniques. During the excavation process, it is impractical to re- cord all potentially significant archaeologically useful data. It is also impossible to record everything and we are forced to make judgments on what data should be pre- served. We plotted all artifacts identifiable in the field and saved all probable and possible lithic artifacts. All detritus was bagged by excavation unit. All rocks greater than fist size were plotted and mapped as were those in- stances of homogeneous "rotted" granite where rocks had been. However, a search of the literature indicates that this combination, the plotting of all rocks and artifacts, has not been the rule and that data useful to compare the relations between these distribution patterns are virtually nonexistent. ‘While some might argue that hindsight is un- fair, this neglect of potentially useful data (see Chapter 5) is no longer professionally acceptable. 184 The analysis and presentation of archaeological data are also subject to bias which may influence the conclu- sions of the investigator and the interpretations of those interested in the data. One of the major sources of such differences in data is in the categorization process. It is essential that categories of data be strictly defined and any lumping of data (from excavation units, different areas of the site, different sites, etc.) be made explicit (Taylor's is an admirable example of such explicitness-- 1968a:Chapter 3). Sub-sets of artifacts from possibly relevant units of sites should be kept separate and de- scribed separately, even after preliminary analysis has been interpreted that no definable differences among the sets exists, to allow others to make their own conclusions, and to allow a reassessment in the future when we are sen- sitized to different sources of potential variation. Such sensitization is another source of bias in interpretation and presentation, because if the investigator is not aware of the potential comparative significance of an attribute or artifact, it may be neither recorded, considered, nor presented (polish on burins and burin spalls, burin spall artifacts, and the use of pumice are examples) (French 1973 105; Munton l973:686).~ The categorization of the data for publication is another source of difference in data, as the lack of stand- ardization in typologies and descriptive criteria is a problem (compare Taylor 1968a, Nash 1969 and Maxwell l973a 185 on descriptions of burins). Standardization of measure- ments of artifacts is also not yet the rule, and our confi- dence in the measurement techniques of others is based largely on faith, in the absence of replicability studies. The treatment of stone detritus is also not standardized, with some investigators including such a category within the sample size and using the resultant number (N) for per- centage computation, while others cast these data aside (Giddings 1964). While some archaeologists are "lumpers” with data pre- sented in large categories, others are "splitters" with a great many categories (See Dumond l974b). In an operation- al sense, the splitters prepare their data for more facile use by other researchers, and should be emulated in this practice. The selection of illustrative data is also prone to bias, especially when only "ideal type" artifacts are pre- sented. This "replication of uniformity" approach uses se- lected artifacts and conveys a picture of uniformity within the categories used by the archaeologist. When this ap- proach is combined with a limited number of illustrations (Knuth 1967), the data are severely limited and biased by the selector. An "organization of diversity" approach (see Wallace 1961:26-27) includes data on the variation within artifact categories, as well as more "typical" artifacts, providing data of much greater utility to other investi- gators. Perhaps the selection of an artifact sample for 186 illustration should receive as much attention as the exca- vation process itself, in order to insure the representa- tiveness of the illustrated sample. Parametric data are also prone to bias in their selec- tion and depiction, although these are more likely to be the result of "sins" of omission than of commission. A normative approach to measured data involves the depiction of certain parameters related to "measures of central ten- dency", such as means, modes, and medians without the in- clusion of data indicative of the actual dispersion of the variables (ranges, standard deviations, measures of skew- ness or kurtosis, etc.). Perhaps this is because some of these parameters require additional and Often extensive manipulation of the raw data, but the advent of sophisti- cated computers and inexpensive calculators will alleviate the laborious and imprecise aspects of such depiction. The addition of such a fundamental value as the standard devia- tion allows the testing of significance of difference be- tween two sample means to determine the probability that such a difference could have occurred by chance in two ran- dom samples drawn from.the same population. Such an esti- mate is essential before we can attribute significance to such a difference (see Anderson 1970:13). Thus, even parametric data have been subject to a selection biased in the direction of behavioral norms, reducing the presented evidence for behavioral variety. Greater precision and ex- plicitness in the depiction of variation in parametric data 187 will result in a reduction of bias in data depiction, and will allow greater precision in the comparison of data among artifact assemblages. Perhaps the biggest problem of comparison of archaeo- logical data is related to the representativeness of the sample to the population from.which it was drawn. Vir- tually everyone has talked about sample inadequacy, but few are doing anything about it. A concern for sampling tech- niques is relatively recent in archaeology, yet we have all at one time or another wondered about negative evidence and whether the absence of a particular trait from a sample was significant. In this problem, sample size frequently seems to be a major problem, particularly in the Arctic where some sites have thousands of artifacts and others several dozen. Statisticians have lead in the consideration of sampling problems (for example, Cochran 1953) and geogra- phers have followed, from whom we could learn a great deal. Berry presented a clear and succinct analysis of sampling problems and spatial dimensions in 1962 which was elabor- ated and summarized by Haggett (1966). My Figure 20 is based on data and formulas presented in Berry (1962), Haggett (1966), Cochran (1953:Chapter 4), and Tate and Clelland (1957:149-152). It depicts in graphic form probability curves for a random sample having Egg; of a certain category of items found in a given proportion of a pOpulation from which the sample was drawn. It is thus 188 5000 Probability of Somplo Proportion of 0.0 500 Random 250 Somplo Sizo (N) 100 f 50 DJ 0.2 0.3 0.4 0.5 Population Proportion AD Figure 20 Probability Curves for the Absence of Artifacts With Specific Random Sample Sizes and Population Proportions 189 useful in the evaluation of negative evidence (absence of traits), because it allows the statement of the probability that such an absence would occur by chance, if we can specify the size of the random sample (from an infinite population) and the proportion which we might expect in the pOpulation. For example, in a random sample of 50 arti- facts from a large site, there is one chance in ten (P=0.lO) that an artifact type comprising four percent (0.04) of the population would not be found in the sample (see Table 9). In a sample of 250 artifacts (which is approximately the median sample size of assemblages available for comparison in this thesis) there is one chance in ten that the sample would not include items present in the pOpulation in a pro- portion of slightly less than two percent (0.02). A more precise example may be the presence or absence of steatite (or graphite) vessel fragments, of which we found 3 at the Closure site (among 1323 artifacts).r If we can assume that the Closure site sample is representative of Pre-Dorset assemblages on the South Coast of Baffin Island (which we will assume here only for the purpose of illustration), then the pOpulation prOportion is 0.00226. With a median sample size arbitrarily set at 250 (corres— ponding to the median of samples compared in this thesis), the sample proportion should be 0.00226 T 0.003, which means that the probability of not finding such vessel frag— ments in a sample of 250 is approximately 1/6 (P=0.l66). Of six such samples of 250 each, the probability of one of 190 them not having such fragments is unity. To find at least one in a random sample, the minimum sample size is 1766 (.975 probability). This curve is also useful for determining what sample would be necessary to insure that artifacts missing from the sample are not missing due to chance. If we expect micro- blades to have a frequency of 22° 14% (as they are at the Closure site), then a random sample from a pOpulation re- quires a minimal N of 25 to insure that an absence of microblades is not due to sampling error (at the 97.5 con- fidence level). Note that it is essential at this point that each artifact in the population have an equal proba- bility of being sampled, and I doubt that we can say this for any archaeological sample. It is possible to make the following general statement on sample size. If we have an adequately random sample of a population, we can predict that nine times out of ten (P=0.90) we will find at least one item if the minimum sample sizes (N) and the population prOportion are as in Table 9. 191 Table 9 Minimum Sample Sizes to Insure Finding One Item From.POpu1ations With The Following Proportions at the 90% Probability Level Minimum Sample Size Population Proportion 1000 0.01 250 0.02 200 0.02 100 0.03 50 0.04 35 0.07 25 0.09 (After Tate and Clelland 1957) These calculations allow us to assess the probable sig- nificance of negative evidence, but we are still hampered by the fact that the pOpulations of artifacts from.Arctic sites are seldom randomly sampled, and there is no present way to assess the systematic impact of non-random sampling. We can at least make a start towards defining sampling bias, sampling error, and the evaluation of the significance of the absence of diagnostic artifacts from an archaeological assemblage. It may be well to emphasize at this point that the term.bias refers to the non-representativeness of a sample from a population. Sources of bias are many, but the pro- cesses of gathering archaeological data perhaps contribute the most and at the same time are those over which the archaeologist has the greatest control. While some may read this portion of this thesis as overly critical of pre- vious archaeological research in the Arctic, such is not my intent. It is imperative that we be aware of the biases 192 which our field techniques and methods introduce, for only then can we study them, eliminate some, control for others, and otherwise insure that our analyses are of those varia- tions caused by human behaviors of the people under study and not by archaeologists. The Relationships Between Technological Data and Human Behavior at Different Levels of Analysis Technological patterning may be observed and studied at a number of levels of analysis, just as in the geogra- phical study of spatial diffusion. We should not be surprised, therefore, if we re- quire many different models to help us clarify these difficult processes, which Operate at a variety of levels. Shifts in scale are very common in all the sciences, and as men and women become dissatisfied with very general statements giving a very gross overview of a subject, they tend to direct their attention upon smaller and smaller pieces of the problem, shifting along a continuum from.the macro- to the micro-viewpoint. The rise of many subjects with names prefixed by micro- attests to the shifts in scale that have taken place throughout science during the past half century. Stafford Beer, one of the great names in the field of Operations research, has coined the provocative phrase "cones of resolu- tion," implying that problems can be considered at many different levels and in varying degrees of detail. . . . In any subject, the pendulum usually swings back and forth between the ex- tremes, and with peOple working at all scales of inquiry, the "cones of resolution" are eventually filled with a hierarchy of models. Mbre general models, high up in the cones, are supported by a number of others at lower levels (Gould 1969:25). It has been profitable to speak of continuous variation in studies of spatial distributions of populations, recognizing that the construction of discrete and isolated population 193 clusters is an artificial process, and that such constructs are best seen as heuristic devices (see Haggett 1966:100). The study of spatial and temporal variations in human be- havior may be subject to the above characterizations of the heuristic nature of units of analysis, as may be the analy- sis of the dimension of social complexity. We are faced with the problem.that our most fundamental and useful cate- gories of human behavior and concepts regarding human pro- cesses are analytical constructs (heuristic devices) sub- ject to a great deal of variety in their use and utility to scientific analysis (compare Haggett et a1. 1965). Units of archaeological analysis are also heuristic devices, whether their dimensions be in the spatial, tem- poral, or hierarchal orders. While such attention has been devoted to the categorization of archaeological units in the Midwestern United States (McKern's Midwestern Taxonomic Method--see Griffin 1943:Appendix A), relatively little formal attention has been given to this problem in the Arctic (see Chapter 1 above). Still less attention has been given to the relationship between archaeological units of analysis and other units of anthropological analysis, in particular those dealing with the organization of social and cultural systems. The relationships between patterning in archaeological data (at various levels of analysis) and patterning in human behavior remains largely unspecified and unanalyzed in Arctic archaeology (Dekin l973a:41). 194 Mbst archaeologists begin with sites and arti- facts and attempt to explain what they have found. One may also begin with a cover-law hypothesis and see if one can test it with ex- cavation or analysis. The two approaches are complementary, however, for one can hardly ex- plain what one has found without some theory, and one cannot test hypotheses without data. Thus, no matter which situation stimulated our inquiry, we must deal in the first instance with the relations between artifacts and the behaviour we are seeking to explain. In my Opinion this is precisely the point at which archaeology iS‘weakest. Rather than_ an_ ex- plicit set of_tEeory we have a set of proce- dures: typology, numerical taxonomy, attribute analysis and the'like. And even worse, since we must work with what we have, we tend to grasp at straws, hoping that such artifacts as exist will somehow inform us on the behaviour that we question. Neither established proce- dures nor the artifacts at hand necessarily re- sult in information that is meaningful in terms of the cultural categories we wish to under- stand (Hole 1973:25). One of the clearest depictions of the relations between artifacts and the behaviors archaeologists seek to explain is contained in Deetz's introductory text in archaeology (Deetz l967zFigure 17) where he presents an iconic analogue model in which the nested hierarchies of units of human be- havior and of archaeological analysis are interrelated. I wish to state explicitly that this portion of this chapter is a postulation of isomorphism between specific units of archaeological analysis and specific behavioral systems at several levels of analysis and at several levels of abstrac- tion. It is important to note that such isomorphism does 39; extend to the contents of arbitrarily defined excavation units, with the exception that such excavation units may 195 CUUUII GROUP 'IIIII' ETHNIC OIOUP macsouuo hum... Ioglonol Voriont : IlClOlAtl-D I i ' i . i I I I , g I ACIIVIIV Coupooont A cwsrn \'. > < , . : ArtiIost I I I Clostor ssnsvaor ~.u Artifacts, Bohovior, ArtiIost and Spoco ' a Figure 21 A MOdel of the Nested Hierarchies of Units of Human Behavior and of Units of Archaeological Analysis, and of Their Interrelations 196 contain relevant units of archaeological or behavioral analysis such as artifact clusters or structural features. As I will demonstrate below, the use of arbitrary excava- tion units by itself will not improve the quality of the archaeological data nor will it allow the establishment of meaningful units of archaeological analysis, leaving the selection of arbitrary excavation units open to evaluation, judgment, and criticism, as befits any heuristic device. The structures produced by people are in somewhat of an ambiguous position in this discussion of analytical units, as they are both artifacts themselves and delimita- tions of the spatial dimension of certain activities. They are thus both artifacts and components of the spatial di- mension of other artifacts and behaviors. Mbreover, as we shall see, the remnants of the structure may not delimit the distribution of the artifacts produced by behaviors which occurred within the structure. Figure 22 contains a more explicit depiction of the postulated relations between Artifacts Units, Spatial Units, and Behavioral Units. In general, the Spatial dimension be- comes larger as the complexity of the behavioral and arti- factual units increases. This also coincides with in- creasing levels of abstraction and decreasing degrees of precision and corresponding increasing levels of uncer- tainty. Thus patterning in attributes and artifacts yields the most reliable inferences within a limited spatial di- mension for the behaviors of an individual or relatively 197 mamaaoc< mo mHo>oq pan mugs: poumoz NN ouowwm 0‘ 260.. 3.52 3039.3 £263.05 82 3.36.4 96.0 xmm...\>_.Emu\naotu .308 g in 8.3025 L fil EoEoEercSEoE €95.23 neonates. 280.. 9.20 o.c5w\:3t._.: 533. 9.9.0 9.3.5 no.2. 33.30 8...: 3.0.22.3 8...: .333 «meaning 88.5 5. 60:83.20 Sutt< 3:350 one 232:5 EoEotuombcocanoo Euro) .30.. Enta> .3203. c0368... use: «aunts m_m>_mc< .6 296.. new 8...: .8qu 198 small groups of people. Patterning at higher levels of generalization becomes subject to increasingly complex en- vironmental constraints, and therefore greater uncertainty. The fact that these hierarchies are nested, means that patterning at higher levels of analysis must be based on determined patterns among the units at lower levels, thus insuring the utilization of an inductive approach to speci- fic data at lowest levels at some point in the scientific methods of analysis (Dekin l972a; Thomas 1972:673). We can- not avoid basing our analytic procedures on the observed patterning of attributes of artifacts, moving from them up the hierarchies as depicted in Figures 21 and 22 towards more and more general levels and larger and larger areal, archaeological, and behavioral units. The analysis of the Pre-Dorset and Arctic Small Tool tradition data will pro- ceed with such a strategy. Retrospect and Prospect From the above attempts to specify sources of varia- tion in archaeological data which result from differences in field techniques and sampling, it should be apparent that little attention has been given to this problem, in spite of the realization that such problems exist. In- creased attention to making our field and analytic tech- niques explicit should allow us to control the archaeologi- cal sources of variance in order to study those variations which resulted from prehistoric human behaviors. Instead 199 of ignoring or postulating away sources of sampling error, we are able to make more precise statements of the possible extent of sampling bias, thereby increasing the precision with.which we can deal with interpretive problems. Thus, greater explicitness will lead to greater precision in dealing with the data and in partitioning the sources of variation in archaeological interpretation. The remainder of this part will consist of the develop- ment and testing of three models of Arctic Small Tool tra- dition behavior at three different levels of analysis. They are designed to demonstrate the utility of greater precision and explicitness in archaeological research at several levels of abstraction and with several combinations of data and methodology. The first is a model of behaviors within the Closure Site which is proposed for further testing on comparable data sets. The second is a model of the Arctic Small Tool horizon, which is seen as an example of the dispersal of human pOpulations into an unoccupied ecological niche. The third is a model of culture growth, atrophy and stability in the Pre-Dorset, relating ecologi— cal variables in models of trajectories of change (Clarke 1968:75-77; Flannery 1973:47). Model one is principally a contribution to methodology in which the proposed model is developed using a novel data structure, and suggested for further testing. The methodo- logy may be of wider applicability than the model itself. Mbdel two may apply in other archaeological horizons 200 where a newly formed or previously unoccupied ecological niche is occupied by the horizon. Potential horizons which might fit this model are the Thule horizon across the Arctic gs. 1000 A.D. and the early Paleo-Indian horizon into North America (Haynes 1974:381). MOdel three may be applicable in a variety of situa- tions where there have been significant environmental changes resulting in pOpulation and behavioral changes. As a similar model was useful in the Southwest (Plog 1974), the continued deve10pment of this model may be of wider utility. Chapter 5 Pre-Dorset Tent Structures and Internal Activity Areas From the Closure Site One of the research goals of the Closure Site excava- tions was the precise recording of excavation data to allow the consideration of hypotheses regarding the within- site structure of human behaviors. Inasmuch as such pre- cise data preservation and presentation have not character- ized much of Arctic archaeological research, we are faced with the inability to test hypotheses on the internal dif— ferentiation of behaviors within structures or find spots, since virtually none have been suggested. This is particu- larly true of find spots such as those from the Closure Site where obvious house outlines are not present. I will approach these data from the Closure site with several hypotheses for testing regarding the nature of the behaviors reflected by the artifacts, rocks, and their dis- tribution patterns. The following discussion is similar to one presented somewhat less formally at the School of American Research Advanced Seminar on Pre-Dorset -- Dorset Problems in 1973. Arctic Small Tool structure forms vary from roughly 201 202 square (Dumond l97l:8; Meyer l970:Table 2) to square with rounded corners (Anderson l970:9; Irving 1964; Rousseliere 1964:Fig. 3; Gosselin et a1. 1974:P1anche 3) to round (Anderson l970:9; Meyer 1970:Tab1e 2) to oval or ellipti- cal (Knuth l967:49,Plate 7; Harp l96l:Fig. 5; Rousseliére 1968:Fig. 5; Taylor 1968azFig. 5; Larsen and Meldgaard 1958:Fig. 18; Meldgaard 1962:P1ate 5; and Meyer l970:Table 2). The major dimensions of these structures vary from 93. 2.5 to 4 or 5 meters, although it is impossible to discern clear outlines in many cases (Anderson 1970; Knuth 1967; IMcGhee 1973; for example). Several archaeologists have de- scribed changes in house form as significant evidence for behavioral change, being coincidental with changes in other behaviors (Anderson l970:9; Meldgaard 1962:93) but there does not seem to be any universal trend in changes from square to round or oval to rectangular. The Seahorse Gully site contains houses with variable outlines from square to circular (Meyer l970:Table 2) although the house «depressions are not always clearly defined (Meyer l970:Fig. 26,b, Map 21). Internal features, such as central passages, flagged :floors, cleared and smoothed areas and fire places are also highly variable in both space and time. Some arti- fact scatters or central passages have no associated 'Wuouse outline" (Knuth 1967; McGhee 1973) while other airtifact and rock "clusters" are without apparent struc- tnire (Taylor 1968; Tuck 1973). A further complication is 203 unwound“; was .3093... min... MN ouswwm isl- . zflfl on: 30.: .3... . >on xtoz E a a .. .. :2; costs: 2 ‘Ivco... .02.. av . . . .. O o . .. . ”Wm... OAI‘I‘_O—.¢°b IBOU 2 no a! 204 "r. t In it; 1“- .1; ) ' s loos-rod Iroi Urban Growth. 1‘ ‘s ’ .y Iotornl Is I In . .‘ Figure 24 Kqu-ll, The Closure Site 205 that there has not been consistent data collection and pub- lication. Criteria for the selection of "peripheral rocks" from the general "background scatter" prevalent on many sites have not been made clear and we are forced to rely on the judgment of the archaeologist, being unable to evaluate his field interpretation (see Dekin l972c; Meyer l970:Map 21; Knuth 1967; Gosselin et a1. 1974:P1anche 3). Applying the behavioral paradigm as set forth in this thesis, there may be several sources of variance in the re- ported behaviors reflected by the structural remains and internal rock and artifact distributions in Pre-Dorset sites. These may be listed as follows: the environment (including season of occupation; variegation in suitable rocks--flagstones; availability of heat sources--raw materials such as wood, animal fat, etc.; and availability of subsistence sources, which influences permanence of oc- cupation or duration); behavioral norms (culture) or ideals and values; idiosyncratic choices (such free variation may approach randomness, at this level of analysis -- see Scheidegger and Langbein 1966 and McConnell and Horn 1972 for a discussion of macroscopic randomness, even though nothing is truly random at the microscale); and post- depositional processes (such as the removal of rocks for use by subsequent human populations) (see Figure 19, above). It should be apparent, that by not assuming a priori that most of the variation in these behaviors is cultural, we have increased the explicitness of our consideration of 206 possible sources of variance, and, at the same time, have allowed the formulation of hypotheses which would allow the partitioning of sources of variance. As an example, the shift in house form at the Jens Munk (Kapuivik) site at Igloolik coincides with a shift in geographical house loca- tion. All of the Pre-DorSet (called Sarqaq) houses above 22 meters in elevation are sheltered by adjacent rock out- cr0ps and are exposed to the sea in one direction (see Meldgaard 1962:P1ate 1). These are "rounded, oval houses with central fireplace" (1962:93). The Dorset houses from 22 to 6 meters in elevation are on what Meldgaard infers to have been peninsula, with expo- sure to water on two sides and general location near the water on heights of land. They are located with respect to the relief and exposure of the site much as modern Eskimo winter houses are located (William Kemp, personal communi- cation). Meldgaard describes them as "large, rectangular, dug-down houses with side benches" (1962:93). This apparent shift in settlement location, with re- spect to geographic features, may reflect a shift in the season of occupation and in the activities which occurred on Jens Munk Island. Thus, the break which Meldgaard in- ferred from his 1954 and 1957 researches may represent a change in site utilization, rather than a change in culture or population. The changes in needle form, harpoon form, and the addition of multi-barbed fish spears, caribou bone awls and knives, and sledge-shoes and snow knives may 207 reflect seasonal behaviors not previously practiced at this site. The subtraction of arrow points may be a simi- lar change. Flint flaker points are all seal penis bones in the late Pre-Dorset, but shift to walrus penis bones in Dorset (Meldgaard 1962:93). A shift in seasonal occupation may be accompanied by a shift in hunting patterns, possibly with greater reliance on walrus from this location, and a shift in the available raw materials. The sea mammal populations in the Igloolik region may also have changed in distribu- tion and abundance, resulting from changed marine condi- tions. Processes of environmental change include climatic change and the continuing uplift of the land both of which may have caused significant changes in marine conditions. There may have been ample environmental reasons for a shift in the utilization of Jens Munk Island by a late Pre-Dorset occupation, and this may be reflected in changes in site location and artifacts found at Kapuivik (compare Knuth 1968:72-73 regarding seasonal variants of Independence 11). This hypothesis is also in accordance with the changes in Meldgaard's interpretations. His 1962 inference of a Dorset migration into the Igloolik area was apparently largely based on data drawn from the Jens Munk site, as he chose to discuss it at some length in support of his inter- pretations (1962:93). As a result of his later excavations at Igloolik, Meldgaard revised his views to see behavioral continuity as the basic theme, with rapid change and 208 ig_si£u_deve10pment of Dorset from Pre-Dorset. It seems likely that, with an increased regional perspective and in- formation, the coincidence of culture change with settle- ment pattern change at Jens Munk no longer dominated his data and conclusions. Testing of such hypotheses regarding the causes of ap- parent changes in house form and location at the Jens Munk site would involve the analysis of faunal remains and in- ferred behaviors across this period of apparent change. We should be able to ascertain whether this variation re- sults from changes in the scheduling of behaviors within an essentially unchanging system of behaviors and behavioral norms, or whether there is such a cultural change. This example has been provided to demonstrate the utility of the behavioral paradigm which I have proposed. By focusing on the study of human behavior, and by refusing to assume changes as cultural (a priori), we facilitate the formulation of multiple working hypotheses and suggest avenues of future research regarding Pre-Dorset structural change through space and time. The skimpy data on the seasonality of site occupation, house contents, and within-structure artifact distribution do not allow the formulation of general hypotheses re- garding the spatial distribution of behaviors within Pre- Dorset structures. It is, however, useful to derive a model to describe the structural and artifactual evidence from the Closure site, with the hOpe that such a model 209 could be tested against similar data from other Pre-Dorset sites, when the data permit. From our general knowledge of Eskimo behaviors in the Eastern Arctic during the historic period and from previous ethnographic and archaeological research we can generate a number of working hypotheses for testing on the data from the Closure Site. Because our field techniques included the careful troweling and three dimensional recording of all rocks greater than fist size and all artifacts recorded and recognized in the field, we were able to prepare dis- tributions of the in situ locations of artifacts and rocks for several loci of the Closure Site. These techniques were used in loci 11-6 + 50, 1146 + 30, 11-7, and 11-8, re- sulting in the plotting of rocks and artifacts distribu- tions (see Figures 29, 30, 31, and 32). Field interpreta- tions were inconclusive (see Maxwell 1967) and we were not convinced that such accuracy was worth the large amounts of excavation time involved in obtaining such data. In loci ll-B and Kqu 23 and portions of locus 10 we plotted rocks only, recording artifact locations only by contents of quadrants of five-foot squares, unless an unusually signi- ficant artifact were encountered (see Figures 26, 27, and 28). Additionally, locus 11-6 + 30 has a large test pit (with unplotted artifacts) in the center of an apparent cluster of rocks and artifacts, so that the constraints of time and the needs of reconnaissance have influenced the comparability of data from.within this site (minor 210 Burin, not poli Burin, polishod Burin Spa" Burin-like Iool Nosod Sido—Srl Sido—Srropor ‘h End-Scrap" End gdo, iria End’glado, Sido r Enthod.,ooul ArtiIast Iaso Biiau Fragrnol UniIaco Fragmt Microblado Misroblada (of Adso, Chippod Adu, Ground Sid. Blade Slat. Fragmont Punk’s lamp .Fragmon- Iosk fl Rock'tarbon E Figure 25 Key to Symbols Used on Distribution Maps and Profiles 211 a: .'- 4 .‘ 1‘ Figure 26 Kqu ll-B 212 Figure 27 Kqu 11-10 213 8c 09. 0 02C f a If“' ._-.___ —__ .‘ _____ _-5 Figure 28 Kqu 23 214 Figure 29 Kqu ll—6+30 215 An: U0 1v. Figure 30 Kqu 11-7 216 . _ 2‘ .. .fixs . PR... a $4.12!. .1 .. «1.- .iawu.¢1roe.m an: an 3. Figure 31 Kqu 11-8 217 Figure 32 Kqu ll-6+50 218 KquI Figure 33 Kqu ll-6+50, Plus—Minus Profiles 219 Figure 34 Kqu ll-6+50, Left—Right Profiles 220 discrepancies in the number of plotted artifacts and the summary distribution chart presented on each distribution figure are the result of the recovery of artifacts in the laboratory for which precise field data were not recorded). In the course of the analysis of these distributional data, we considered several hypotheses regarding the general nature of the activities represented by the finds from the two loci with the best evidence for artifact clustering -- 11-6 + 50 and 11-8. These were selected because I felt that the combination of excellent distributional data and observable rock and artifact clustering would yield the most reliable inferences regarding probable behaviors. I. Hypothesized general activities reflected by the clusters 11-6 + 50 and 11-8 a. dwelling structure, complete (preferred) b. portion of a larger structure c. exterior activity area (complete) d. portion of an exterior activity area e. secondary deposit. There is no soils evidence that this could be a secondary deposit, and we can safely infer that we are dealing with generally i§_si£u artifacts and rocks. The lo- cation of the artifact and rock clusters in the centers of slight depressions in the bedrock, bounded closely by rock outcrOps suggests that these clusters are discrete enti- ties, not directly related to adjacent depressions which are a minimum.of ten feet away from the margins of these 221 clusters. There is no evidence for continuous activity and artifacts which would link together such clusters. It is thus unlikely that these are portions of structures or that they represent a portion of an external activity area. The diversity of artifacts, their concentration in a cluster, and the ring-like arrangement of rocks suggest that these are not specialized activity areas outside of a dwelling, but rather support the hypothesis that these are complete dwelling structures in which a variety of activities oc- curred. II. Hypothesized nature of these dwelling structures a. windbreak b. snow house c. stone house d. sod house e. tent (skin) (Preferred) There is no evidence that Arctic peoples have made ex- tensive use of windbreaks as dwelling structures, especial- ly given the nature of the weather and environment of Hudson Straits. A windbreak would not be sufficient pro- tection. The present environment does not provide suffi- cient snow suitable for the construction of snow houses at this location (William Kemp, personal communication) and the clustering of artifacts in the general centers of de- pressions suggests that the micro-relief of the bedrock ex- posures was an important constraint in the location of these activities, which would not have been the case if 222 there had been several feet of snow over them into which snow houses were built. Other sites in the Cape Tanfield area where sod houses were apparently constructed revealed large amounts of sod accumulation, which is not the case in these loci. The stones are not of sufficient size or fre- quency to support the hypothesis of a house built largely of stone. The evidence is in accordance with the hypothe- sis that some form of skin tent was the dwelling structure. III. Hypothesized shape of the skin tent a. rectangular b. circular c. elliptical (preferred) By inspection, the best geometric shape that would conform to the pattern of rock distribution would be an ellipse. There is no evidence for rectangular tents in the historic period, and elliptical tents (occasionally grading into more circular tents) are frequent across the Arctic (see Spencer 1959:44; Balikci 1970:26; Thomsen 1928:294). In an attempt to model such a shape, an ellipse was con- structed to enclose the majority of rocks centrally located within the clusters (Figures 35 and 36). Several trials and orientations were necessary to obtain a "best-fit". Attempts to fit other shapes to this cluster support the hypothesis that the shape of the tent was elliptical. 223 " *”— ’T— if m —-—-—-_'-'—‘—" Figure 35 Kqu 11-8, Ellipses Constructed .10 I. l“ / / — I l 4 . / ‘ O l .1 _ xi . ._ ._ 224 Figure 36 Kqu ll-6+50, Ellipses Constructed 225 IV. Hypothesized configuration of the skirt of the skin tent a. not tucked (stops at ground level) b. tucked inside (preferred) c. tucked outside There is not ethnographic evidence for tent skirts which stOp at ground level, without tucking under rocks or gravel, as this would not be a desirable design where the weather is wet and windy. The Netsilik used tents with skirts of both types (Balikci 1970:26) holding down the skirts tucked inside with rocks on the inside or holding down the outside skirts with rocks on the outside periphery (see Boas 1964:145,Fig. 117). When these tents are taken down, we might expect two characteristic patterns of rock and artifact distribution, coincident with the type of skirt construction. With the skirt tucked inside, we would expect arti- facts to be distributed within the living space, with some falling between the inside rocks into the skirt itself. When the tent is to be taken down, the inner rocks are rolled toward the center of the tent, clustering on arti- facts previously lost or discarded. When the skirts are lifted, artifacts that had lodged there are dropped off, some under the area where the skirt had lain, but most would be distributed towards the interior where they would concentrate among the previously deposited rocks and arti- facts. Additional artifacts might be strewed or dragged in 226 in the direction in which the tent was taken off its sup- porting structure. The area under the skirt would be generally devoid of rocks used in the tent construction, with the possible exception of the entry where rocks might have been rolled into the entry area and rocks used for holding entrance flaps or for seating might remain. With the skirt tucked outside, we would expect arti- facts to be distributed within the living space, with no "wall" rocks used inside the living area. When the tent is to be taken down, the outer rocks on the skirt are rolled away from the tent off the skirt (see Thomsen 1928:293-294). The area under the skirt would generally be free from arti- facts and rocks, with artifacts concentrated in the center of the distribution, a peripheral area free from artifacts and rocks, and an accumulation of rocks at the margins of the cluster. To test these alternative models, I constructed another ellipse about the first ellipse (Figures 35 and 36), so that the outer ellipse was twice the absolute area of the first ellipse, and so that both ellipses had common axes. These ellipses were constructed as heuristic devices to structure the data for the following tests of the models as proposed above. Note that the ellipses and their axes divide most of the excavated portion of the loci into eight samples of equal area, with peripheral areas containing fewer artifacts and rocks. With the skirt tucked inside, artifacts should continue 227 ouwm onsmoao one Scum mouduodhum name nonuonuoum mo Houoz < mm ouswwm 0‘ aJvol/Z ..uo..:<\z c.— .u o.» . _ .. an .E.o..o.a oa.aoo.m. .. . an a. O.“ 0.. cd c. an 2 o n. w ... . . m o« o. o... s 004 f I. as Onto-.. copy. .woOE 0.: Unix Kqu 11-8 1. II. III. IV. VI. 228 Table 10 Statistical Tests of Significance of Artifact/Rock Distributions Artifact and rock distribution, excluding Minus- Left quadrant Artifacts Rocks N Inner Ellipse 46 47 93 Outer Ellipse 26 9 35 128 2 X =6.37 df=1 p=less than gs, 0.01 Null hypothesis rejected. Rocks in outer ellipse, minus quadrants Rocks, Minus-Left Rocks, Minus-Right N 11 l 13 2 X =6.2 df=1 p=ga. 0.015 Null hypothesis rejected. Rocks in inner ellipse, left quadrants Rocks, Plus-Left Rocks, Minus-Left N 20 10 30 2 X =3.3 df=l p=ca. 0.07 Null hypothesis rEjected. Artifacts in inner ellipse, right and left halves Artifacts, Left Half Artifacts, Right Half N 40 23 63 2 X =2.29 df=l p=ga. 0.12 Null hypothesis probably rejected. Rocks in inner ellipse, right and left halves Rocks, Left Half Rocks, Right Half N 30 37 57 2 X = less than 0.16 df=1 p=ga. 0.70 Artifacts in inner ellipse, plus and minus right quadrants Artifacts, Plus-Right Artifacts, Minus-Right N 17 6 23 X2=5.26 df=l p=ca. 0.02 Null hypothesis rEjected. VII. 229 Table 10 (Cont'd) Artifacts in right outer ellipse, plus and minus quadrants Artifacts, Plus—Right Artifacts, Minus-Right N 14 5 l9 X2=4.26 df=l p=ee. 0.03 Null hypothesis rejected. Kqu ll-6 + 50 I. II. III. Artifact and rock distribution Artifacts Rocks N Inner Ellipse 164 107 271 Outer Ellipse 127 52 179 450 x2=5.13 df=l p=ca. 0.03 Null hypothesis rEjected. Artifacts in inner ellipse, right and left halves Artifacts, Left Half Artifacts, Right Half N 112 52 164 x2=21.95 df=1 p=less than 0.001 Null hypothesis rejected. Artifacts in left Hemisphere, plus and minus quadrants Left Hemisphere, Plus Left Hemisphere, Minus N Quadrant Quadrant 75 125 200 2 X =12.5 df=l p=gg. 0.001 Null hypothesis rejected. 230 into the outer ellipse from a concentration in the inner ellipse while rocks do not. A test of significance (Chisquare) demonstrated that the difference in the outward extension of rocks and artifacts was not significant in locus 11—8, when we included the entire periphery of the ellipse. By inspection, there is a cluster of rocks in the outer ellipse at Minus fifty-six Left fifteen (-56L15). This is the only portion of the cluster that could be in- terpreted as an entrance structure. If this is correct, then removing the concentration of rocks in this quadrant from the sample tested would allow consideration of the hypothesis along wall areas only, stating that along wall areas, artifacts continue into theouter ellipse while rocks decline significantly in frequency. Thus, excluding the minus-left quadrant, the observed distribution is statis- tically significant and the hypothesis that tent skirts were tucked inside is supported. Locus 11-6 + 50 demon- strates such a significant decline in rocks in the outer ellipse along the entire periphery, and also supports this hypothesis. Summarily, the lack of rocks outside the constructed ellipse and their concentration within the inner ellipses while artifacts continue to be found in the outer ellipse, support the hypothesis that the tent skirts were tucked in- side with rocks piled on this tucked skirt to provide ten- sion and support. 231 V. Hypothesized entrance location a. Locus ll-8, in the minus left quadrant of the outer ellipse (preferred) b. Locus 11-6 + 50, in the plus quadrant of the outer ellipse (preferred) c. elsewhere There is ethnographic evidence for rocks around en- trances of tents, used both for support of the tent flaps and for seating (Balikci 1970:27). Additionally, at an en- trance where there is a break in an otherwise continuous wall, the rocks there will not have to be rolled Off into the interior for removal of the tent. They may, in fact, be rolled into the entrance passage proper. This kind of a rock concentration on the periphery of the cluster is not found at any other location on the ellipse periphery, thus there is no support for alternative locations. The concen— tration of rocks in and outside the outer ellipse at these locations is significant when compared with adjacent por- tions of the periphery, thus the data support the hypothe- sis that the entrance was in these quadrants. VI. Hypothesized internal activity organization a. no internal differentiation of activity location (activities and artifacts dis- tributed homogeneously and randomly) b. internal activity areas (preferred) 1. hearth and associated work area 2. entrance work-butchering area 3. sleeping platform divided into two areas -lithics area, possibly male activities -non-lithics area, possibly female activities In locus 11-8, evidence for extensive fires in the 232 form of charcoal encrusted rocks (one of which was dated by the encrusted charcoal at gs. 4690 t 380 radiocarbon years -- GSC-l382) is found almost exclusively in the Plus-Left quadrant. In the absence of alternatives, we can safely hypothesize that this area was a cooking-heating area, probably heated with a seal-oil lamp, as we have no evi- dence of wood charcoal and there were abundant rock encrus— tations usually interpreted as resulting from the burning of overflowed seal fat. There is also a significant concentration of rocks in this Plus-Left quadrant of the interior ellipse, when come pared with the adjacent Minus-Left quadrant interior to the entrance feature, which is in accordance with the use of rocks to support lamps or other associated objects (cooking vessels, etc.). There is considerable ethnographic evidence for the utilization of a large portion of the interior space of a tent as a "sleeping platform" (which is kind of a misnomer, as many activities were conducted there in addition to sleeping). Throughout the historic period until the pre- sent day, Eskimo tents usually have roughly one-half of the interior space as "sleeping platform". A comparison of artifacts and then rocks in the two halves of the inner ellipse indicates that there is a significant decline in artifacts in the right half while rocks continue. Further, within this right half of the inner ellipse, there is a significant decline in artifacts in the Minus-Right quadrant. 233 This pattern is mirrored in the adjacent outer ellipses, as there are more artifacts in the Plus-Right area than in the Minus-Right area of the outer ellipse. These data support the hypotheses that there was a sleeping platform and that it was divided into two activity areas, one of which did not involve as many stone tools as the other. All of the end-blades found in this locus are associ- ated with the entrance feature or are adjacent to it. This is consistent with the activity of bringing in game and possibly processing it in this area. The distribution of burins, burin fragments, and burin spalls in the areas adjacent to the hearth and on the Plus half of the hypothesized sleeping platform may indicate their use by men, as these grooving and engraving tools are thought to be used in the repair and construction of equip— ment used in the hunt (harpoons, toggles, etc.). The de- cline in stone artifacts frequency in the Minus half of the sleeping platform may indicate that this was an area fre- quently used by women whose tools were of generally perish- able organic materials (antler, ivory, wood, bone, skin, etc.) and who did not use stone tools to as great a degree as the men. While the division of such behaviors may not have been based on sex, this seems a reasonable inference, in light of ample ethnographic evidence for such a division in the historic period in this area. In locus 11-6 + 50, within the hypothesized structure charcoal encrusted rocks are found only in the Right 234 quadrant, supporting the interpretation that this area was a cooking-heating area, similar to that from 11-8. A simi- lar area may have occurred outside the structure between +65 and +70. One of these "cooking rock" encrustations was dated gs. 4460 t 100 radiocarbon years (Gak-1281). There is a probability of 0.55 that this date is coeval with the date from 11-8 above (see above). The pattern of rock and artifact distribution within the elliptical model of 11-8 is also observed here in the distributional data from 11-6 + 50, with rocks fairly evenly distributed, but artifacts concentrated in the Left quadrants (left hemisphere), supporting the hypothesized location of a "sleeping platform". While the artifacts in the Left inner ellipse are evenly distributed, there is a much larger number of arti- facts in the Minus outer ellipse quadrant, which supports the above contention that significantly more lithic utiliza- tion and working occurred in the quadrant of the sleeping platform adjacent to the hearth area. These data support the suggestion that there was differential use of the sleeping platform, and that such a differential may have been based on sex. By using ellipses (a geometrically-specific construc- tion) constructed about a rock cluster, the data can be structured to test hypotheses on the nature and design of the structure itself, the division of its activity areas, and on the definition of behaviors and possible divisions 235 — 25o: Figure 38 Igdluluarssuk 236 of labor and tool use. The ellipse constructed and divided for the above analyses provide a model of Pre-Dorset tent structures at the Closure site. Fortuitously, the axes seem to coincide with internal divisions of behavioral areas within the tent. Thus, the ellipses constructed originally to test hypotheses regarding the design of the dwelling, end up providing a useful iconic model of the structure itself and of its activity areas. The degree to which this model holds for other Pre-Dorset structures is a matter for further testing, but the analytical model presented here should be considered whenever we are dealing with data of artifact and rock clusters. The analytical techniques introduced here (use of geo- metric models for structuring the data) may be applied to a variety of archaeological complexes in widely scattered geographic areas wherever rock and artifact clusters com- prise the available data (Fitting 1965; MacDonald 1968). This analysis suggests that precise excavation techniques ‘may yield data useful to the study of the spatial distribu- tion of behaviors within archaeological sites, particularly in the Pre-Dorset, and also suggests that more attention should be given to possible internal differentiation of be- haviors within Pre-Dorset structures. It seems likely that this model should be tested wherever the data are relevant, such as in North and Southwestern Greenland (Knuth 1967; Larsen and Meldgaard 1958) and elsewhere (Maxwell l973a; 237 Taylor 1968a). The major contribution which explicit and precise ex- cavation techniques combined with the use of novel data structures make is that they allow the focus on the Egls- EEQEE among artifacts and the testing of hypotheses re- garding the nature of dwelling structures. The modeling of these data as tent structures may solve an interpretational problem which had not been resolved by traditional tech- niques and methods of analysis (Maxwell 1967, 1973a:303-4, 310). Chapter 6 The Arctic Small Tool Horizon: A Behavioral Model of the Dispersal of Human Populations Into an Unoccupied Niche There can be little doubt that the majority of the area through which the Arctic Small Tool horizon spread was unoccupied by human populations. This is particularly true of the Eastern Arctic, with the exception of the southern fringes where they may have come into contact with Archaic peoples near the ecotone between the tundra and the taiga. In the Western Arctic, it is apparent that there had been earlier human populations in many areas, but these later migrants appear to have occupied a previously unoccu- pied ecological niche. Geographically, they spread along the Arctic coasts until they had reached the maximum.extent of seasonally frozen coasts, usually with adjacent tundra. While it may be that the rising sea level had flooded earlier evidence of such a coastal occupation and that this habitat had not been unoccupied as it now appears, the pre- sent data support the inference that Arctic Small Tool pOpu- lations were the first to accomplish a successful adaptation to these particular conditions in the American Arctic. For 238 239 the present, I wish to focus on this coastal aspect of Arctic Small Tool peOples, ignoring (or not modeling) the expanded distribution of sites in interior Alaska (Figure l) which also occurred at this time. This model may apply best to the eastern extension of the Arctic Small Tool tradition (the Pre-Dorset) but it may be useful to consider a wider data base during the initial formulation of the model. The methods used in the formulation of this model rest on the utility of borrowing theoretical models from similar academic disciplines, in this case, from the study of migra- tions of animal species. ‘While Service has cautioned us against the uncritical borrowing of such concepts and methods from other disciplines (and named the faddistic malady, "MOuthtalk") he suggested that the surest way to avoid problems of uncritical borrowing was to follow the sequence: "problem.now: borrow later" (Service 1969:79) (see Hagerstrand 1967 for an example from geography; Pred 1967:324). The problem.is in trying to explain variations in what appears to have been a rather rapid expansion of people across approximately four thousand miles of Arctic coast. These data have been modeled as representative of two mi- grations spreading throughout the Eastern Arctic separated by several centuries in time (McGhee 1973, 1974), or as a relatively simultaneous migration of two related but dif- ferentiated pOpulations, one into the high Arctic and one 240 into the middle and low Arctic (Maxwell 1967). Thus, the variability in artifact assemblages from early occupants of the Eastern Arctic has been interpreted differently by two essentially incompatible models which have been derived in- ductively from considerations of the archaeological data. Since there is essential agreement that we are dealing with the migration of human pOpulations into the Eastern Arctic which formed archaeological data which could be considered a horizon, it seemed appropriate to attempt the derivation of a model of migrations which would account for the diversity in artifact assemblages found within the horizon. In attempting to model this horizon, I initially considered concepts and analyses used in general systems theory to sensitize me to previously used techniques and methods for the analysis of changing systems. This was followed by a consideration of ecological models from.which the following model is deductively derived (see Redman l973b:l6; Spaulding 1973:346). Such procedures for the derivation of hypotheses and hypothetical models are widely accepted as part of the cycle of induction and deduction which charac- terize the prevailing scientific paradigm (Hill 1972:94-97; Pelto 1970:17; Kemeny 1959:86). From an ecological systems perspective, the adaptive relations between a species and its ecosystem and between a human behavioral system and its environment system may both be seen as examples of coupled systems (Clarke 1968:50-51) where the adaptive response of one system to the other 241 .is merely a particular case of jointly correlated transformations" (Clarke 1968:57). It is thus possible to hypothesize that explanatory models regarding these trans- formations in the coupled systems of species and environ- ment might also be appropriate models of the coupled sys- tems of human behavior and environment. Wilkinson based part of his analysis of human preda- tion and migration into North America on the assumption That human populations colonizing previously unoccupied areas will tend to follow the pattern observed for other colonizing species; namely, an initial spread leading to the establishment of an equilibrium.with respect to resources (1972:554). I wish to make explicit the deve10pment of a model of such colonizing. Valerius Geist developed a dispersal theory to explain the Pleistocene evolution of one group of ruminants, and, believing that such a theory might be useful and relevant to the understanding of similar evolutionary change in other species, deveIOped a more generally applicable dis- persal theory describing pioneering populations. The further the species disperse, the greater the difference between individuals of the parent population and the pOpulation at the fringe. Since selection for neoteny comes to an end when the habitat is filled to carrying capacity, the gene flow is broken by a frag- mentation of the species into isolated pOpula- tions. . ., and individuals from the pOpula- tions between parent population and the colonizing fringe will form a cline. This cline is not related to clinally varying en- vironmental factors, but to the dispersal history of the species. . . . The. . .theory thus explains the formation of certain clines, 242 and the rapid evolution of Ice Age ungulates and their giant horns, antlers, and tusks. It links rapid evolutionary change to coloni- zation and suggests a long period of evolu- tionary quiescence thereafter. Thus it challenges the concept of continuous, progres- sive evolution (1971:288-89). Geist went on to relate this theory to patterns of morpholo- gical diversity and zoogeography in other ungulates con— cluding, with particular relevance to the environments with which we are interested: In the periglacial zones, differentiation is re- lated to colonization of vacant habitat, not to ecological differentiation. Hence, when the "advanced" form meets a "primitive" one, they are not compatible since they have the same eco- logical adaptations. They cannot live sympatri- cally. Therefore, one form may replace the other, they may mix, or they may remain largely separated and joined by a stable hybrid zone. This implies that although evolutionary dif- ferentiation may proceed at a high rate in the northern zones, the differentiated forms cannot survive sympatrically. The rate of extinction is high, and hence, species' diversity in this rather unproductive land remains relatively low (1971:311-312). Applying these ideas to migrations of human populations we would expect behavioral differentiation to occur as a direct result of the dispersal of the pOpulation and that clines of such differences would occur in proportion to space and time in spite of environmental homogeneity. If there were habitat differences, such adaptive variation would be in addition to clinal differentiation. Implicit in the extension of Geist's ideas to human populations is the equivalence of the transmission of genetic information with the transmission of behavioral in- formation, since such a migration would involve clines of 243 communications and behavioral similarity. As these clines deveIOp as a result of the dispersal process, they may be subject to other characteristics of - diffusion (Gould 1969). For example, we might expect that as the habitat is filled, the expanding frontier would be subject to clinal compression, because of the inertia of the expanding pOpulation against the boundary (or barrier, see Gould 1969:12; Haggett 1966260; Hagerstrand 1967:293). At the point of origin of expansion, there may be clinal expansion, again as a result of inertia of expansion as population growth ceases. Barriers in general may truncate frontal clines in advance of the barrier and stretch such clines after permeation of the barrier, and we might expect such changes in clinal variation at significant geographic barriers, such as Smith Sound or Northeast Hudson Bay (Gould 1969:12). Explicitly, the model states: With human pOpulation dispersal into an unoccupied ecological niche, behavioral variation will occur in prOportion to the dispersal variables of space and time. Pre-Dorset artifacts would beexpected to vary clinally in proportion to their spatial distribution, if we control time by considering early sites as a horizon and if we can examine data for non-functional variability. By studying artifact variability which approaches normal dis- tributions and is thus not tightly constrained by function- al requirements, we can examine the evidence for clinal 244 variations which the model predicts. The following test conditions are a more explicit statement of the data neces- sary to evaluate the model's predictions. Test Conditions The following is a stipulation of the data structure which we would expect if the above model is appropriate. Note that this is not a test to establish if the Arctic Small Tool horizon is a migration of people, as no reason- able alternatives exist. Only the southern boundaries of the horizon have earlier human occupations (Noble 1971; Dumond 1971; Fitzhugh 1972; and Tuck 1973) while most of the horizon extended into unoccupied territory. If this model is apprOpriate, we should expect two general types of variation in prOportion to space and time. Single variables should exhibit clinal variation and there should be increased variability among functionally-equiva- lent artifacts. These should be seen in attributes usually categorized as "style" or as "free variation" among functional equivalents. Adaptive variation resulting from the exploitation of new and diverse ecological niches should be in addition to the above variations and should be habitat-specific. Two fundamental problems face this analysis: 1) there is no agreement on what constitute stylistic variations in lithic artifacts; and 2) there is no agreement on which artifacts are functional equivalents. Moreover, Maxwell 245 has suggested (on a largely intuitive basis) that the arti- facts from the Pre-Dorset and Dorset occupations of Southern Baffin Island ". . .have a high degree of functional speci- ficity" (1973a:344) in which there is a close relatiOn be- tween morphology and function, mitigating against signifi- cant free variation in artifact morphology. Nevertheless, we might expect that artifact size would be subject to a degree of clinal variation (possibly as discrete variation, due to sampling) as demonstrated in such artifacts as burins and microblades (selected in part because they are ubiquitous in these assemblages and be- cause they have most frequently received descriptive atten- tion). We would thus expect continuously varying attri- butes (burin length, burin width and microblade width) to demonstrate considerable variety through space, if we are able to control for the variables of time (done by limiting the data to the horizon of 4100-3700 B.P.) and environment (raw materials, functional difference, etc.). We might also expect that new forms of tools would ap- pear, either replacing earlier forms or in addition to them. We will consider end blades and end scrapers, as they too have received considerable attention, focusing on varia- bility in shape (the easiest to obtain from the published examples). If the above model is apprOpriate, we will find in- creased variation in burin length and width and in micro- blade width in the Eastern Arctic and from the high Arctic 246 to the low Arctic. We would also expect to find new forms of end scrapers and end blades in the Eastern Arctic, with some differences between northern and southern assemblages. Testing the MOdel The data on which the model will be evaluated are sum- marized in Tables ll, 14, and 15 which contain data pub- lished with regard to sites in the Arctic Small Tool hori- zon between 4100-3700 B.P. (see Figure 6). Decisions to include sites within this horizon for purposes of testing this model were based on the adjusted radiocarbon chronology of Figure 6, with the following exceptions. The Annawak site (KeDr l) was included on the basis of Maxwell's place- ment of the site within the Lake Harbour Pre-Dorset at a time earlier than its radiocarbon date (Maxwell l973a:308; based on the overall crudeness of the artifacts and lack of sidenotching). The Loon site and site Kqu 13 were ex“ cluded on the basis of later radiocarbon dates and the ap- parent extent of their ground slate industry. While it is possible that the Arnapik site was occupied somewhat after the arbitrary limit of this horizon (3700 B.P.), the general similarity of the Arnapik artifacts to those from Igloolik and the southern Baffin Island sites under consideration here suggests its inclusion. The Closure site itself was divided into three groups for purposes of this analysis: Kqu 23; Kqu 11-10; and Kqu ll—B, 6, 7, 8. Site 23 was considered separately 247 because of its Spatial separation from the rest of the Closure Site (see Figure 24) and because none of its burins are polished. Kqu 11-10 was considered separately because of its prominence at one end of the site, its slightly higher elevation, and extremely low frequency of polished burins (1 out Of 40). Kqu ll-B, 6, 7 and 8 were lumped together because they are all at approximately the same elevation and are contiguous, moreover, the radiocarbon dates from Kqu 11-6 and Kqu 11-8 have a probability of 0.55 that they date the same event. Their contents are similar. There is considerable variety in sample size on which the means are computed (see Chapter 5 above) and there are several collections on which no data are available (the data from the Independence I sites in North Greenland are computed from the illustrations in Knuth 1967 and Giddings 1967, although Knuth's narrative confirms their large size -- Knuth 1967:34). Perhaps the most striking aspect of these data is the extremely small size of the burins from southern Baffin Island (the Closure site) and the extremely large Size of the burins from Independence I, both when compared with those from the Western Arctic and other sites in the East. Without complete parameters for any of these series it is impossible to assess the levels of significance of the dif- ference among these samples, but the relative disparities are clear. Note that the clinal variability in these 248 samples is appropriate to the model's predictions with re- gard to mean burin length and apparently with regard to mean burin width, although somewhat less clearly. Perhaps this is a case of even greater variety in this latter di- 'mension, such that the stochastics approach randomness about a value constrained by limitations of hafting tech- nology. As Table 12 demonstrates, the dispersal around the ‘mean burin width is indistinguishable from the normal. This normalcy of dispersal is also true of the distribution about the mean burin length. This is noteworthy because to reflect "stylistic" variation, we would expect the para- meters not to be tightly constrained and to vary normally. This is particularly true of those chosen from.this series, especially when compared to a Similar variable, haft length, which is skewed in the direction of somewhat longer hafts with a highly peaked curve. This implies that hafting re- quirements demand a haft length of a certain minimum.size and that Pre-Dorset craftsmen held closely to this norm, allowing only little variation in the direction of in- creased haft length. To reiterate, these parameters of the series of burins from the entire Closure Site reinforce the suitability of burin length and burin width for the above analysis, be- cause of the apparent lack of severe manufacturing con- straints. The interpretation is less clear with regard to micro- blade width, except that there is considerable variety 249 among the samples which does not seem to reflect a dis— persal vector. However, in order to assess the variability in this horizon, we can compare the variability of the means of this Set of samples with the variability of the set of samples reported by Anderson (1970:14) which he in- terprets to represent the same technological tradition (Proto-Denbigh, Classic Denbigh Flint and Late Denbigh Flint) through time in the same area. If the above model is apprOpriate, we Should expect greater variability in the horizon than in the tradition. The mean of the sample means from the Denbigh Flint Complex (Anderson 1970:14, Figure 5) is 7.17 mm.with a standard deviation of 0.65 mm.(coefficient of variation is 9.06). This average mean is slightly smaller and has a smaller standard deviation (as a measure of dispersal about the mean) than that for the Arctic Small Tool horizon (from Table 11 samples greater than 4) which is 7.81 mm with a standard deviation of 0.93 mm (coefficient of varia- tion is 11.90). There is somewhat greater variability in the horizon, which is suggestive that the model is sup- ported by these data. To accept such a difference as proof of the model with- out testing the statistical Significance of these differ- ences would be contrary to the stated objectives of this thesis, even though such inferences have been made in the past. If we assume that these two samples were drawn ran- domly from a universe of all Arctic Small Tool assemblages, 250 we can assess the probability that they could have been drawn from.the same population (see Chapter 2; Wallis and Roberts 1956:418-419). Using the above standard deviations, their standard error is 0.316, which divided into the difference in the means (0.64 mm) yields a value for Z of 2.02 which we might expect to have occurred by chance from the same pOpulation with a probability of gs; 0.04 (Wallis and Roberts 1956: 365). A more precise test of significance used for small samples (Yeomans 1968b:104) is a t test of these means, which has a value of t=l.81 (df=23) and a probability of gs. 0.08 by chance (Hodgman 1958:216-217). In brief, these data are highly suggestive that the horizon has signifi- cantly greater variability than does the tradition (in spite of the fact that they do overlap), and that the model is appropriate. Since these samples overlap, they are not entirely in- dependent. If I removed the overlap in these samples, by removing four samples from the tradition and three from the horizon, the sample means would diverge somewhat more (Since the overlap brackets the mean of the tradition but is below the mean of the horizon) becoming 7.95 mm for the horizon and 7.15 mm for the tradition. The standard deviations would then become 1.09 mm for the horizon and 0.81 mm for the tradition (coefficients of variation are 11.32 for the tradition and 13.71 for the horizon), but the value of z drOps to 1.667 with a probability gs. 0.10 by chance. 251 Applying the more precise t test for small samples, t=1.39 (df=15), thus the differences between these sample means could have occurred by chance with a probabilityof almost 0.2 by chance (Hodgman 1958:216-217). This probability is still suggestive, even with the overlap removed. The model is supported by these data, at a Statistically Significant level. The establishment of Significant difference in the variation within the Arctic Small Tool horizon and the Arctic Small Tool Tradition (Denbigh Flint Complex), raises the question of assessing the significance of differences. In point of fact, the different trends seen by Anderson within the Denbigh Flint tradition may be insignificant variations within the same population. A comparative summary of the microblade dimensions also reveals a change from wider forms (Proto-Denbigh) to narrower ones (Early Classic Denbigh), and again an increase in width (Late Denbigh). At present this seems to be the only consistent change through time in microblade measurements at the major Denbigh sites (Fig. 5) (Anderson 1970:13). Such testing of significance should become a normal tech- nique to reduce the degree of spurious inference which comes from imprecise and implicit methods of analysis. While the above data support the model at a signifi- cant level, there may be adaptive reasons why these dif- ferences occurred. The buff chert used for the majority of artifacts from.the Closure Site may be available in smaller dimensions than in other areas (the high Arctic in particu- lar), thus influencing the Size of the artifacts made from 252 Table 11 Measurements of Selected Chert Artifacts, ASTh Brooks River Gravels Iyatayet Onion Portage, DFC Onion Portage, Proto Denbigh Itivlik Dismal 2 OhPo 5 Oqu 4 St. Mary's Hill Igloolik Arnapik Ivugivik KeDr l Kqu 11-10 Kqu 23 Kqu ll-B, 6, 7, 8 Kan2 Kqu 11 Cape Sparbo Port Refuge Ind. I Saglek, Site K Independence I Mean Burin Length 2.64 (2.2 (3.5 .93 .08 .12 .98 .91 .15 .94 .23 NI—‘NI—‘t—‘NNNN 3.95 Measurements in Cm. N 168 2 1 256 49 16/20 29/30 10 50/72 98 10/14 13 7 Mean Mean Burin Microblade Width Width 1.31 .69 - .72 - .77 1.7 .90 1.8) .90 2.1 .80 1.46 .78 1.26 .76 1.28 .91 1.31 .88 1.15 (.80 1.23 .76 1.24 .76 1.23 .74 — .62 - .66 1.88 .88 a” N 380 411 102 15 2) 143 25 .37 6 .29 14 .36 4) .24 95 .17 189 .20 20 27 56 253 Table 12 Metric Parameters for the Sample of Burins From the Closure Site Standard Mean Deviation Skewness Kurtosis Length (L) 1.98 0.42 0.59 3.1 Width (W) 1.27 0.24 0.04 3.8 Thickness (Th) 0.44 0.12 0.67 3.6 Haft Length (HL) 1.07 0.36 0.80 5.0 Spall Length (SpL) 0.85 0.37 0.67 3.3 Mid Spall Thickness 0.30 0.10 1.20 5.7 (MSTh) Number of Spalls 3.19 1.90 0.98 3.3 (#SP) 0.0 3.0 N=173 Cm. Cm. Normal Variables Table 13 Metric Parameters for Burin Samples Within the Closure Site L W Th ‘m o— N n. 0' N 'm HuH nmnnamn x nououm .owmuuom condo x 0mm .owmuuom aoHco N um%mumxH x mHm>me Ho>wm mxooum om D nem< 6:“ an sumaum> umamuom 66m ma 64669 260 base (i.e., side-notched) to parallel-edged (i.e., stemmed) (see Table 14). There seems to be a constraint operating on haft width, such that an increased point width (or size to increase the shock of impact?) still had to fit within some constraints on hafting, but this possibility requires additional study on assemblages with large numbers of end blades. There are thus several possible sources of variation in end blades which may need to be modeled in addition to the general increased variety predicted by the model. Once again we find that the model will not necessarily predict all of the variation which we observe and that additional models will be necessary. However, the data presently available on end blade variety in the Arctic Small Tool horizon fit the model as presented. End scrapers are also apparently subject to variation in hafting morphology, but apparently to a lesser degree. The end scrapers depicted in Table 15 vary somewhat in size, but the major variation is in the treatment of the proximal or hafted end. Again, this varies from a concave- converging stem to a straight-converging stem to parallel- edged (stemmed) to concave-edged with straight or convex 'base (i.e., side-notched). While the concave-converging stem (often called flared end or even eared) is found throughout the horizon, as is the straight-converging stem, the parallel-edged (at least in its small form) is ap- parently found in the West, while side notched varieties 261 appear in the East. Again, we must be careful because some assemblages have extremely few end scrapers and the vagaries of sampling error are difficult to control. In general, the increase in side-notched artifacts in the East may reflect the operations of the proposed model, or may have an as yet unrevealed adaptive significance. There is certainly an increased variation in end scraper forms in the eastern extension of the horizon, and these data support the model as presented. It should be apparent that, while we cannot fully ac- cept the model as proposed, we have found no data which would cause us to reject it or to make substantial altera- tions. Rather, it would seem appropriate that we build other models upon it in an attempt to increase our ability to predict artifact variability in the spread of the Arctic Small Tool horizon. To accomplish this, it will be neces- sary for other researchers to produce their data in a manner which would facilitate this kind of analysis, with its concern for the comparability of sample variation and for the assessment of statistical significance of infer- ences. Obviously, such analyses will require the explicit use of rigorous and precise sampling and analytical proce- dures far beyond the present capability of the data available in print. The data considered above support the prediction of the model that the dispersal of human pOpulations into an unoccupied niche will cause behavioral variation in 262 proportion to Space and time. Specifically, by holding the time dimension relatively constant by limiting the compari- son to the Arctic Small Tool horizon, the model adequately predicted that some of the variation in artifacts was a re- sult of the dispersal vector. It is important to note that this model is not seen as explaining all of the variance in the ASTh nor is it seen as Operating to the exclusion of other processes. Rather, the model demonstrates one source of variance and must be used with other models in an at- tempt to explain variations in human behavior. It requires further testing. However, the formulation and testing of such models must be based on a research strategy based on explicit and precise methods. Such clarity should reduce the large amount of unspecified speculation and insignificant inter- ence and pave the way for somewhat slower but hopefully surer steps towards a behavioral model of Eastern Arctic prehistory. Chapter 7 An Ecological Systems Model of Culture Growth, AtrOphy, and Stability in the Pre-Dorset Previous Mbdels of Pre-Dorset Cultural Change Only in the last several years have students of Arctic archaeology attempted to develop explicit models, while the history of such studies is replete with the use of largely implicit models. Inherent in the concept of Pre-Dorset was the idea that such a culture develOped into Dorset (Collins 1954bz304), although the exact nature of the development was unspecified. Model No. l -- General Development of Pre-Dorset cul- ture into Dorset culture. Mbdel No. 2 -- General DevelOpment of Pre—Dorset cul- ture into Dorset culture in some areas, with subsequent migrations to other areas and replacement of cultures there (possibly following local extinctions, but this was not specified) (see Larsen and Meldgaard 1958; Mathiassen 1958). Model No. 3 -- General Development of Pre-Dorset cul- ‘ture into Dorset culture in some areas, with subsequent cultural diffusion to other areas influencing cultures al- ready existing there (Taylor 1968a). 263 264 Two variants of Models 2 and 3 existed, depending upon the nature of the deve10pment used in the model. Meldgaard saw significant acculturative influences from.other cul- tural systems involved in this deve10pment, particularly from southern "boreal" cultures (Meldgaard 1962). Taylor (1968a), Maxwell (1962) and others saw these deve10pments as evolutionary, with minimal influences from outside a Pre-Dorset cultural system. While many of these ideas were mainly concerned with the transition from Pre-Dorset to Dorset, such a transition was seen as the culmination of a series of changes within Pre-Dorset which may have been ac- celerated by environmental or acculturative influences (Taylor 1968a:103). While Taylor pointed out continuities in regional se- quences of deve10pment in Greenland (l968a:93) and at Igloolik (l968a:99), Maxwell demonstrated the evidence for continuous cultural development in the Lake Harbour region through the Pre-Dorset and Dorset periods (1967, l973a). Maxwell suggested an alternative model which went beyond the demonstration of cultural continuity through time. Model No. 4 -- Pre-Dorset -- Dorset as a closed cul- tural system, with internal variations in technological style which are seen as minor stylistic drift (Maxwell 1967: 10). Maxwell emphasized the essential conservatism of the Pre-Dorset -- Dorset cultural continuum, and characterized the Eastern Arctic region as participating in a sphere of cultural interaction with varying levels of intensity 265 through time. The deve10pment of these four models led to a number of attempts to reconcile the data available from across the Eastern Arctic and to a heightened interest in archaeologi- cal systematics and models. Following Maxwell's lead (Maxwell 1960) and the accumulating evidence that the Eastern Arctic environment had been subjected to a number of changes of potential significance to prehistoric hunters, I attempted to assess the potential relations be- tween changing environments and changing cultures, hypothe- sizing that climatic changes caused cultural changes by in- fluencing the rate of cultural interaction among Arctic peOples (Dekin 1968, 1969, 1970, l972b). These studies focused on the coincidence of climatic and cultural changes in a number of different areas of the Eastern Arctic, and can be seen as initial steps in the deve10pment of a model of cultural change as a response to changes in environment. This model was made more explicit with specific reference to the prehistory of Greenland (Dekin 1972a) in which some of these hypothesized causal relationships were tested. Similar studies by McGhee (l970c), Fitzhugh (1972), and Bookstoce (1973) suggested the general utility of such an approach to the study of cultural change which might be summarized as follows. Mbdel No. 5 -- Changes in the development of Pre- Dorset -- Dorset culture result from the adaptive response of a cultural system to changes in its environment. This 266 model makes explicit the cause-effect relationship between environmental change and cultural change, such as discussed with regard to the prehistory of Greenland (Dekin l972a). Fitzhugh's pOpulation compression model prOposed in 1973 (1973; personal communication) was related explicitly to the Hudson Bay region and to environmental changes and may be seen as an example of an attempt to apply such a model to a specific region. However, the utility of general models for the under- standing of culture change throughout the Eastern Arctic was questioned by McGhee (1973; personal communication) who proposed a model based on historical particularism, seeking no link among the apparently discrete historic events in human populations about the periphery of the Eastern Arctic. McGhee considered the extinctions of such marginal popula- tions as a "normal" occurrence throughout much of the human occupation of this region and presented a general re- vision of Model No. 2 (above) to include his data on the juxtaposition of Pre-Dorset and Independence I cultures on Devon Island and Noble's Canadian Tundra tradition (Noble 1971). Thus, the last several years have seen an increased interest in the formulation of explanatory models in Eastern Arctic prehistory, and several attempts at the derivation and testing of such models. It is in this cli- mate and with this background that the following models are proposed. 267 An Ecological Systems Model of Culture Growth, Atrophy, and Stability in the Pre-Dorset The utility of an analytical approach involving a con- sideration of the ecological relationships between foraging societies and their environment has been amply demonstrated in the literature of anthrOpology. Recent attempts by McGhee (l970c), Bockstoce (1973) and Fitzhugh (1972, 1973) have utilized an ecological framework to relate changes in environment to changes in culture, with Dekin's more formal attempt to depict the nature of these changes in linked sys- tems using data from the prehistory of Greenland being the more explicit and complete statement of the methodological aspects of such studies (l972a). The model deveIOped here is a direct outgrowth of this previous research, based on similar definitions and postulations using theoretical de- ductions from ecological and systems theory (see Rosen 1972: 47-48) (for comments on the applicability of such methods, see Hill 1972:97). The model of environmental (ecological) change is a product of the researches and analyses of a number of scientists, including Odum (1959), Levins (1966), Dunbar (1968), Margalef (1968), Dillon (1970), Geist (1971), Slobodkin and Sanders (1969), Irving (1972), Barry and Chorley (1971), Bryson and Wendland (l967b), Barry and jPerry (1973), Dansgaard et a1. (1969, 1970, 1971), Johnsen et,al. (1970), Lamb (1966), Langway (1970), Vibe (1967, 1970), Bryson (1974) and Dekin (1972a, 1972b), much of 268 which has been summarized earlier in this paper. The model of human behavioral change is a product of a great deal of anthrOpological research during the last cen- tury, with intellectual roots going back to Durkheim (1933), who prOposed that increasing pOpulation was the general cause of increased division of labor (1933:262; see also Allen 1971:Ch. 6). The role of increased population and in- creased social scale (Siegel 1952:138) in causing other changes has been prOposed by several anthrOpologists, in- cluding Siegel (1952), Redfield (1950:67), Wilson and Wilson (1945:83-88) and Swadesh (1951:3-4) (see also Kushner et a1. 1962:Category Five, pp. 20-22; Dekin 1968, 1969). However, it is only recently that archaeologists have turned their attention to the consideration of these prOpo- sitions and to their testing. An explicit model of cul- tural growth has been used by Plog (1974) in the study of prehistoric change in southwestern prehistory, with speci- fic reference to the Basketmaker--Pueblo transition. The incorporation of certain aspects of this model of grwoth may be seen as my attempt to expand the utility of his general model, by relating it to foragers in a changing en- 'vironment. The addition of a model of system stability is derived from.Maxwell's earlier suggestions (1967) and IPlog's point that ". . .in order to understand change, we unist be as concerned with nongrowth as with growth" (1974: 70). The conceptual scheme which I will use is similar to 269 that which I used in my previous paper on Greenlandic pre- history (Dekin 1972a), following trends in ecological and systems theory as exemplified by Ashby (1963), Buckley (1968), Harvey (1968), Clarke (1968, 1972b), Renfrew (1973) and Watson et a1. (1971). These, then, are the intellectual foundations on which the following model is based. A more complete statement may be found in Dekin, 1972a, and in Plog's Chapter 6 (1974), but the above outline should be sufficient to indi- cate the underpinnings of the following statement of a model for testing. 270 |—- — _ — . _L GROWTH TRA lncreasir‘ lncreasir Inc reasi n Decreasir lncreasin 1 P ,1, , _ Figure 40 An Ecological Systems Model of Culture Growth, Atrophy, and Stability in the Pre-Dorset 271 In general, the Pre-Dorset Behavioral System is per- ceived as an open system subjected to numerous inputs from other systems, in particular that of the biophysical en- vironment. Specifically, it is a complex adaptive system (Buckley 1968) coupled with an environmental system (Ashby 1963:Ch. 4/6) with linkages such that, in such a foraging behavioral system, the human behavioral system is more likely to be adapting to changes in environment than it is to be causing the environmental system.to adapt to changes in human behavior (see Dekin l972a for a more extensive dis- cussion of such linkages) (note that such may not be the case in 1974 or with other more highly evolved behavioral systems) (Clarke 1968:50,51,57). It is important for this perspective that we not confuse this approach with those making use of other types of systems, particularly homeo- static or deviation amplifying, neither of which is implied in this particular discussion. Explicitly, the Pre-Dorset behavioral system is not herein modeled as self regulating, but as complex and adaptive (Flannery 1973:52). By modeling the Pre-Dorset behavioral system, we are not partitioning the behaviors of "Pre-Dorset people" into those which are genetically programmed, idiosyncratic acci- dents, purposeful inventions, or learned as systemic norms. Implicit may be the attitude that much of the variation 'which we intend to study is the result of purposeful learned behavior. Much of this is the study of Pre-Dorset culture, in the many senses in which this concept has been 272 used, but any attempt at building a priori normative sys- tems will be avoided in the deve10pment of this model. While it is my explicit intent to model the relations be— tween the Pre-Dorset behavioral system and its biophysical environment, neither of these systems can be considered closed systems. However, the inputs to these systems from other similar (or dissimilar) systems may be controlled in the course of our analysis, or may be accepted, although unspecified or substantiated, and postulated. Specifically, the ultimate cause of climatic change in the Eastern Arctic is seen as external to the specific sys- tems under consideration, being radiational or orbital changes (inputs) from our solar system and universe. While there is acceptance of this position (Willett 1961:93; Schurrman 1965; Bray 1965, 1966, 1968; Bray and Struick 1963; Dansgaard et a1. 1969:378; and Dekin 1970) we will postulate that the ultimate cause of many changes in en- vironment is external to the systems under consideration here. Other environmental systems may have significant in- puts to the Eastern Arctic during the time periods under consideration here, however these would be difficult to de- termine, given the data at hand. 'Migrations of biota into the Eastern Arctic are but one possible input of potential significance. It is also possible that other sub-systems of the earth's biosphere would have influenced the Eastern .Arctic environment (such as volcanic eruptions increasing 273 the particulate matter in the earth's atmosphere, causing weather changes, etc. Kennewell and Ellyett 1964:356), and we must be careful not to preclude such inputs into these systems. It is also likely that other human behavioral systems could have provided inputs into the Pre-Dorset behavioral system, particularly along the southern fringes of its dis- tribution. For our present purposes, acculturative innova- tions are not unlike other innovations. They may be re- garded as derivations from other systems and their subse- quent spread within the Pre-Dorset may be studied just as if they had been innovations within a geographic sub-set of the Pre-Dorset behaviors. The concept of Vector of Environmental Change places emphasis on the direction of the change, or of the processes of change, through time. These are statements of tendency (Plog 1973b:656) designed for a diachronic study, rather than a study of the transition through several clearly de- terminable states. While it may be useful to categorize and model some changes as transitions between periods of relatively steady-state, such studies frequently focus at- tention on the states, rather than on the processes of change (Sopher 1973:102; Kuhn 1970:125). A recent example of this was the model of environmental change proposed by ‘Vibe (1967) in which he defined several stages of recent climatic change in Greenland (pulsation and stagnation). ‘While my own studies of such changes made use of a rather 274 continually oscillating model (Dekin 1968, 1969, 1970, l972a) which seemed to be a better fit with the data used to derive them, Vibe's model of the alternation between two stages did not prove to be as useful as he had first be- lieved, and he altered it to present a better fit with the continuous changes which his data demonstrated (1970). The environment of the Eastern Arctic is a dynamic system.noted for changes not observable or significant in other areas. Arctic weather and climate in this region are among the least stable elements of this environment, and are believed to pass on this instability to other compo- nents of the system.(see Dekin l972a for an extensive dis- cussion of this instability; Ashby 1963:83; Foote and Greer- 'Wootten 1966; Lamb 1966; Dunbar 1968; Vibe 1967, 1970; Nelson 1969:34-53). Thus, the use of the vector concept focuses our study on the processes of change themselves, rather than on the presumably steady states which border them. It should be obvious that the focus of a study on change may make some traditional concepts of less utility than others designed specifically for such studies, and I expect that the initial awkwardness of such concepts will hopefully be overcome by an appreciation of their utility. The essential points to this model of Vectors of Environmental Change are as follows. 1. While the dyna- mics of the earth's atmosphere may be seen as a large Inechanism for heat exchange between the poles and the equa- torial region, driven by the differential solar energy 275 owsmno amquBCoHH>cm mo muouom> mo Hmpoz < H8 ounwwm 43‘ ‘45 05.00300 . _ a 05.90300 — a 3:32.23... 68 3:36; Jan co..o.:u:u gang—.of J...— a 9.300300 023000 mucosu »:>:u=oo= fi 2:33.... LJ » 3:33:10... .6... 20.h.m__ao.m .2.- ,ll. Uzwxmm0<~h< co_.o_:u..u .0:ON dc. 20:<.O(u 05:00:... J(.¢hmwu-upa(¢—Xw 02.2u(3 _UwCOECObm>-um *0 asOqu> 276 available in these regions, the driving forces behind this exchange mechanism are extraterrestrial in origin, and changes in the dynamics of such exchange systems may be seen as being ultimately caused by variations in the kind and amount of radiation reaching the earth, particularly in the high latitudes (Dansgaard et a1. 1969; see above). 2. Such variations in radiation cause immediate and lasting changes in the earth's atmospheric circulation (King-Hele and Walker 1960; King-Hele 1964; Tucker 1964; Wexler 1956; Wilcox et a1. 1973; and Willett 1951), leading to variations in air mass movements. In northern regions, these have been characterized as periods of zonal circula- tion or meridional circulation, depending on whether the dominant pattern was of east-west air mass movement (zonal) or north-south air mass movement (meridional). Zonal cir- culation implies more stable and predictable weather condi- tions with a decreased exchange of air between Arctic and temperate regions, while meridional circulation implies less stability and predictability with fluctuating air mass interfaces and subsequent weather conditions (Bryson and Wendland 1967b; Barry and Perry 1973:364; Barry and Chorley 1971; Barry 1967). Increased extra-terrestrial radiation causes increased zonal circulation, while a de- crease in solar radiation causes increased meridional circu- .1ation, particularly with regard to northeastern Canada and the Eastern Arctic . 3. A change in solar radiation (or other forms of 277 radiation) results in changes in the energy available at all trophic levels and to changes in the general productivity of the ecosystem (Odum 1959:Ch. 3; Dunar 1968:1,98). In— creases in solar radiation result in an increase in energy available and increased productivity, while Opposite changes are similarly linked. Changes in the stability and predic- tability of weather and climate also influence changes in productivity as a result of the changing frequency of de- struction of "marginal" communities (see MacArthur 1972; Dunbar 1968; Margalef 1968:63). 4. Changes in predictability, stability and produc- tivity cause changes in the biomass, with Obvious implica- tions for the human exploiters in the food change. In- creased predictability, stability and productivity cause increases in the biomass, and Opposite changes are simi- larly linked (Dunbar 1968:98; MacArthur 1972:183; Odum.1959: Ch.3). Inasmuch as the Arctic ecosystem is "immature" (Dunbar 1968; and Others) with a relatively "simple" food web, fluctuations in the general stability, predictability, and productivity greatly influence the internal dynamics of the ecosystem, providing a useful example Of Ashby's Power of the Veto (Ashby 1963:83). We should expect the human in- habitants of this region, especially at a foraging level of subsistence capability, to adapt to such changes. The thrust of environmental vectors such as I have described should also influence human behavioral systems, and the 278 following is a model of such changes. The use of explicit models of human behavioral systems has been increasing in the last decade and we may choose from any number, varying in sc0pe and explicitness. By modeling the Pre-Dorset Behavioral System as a complex adaptive system (and an open system -- Chorley 1962:3) I am postulating that the relations among aspects of human be- havior which we may wish to study are complex and that such behavioral interrelations are multivariate (Spaulding 1973: 344-5,352; Chance 1968:5). Moreover, I also postulate that the Opportunity for continuing feedback among behaviors makes the search for unilineal causal relationships diffi- cult, if not impossible, particularly at the level at which we are forced to operate because of the limitations in- herent in data which are archaeological (see above). This postulate is supported by Plog's failure to substantiate a model in which simple linear relations were hypothesized, having to revise the model to include more complex relation- ships (Plog 1974:155-7). By modeling the system as an Open complex adaptive system and by paying particular attention to the environ- mental inputs into this Open system, we are avoiding several limitations which have plagued previous investigators. This model eschews simplicity (Spaulding 1973:344-5,352) .and seeks to model complexity. This model does not force ‘us to consider human behaviors as a closed system, nor does it assume that such behaviors are homeostatic 279 (Flannery 1973:52), nor does it assume that environmental changes are "accidents" (Plog 1974:33). Rather, it is a model which attempts to make explicit the relations which occur within two linked systems, both of which may be changing. It is an adaptive system, such that a change in one (environment) may result in input to the other, with re- sulting response. It is apparent that there is continuous variation between daily weather and its long-range counter— part, climate. While the day-tO-day changes result in day- to-day responses, it is not this level which we wish to model. By choosing a scale which encompasses those various archaeological phenomena from the Eastern Arctic called the Arctic Small Tool tradition which span the time period be- tween 4050 and 2750 B.P. (radiocarbon years), we wish to examine changes in the entire system.during this time span. This point is significant with regard to the input thres- holds from the environmental system to the behavioral sys- tem. It would be difficult to imagine a change in input which Occurred in the time span of one day and it would be difficult for us to determine relevant data, given the limitations with which.we Operate (see above). On the other hand, a change in input which took the entire time span of 1300 years to occur would also be difficult to de- termine and analyze. There are probably significant differences in the zmmounts of changes in particular variables which are suf- ficient to cause different responses, and there may be time 280 delays as well, both of which are, unfortunately, not well enough studied to be modeled at the present time (see ~ Bryson and Wendland 1967b for a tentative model of thres- hold and lag among environmental variables; also Barry and Perry 1973:350). Minor changes in environment may have an immediate ef- fect on movements of peOple (halting movements in a storm, for example) and minor changes in the frequency of good traveling weather may affect the ease with which people come into contact with other social groups (Dekin 1969). Thus, I have modeled trajectories of change which have low input thresholds and which relate such minor changes direc- tly to changes in social scale. Major changes may also cause large-scale fluctuations in pOpulation density and distribution, which will ulti- mately also affect the frequency of social contacts. These are modeled as growth or atrOphy trajectories. In this model, there are two levels at which environmental change may affect social contacts. These are modeled as complex, multi-varite, and subject to different thresholds and re— sponse rates. This proposed model of the Pre-Dorset behavioral sys- tem and its changes considers changes in five dimensions of human behavior: population; behavioral differentiation; social scale; technological variegation; and technological evolution. Just as I modeled two vectors of environmental change (above), there are two significant trajectories of 281 linked changes which we can model and two somewhat less complicated trajectories resulting from changes in social scale (Clarke 1968:75-77; Flannery 1973:47). The results of the warming vector Of environmental change are initial increased stability and predictability of weather patterns, with zonal circulation and fewer blocking highs and far northward storm tracks. These changes alone would result in improved hunting conditions and in greater ease of travel, thus influencing social scale directly (Morrill 1970). However, prolonged and increased warming vector change results in significant ecological systemic changes such as increased biomass and an increased ability to support greater numbers at higher trophic levels. Without marked human behavioral change, there is an increase in efficiency of the food quest which may result in fewer hard times and an ultimate population increase (Freeman 1971:1013; Balikci 1970:151). With this increased effi- ciency and pOpulation increase (increased population density or extent) there will also be an increase in behavioral dif— ferentiation, marked by increased specialization of be- haviors (rise of specialists and increased divisions of labor) and specialized locations (increased spatial varie- gation of behaviors -- single purpose houses, sites, regions, etc.) (Struever 1968:138; Plog 1974:62—5, etc.; Durkehim 1933; White 1959:293; Allen 1971:Ch. 6). The increased population density and differentiation lead to increased social scale (Pothier 1968:38), whether llllllllllll‘ ll llllllll III 1411 D. O 11 .1 E .v 282 purposeful from increased trade or other exchange mechanisms or simply from the increased pOpulation dynamics (Morrill 1970:147-152; Sopher 1973:112). However, as mentioned above, increased ease of transportation may also lead to increased social scale (Morrill 1970:150 reaches the same conclusion with more modern data). If changes in the environmental vector are the Oppo- site to those described for the warming vector, I have modeled a cooling vector which results in decreased stabi- lity and predictability of weather patterns, with meridional circulation and more frequent blocking highs and a northward movement of depression tracks. These changes alone would result in less Optimal weather conditions for hunting or travel (Boas 1964:19; Foote and Greer-Wootten 1966), re- ducing social scale. Prolonged cooling vector change re- sults in significant changes in the ecological system, with decreased biomass and a decreased ability of the food chain to support higher trophic levels. Without marked human be- havioral changes, there is a decreased efficiency of the food quest which may result in more frequent hard times and an ultimate population decrease (possibly both gradual and catastrophic) (Freeman 1971:1013; Balikci 1970:151). These changes result in a corresponding decreased specialization of behaviors within society and decreased specialized loca- tions (again, Plog 1974; Durkheim 1933; White 1959; and Allen 1971) . These changes result in decreased social scale as a 283 result of decreased exchange among specialized locations or simply the result of decreased population density (Morrill 1970:147-52). Thus, both the warming and cooling vectors are seen as causing changes in social scale, both directly and through the intervening changes in pOpulation and behavioral dif- ferentiation. Social scale .refers to the number of persons consciously related in a society and the intensity of their relations (Wilson and Wilson 1945)" (Kushner et a1. 1962:20). There need not be face-to-face contact and communications frequency may be one measure of intensity. Increases in social scale result in increased communi- cation among social groups Of various sizes and at various levels of analysis. These communications tend to maintain the integrity of the social unit and to spread ideas and behaviors throughout the social system engaged in this interaction (Yellen and Harpending 1972:248). As long as conditions of close internal communi- cation prevail in the society, the culture tends to remain uniform throughout; changes appearing in one part of the area either spread quickly through the whole territory or are drOpped be— cause Of the cultural influence of the bulk of the society (Swadesh 1951:3-4). Reductions in social scale result in decreased communica- tion and in a decline in behavioral homogeneity in the social system.(Sopher 1973:106; Yellen and Harpending use the concept of social nucleation much as I used social scale, 1972:249,251). 284 Where the size of the territory or other circumr stances prevent the fullest internal contact of the group, there is a tendency to develop local variations of the culture which may eventually amount to major differences (Swadesh 1951:4). Changes in social scale lead to changes in technological variegation, with increased social scale causing increased exchange of ideas and technological similarity. Such simi- larity in technological systems results from the behavioral equivalent of the Law Of Competetive Exclusion (Plog l973a: 657; Odum 1959:231; May 1973:139; and MacArthur 1972:179) in which increased communications leads to increased compe- tition among behaviors and the expansion of the most effi- cient (read adaptive or apprOpriate) behaviors (McGhee 1973 discusses this at an intuitive level). In general, such changes result in increased behavioral complexity (techni- cal complexity) and diversity which we call technological evolution (Wilson and Wilson 1945:88). Decreased social scale, then, leads to decreased com- munications among spatial units, decreased competition among behaviors, maintenance of diverse behaviors among spatial units (heterogeneity) (increased technological variegation) and a decreased rate of systemic technological change. Several aspects of this model must be made explicit. It does not attempt to account for all causes of increased social scale nor for all resultsfiof increased social scale. .As Plog has pointed out, we may need many models of be- ?havioral change in order to understand such complexities 285 (Plog 1974:156). There is not explicit modeling of feed- back Or lOOped changes in this model, although some could be suggested (increased social scale leads to increased occupational specialization -- Kushner et al. 1962:21; Wilson and Wilson 1945:83). The linear sequence of linked changes in the behavioral system could be seen as but one of a large number of similar sequences possible in complex adaptive systems. The universality of this particular se— quence remains to be assessed. Test Conditions The following is a stipulation of the data structure we would expect if the above model is appropriate. With reference to Chapter 4 and particularly Table 8, the inputs into this model would be as follows. The migration of Arctic Small Tool tradition popula- tions into the Eastern Arctic occurred during the waning phases of a warming trajectory during a time when there is no evidence for marked environmental change. The first major change in the Eastern Arctic environment occurred after most of the Eastern Arctic littoral had been occupied (with the exception of southern Greenland). This change was the advent of a less stable cooling trajectory in both the land and marine ecosystems at approximately 3500—3600 ‘B.P.. This trajectory involved increases in marine ice, changed less-Optimal marine conditions, a southern movement of the tree line, increased meridional circulation and 286 concomitant changes in the distributions of marine and land fauna (see Chapter 4). This change is followed by a more marked change 22- 2750—2500 B.P. which again involved a cooling trajectory. Were it not for the relative recency of the AST migra- tion, we would expect the above model to apply to Pre-Dorset behavioral systems prior to the marked change EE- 3500-3600 B.P.. This might have been modeled as the waning portion of a trajectory that resulted from a warming vector (see Figures 23 and 24), but the dispersal trajectory and re- lated changes (see Chapter 7) are a more apprOpriate model. If the above model is apprOpriate, we would expect ob- servable changes in the archaeological record to follow the marked environmental change 23: 3500-3600 B.P.. These could be of two general types, depending upon the strength of the cooling vector. With a relatively minor cooling vector, we would ex- pect the following: 1) increased technological variegation among regions populated by the ASTh (a process of region- alization resulting from a reduction in social scale); and 2) a general decreased rate of technological evolution (lack of evidence for increased technical . efficiency or complexity following regionalization). With a major cooling vector, we would expect an atrophy ‘trajectory of changes, with the following changes in the archaeological data: 1) decreases in the number of sites, areal extent of sites, or a combination of decreased 287 numbers of sites but larger sites (nucleation); 2) increased homogeneity within sites (fewer areas Of activity specialization within structures and sites), which is decreased behavioral differ- entiation; 3) increased technological variegation among regions pOpulated by the ASTh (a process of region- alization resulting from a reduction in social scale); and 4) a general decreased rate of technological evolution (lack of evidence for increased technical efficiency or complexity following regionalization). If the vector of environmental change is not suffi- ciently strong to exceed the thresholds, we would expect that none of the above changes would correlate with the en- vironmental vector and that we would need to search for alternate models of such changes. The key implications of this model are that such changes exist and that they are correlated to the extent that vectors of environmental change are followed by vectors of behavioral change. Explicitly, we would expect the period following 9&- 3500-3600 B.P. to be characterized by the development of regional differences in artifacts and a reduced rate of artifactual change such that there is artifact stability through time within each region. If there is atrophy, we Would expect sites within each region to decline in fre- quency, extent, or to nucleate. Sites within each region 'would contain artifacts representing a wide range of acti- 'vities with relatively little variation among sites within each region (when site assemblages are compared, they demonstrate similar variety among artifacts). Summarily, 288 the dimensions of archaeological data which we need to Ob- serve are as follows: 1) site frequency and size; 2) variety of artifacts within sites; 3) variety Of artifacts among regions; and 4) artifactual change through time within each region. Testing the Model Before undertaking an assessment of the degree to which the dimensions of archaeological data correspond to the above predictions of the model, it would be advisable to re- iterate some of the cautions expressed above with regard to the sample of data which we have available (see Chapter 5 above). For an accurate comparison of the dimensions of the data we should have an unbiased sample representative of the populations which we wish to study. To insure a lack of statistical bias, we generally take a random sample, in which each element of the population has an equal chance of being selected for the sample (Cochran 1953:11). To my knowledge, such a sampling technique has never been used to derive any data from any population in Eastern Arctic pre- history, thus we are not able to say that we are dealing with unbiased samples. However, we are not able to specify all Of those samples which have been biased by field or ana- lytical techniques (see Chapter 2 and Chapter 5 above), and VNe cannot a priori establish all sources of bias which may have influenced our samples. This problem is most difficult to cope with when we 289 attempt to assess variations in site frequency and site size through time. Without assurances that we have repre- sentative samples of each time period of occupation, we have difficulty in assessing this dimension of the data. Nevertheless, there is no way to convert non-random samples (Tate and Clelland 1957:47) nor to compensate for non- randomness. We must make do with what we have available, being as precise and explicit as possible so that our mani- pulations of the data do not serve as additional unknown sources of variance. With the exception of absolute differences (presence/ absence), it is presently impossible to determine if there are significant population changes within any of the regions in which ASTt sites existed in the Eastern Arctic just after 3600-3500 B.P., and thus it is impossible for us to evaluate this dimension of the model as presented. 0n the other hand, this time period marks significant changes in the distribution of ASTt pOpulations and in the deve10pment of regional differences. The migration of Sarqaq peOples into southern Greenland occurs just after this time. The Independence I culture of northern Greenland apparently abandoned its former range at about this time. There is apparently a hiatus in the ASTt occu- pations of Labrador at this time, and the origins of Noble's newly named Canadian Tundra-Taiga tradition (Noble 1974:162) consisting of derivatives of the ASTh from Victoria Island to Manitoba date from just after this time (Taylor 1967 and 290 my Table 4). Thus, following 3600-3500 B.P. (see Figure 6), we find the development of several distinctive variants from an Arctic Small Tool horizon and several areas from which previous occupants vanished. The regional differ- ences could be categorized into five sub-traditions: Canadian Tundra-Taiga; Pre-Dorset; Sarqaq; Refuge; and Independence 11. The timing of these regional variants is as predicted by the model. If the model is appropriate, within each of these regional variants, we would expect a limited amount of artifactual change through time, as these technological tra- ditions are perpetuated without extensive changes from ad- jacent regions. Such conditions should persist (according to the model) until such time as the level of social scale is increased bringing more extensive contacts with adjacent areas, unless there are other significant inputs from the environments of these sub-systems which the model does not predict. Let us then assess the nature of artifactual change within each regional variant to see if the prediction of the model is in accordance with the data. During the time covered by our consideration of the model (until 22: 2750 B.P.) the northern sites of the Canadian Tundra-Taiga tradition represent a single rela- tively stable tradition, evidenced by the continuity in sites from Banks and Victoria Islands -- Umingmak, ‘Wellington Bay, Buchanan, and Menez (Taylor 1967,1972). 291 This continuity is in "typical" chert burins, side blades, scrapers and end blades as well as the characteristic quartzite bifaces (large and thin) and occasional quartzite end blades of traditional form (Taylor 1967,1972; Gordon 1972). The southern sites in this Canadian Tundra-Taiga tra- dition are three in number (Rocknest Lake Complex -- 92, 1200-900 B.C.) during the time period in which we are in- terested, and demonstrate their greatest similarities with those northern sites just described. These early southern sites are all at or just on the tundra side of the tundra- taiga border and all have an orientation to the adjacent waterways (Noble 1971:107-8). Neither Taylor nor Noble de- scribes any changes in artifact form or stylistic variation through time in either of these sequences, although subse- quent occupations in both the northern and southern por- tions of this distribution demonstrate continuity from this Canadian Tundra-Taiga tradition. It is interesting to note that subsequent developments from this general similarity demonstrate divergence, suggesting that Noble's Canadian Tundra-Taiga tradition be modeled as a diverging tradition (Wauchope 1956:43). These data are in accordance with the model. If the model is to apply in the future in this area, we should not expect to find sites of the Canadian Tundra-Taiga tradition before 3600-3500 B.P.. McGhee's Bloody Falls site is an enigma because it is 292 apparently bracketed in space by sites with artifact simi— larities, but it lacks the large and thin quartzite bifaces which characterize the Canadian Tundra-Taiga tradition. Its radiocarbon date is bracketed by those from Umingmak and Wellington Bay (see Figure 6), yet it contains no quartzite artifacts (McGhee l970a) with a sample of ca. 250 artifacts. However the overall nature of its assemblage (with ground slate adzes and end blades, polished burins, worked copper, etc.) suggests that it belongs in the same category as the Canadian Tundra-Taiga tradition, but that it may represent atypical activities conducted at Bloody Falls (season of occupation, possible idiosyncracies, etc.) It is impossible to evaluate the adequacy of the excavated samples from the Canadian Tundra-Taiga tradition, as McGhee's Bloody Falls site is the only one for which arti- fact totals are published (Umingmak has more than 400 and Buchanan is in the hundreds). The Sarqaq data from.West Greenland also demonstrate continuity in artifact style and form during this period under consideration, but with moderate sample sizes (Larsen and Meldgaard 1958; Taylor l968a:Table 23). If we accept Taylor's seriation of these data, then the continuity is more striking (again recognizing that seriation itself is biased in favor of continuity--Taylor l968a:91). However, the data depicted by Larsen and Meldgaard and Mathiassen (and re-depicted by Taylor) when combined with the radio- carbon dates available from.West Greenland (see Figure 6) 293 demonstrate technological continuity between EE- 3500-3600 B.P. and ca. 2750 B.P., which is a data structure in ac- cordance with the model as presented. In order for the model to continue to be in accordance with the data from West Greenland, we would expect to find no evidence for major technological change during this time and no Sarqaq sites earlier in time than the environmental changes noted above. The recent resurgence of interest in the archaeology of the Queen Elizabeth Islands (McGhee 1973, 1974) has re- sulted in enigmatic data subject to differential interpre— tations. It is, however, clear that this region contained a rather distinctive derivative of the Arctic Small Tool horizon, most closely related to the Pre-Dorset from.the core area. While the following interpretations may not be in accord with those presented by McGhee, they fit both the data and the above model. There was an initial occupation Of this region by people whose behaviors were much like those of Knuth's Independence I occupations of northern Greenland. This oc- cupation has a radiocarbon date coeval with the earliest from Independence I sites from.Creenland, covers a rather limited span of raised beach features (22-24 m above pre- sent sea level), and contains a variety of lithic raw Inaterials, all of which suggest a rather limited duration of Occupation related to the initial migration of Independence I peOples to the northeast. 294 This occupation was followed by what appears to have been a relatively stable homogeneous occupation of this region for a longer period of time by what McGhee sees as a Pre-Dorset variant, characterized by a general hafting of end blades, end scrapers, and burins which involved side notching to a degree not seen in the core area (McGhee 1973). Evidence for a longer and stable Occupation is meagre, but consistent. Sites range in elevation from.19 to 28 meters above present sea level, are not characterized by signifi- cant differences in artifact content or stylistic varia- tion, demonstrate consistent utilization of a blue-gray chert, and apparently last until after 1000 B.C. (1-7344 -- 2900t90 B.P.) (McGhee 1973, personal communication), all of which are in accordance with the model as presented above. It is possible that this occupation overlaps in time with the Independence I occupation, since there is overlap in the elevations of sites (Andrews et al. 1971 suggest that the 9 meter range of these sites may represent 800 to 1,000 years of postglacial emergence in this area). While the tentative nature of these conclusions must be emphasized, these data from McGhee's reconnaissance are indicative of a stable, relatively homogeneous Occupation of this region by a regional variant of the Arctic Small Tool tradition, which we might call the Refuge tradition, to distinguish it from contemporary variants in adjacent ‘regions, one of which developed into Dorset. The Independence Occupations from northern Greenland 295 are interrupted by a marked population decline and possible abandonment at 92- 3600-3500 B.P.. This population dis- placement probably resulted in increased interaction of Independence I, Pre-Dorset, and possibly Refuge populations (if such a tradition were in existence at this early date) contributing to the deve10pment of Sarqaq culture. However, within three or four centuries the Independence I culture had been adapted to the changed ecological conditions in northern Greenland (Independence II), and made a successful reoccupation of the region, which lasted through the re- mainder of the time period under discussion here. Follow- in this re-occupation, there is no evidence of significant behavioral change until the subsequent abandonment of this region several centuries after 1,000 B.C. (Dekin l972a; Knuth 1967, 1968). McGhee's Independence II occupations may result from this abandonment, since the apparently post-date the Refuge tradition (McGhee 1973). The above data are all in accordance with the test con— ditions of the model presented above. The formation of five regional sub-traditions of the Arctic Small Tool tra- dition (Canadian Tundra-Taiga, Pre-Dorset, Refuge, Inde- pendence II and Sarqaq) follows closely the environmental change 22- 3600-3500 B.P. and demonstrates subsequent sta- bility through the remainder of the period predicted by the Inodel. If there were atrophy within these sub-traditions, we :nould expect changes within sites to reflect reduced 296 activity specialization, yet there is no evidence for such changes. The two sub-traditions where data are available are Sarqaq (from Taylor 1968a) and the Pre-Dorset core area (Igloolik and Lake Harbour). In both of these, there is marked continuity in the contents of sites, with no evi— dence of regular change in the degree of variety (hetero- geneity--see Whallon 1968:227-228) within each site or among sites. Table 16 depicts the variety in site contents in Sarqaq (from Taylor l968a:Tab1e 23) and Table 17 depicts the variety in the southern Baffin Island Pre-Dorset. An inspection of these tables demonstrates that statistical analyses to assess variability are not warranted, due to a lack of observable regular variation, except for sample size. Thus, it appears that there is no evidence in support of an atrophy trajectory resulting from the environmental change 22- 3600-3500 B.P.. The regionalization (Yellen and Harpending call this macro-fragmentation--1972:251) pre- sented above suggests a reduction in social scale, as pre- dicted by the model, and we should assess the evidence for gradual evolutionary change within these regions. Once again we are faced with a lack of published evidence of changes in artifacts and attributes with which to test this prediction, and we are forced to turn again to the Pre- .Dorset core area for data. There are two long series of data from the core area, one at Igloolik which is not yet published and one at Lake 297' xNNmV mm mm ma 0H m mm unuOHH Amm aaaaeamwema aoaame aoamv ANqu ANomv mm mm no on no on HH H o H w o n m HH «a o 0 ma 0 o n ma ON uwmwmuoapmH xsmmumcuovmfiz ANNmV ANMAV om an an 34 an ea c a a o a a N m ea m m N ea Hm m a an em mcHOHmH H uwommuoapwH vmvumm dommmsaoapwH Aquv me am mm ma aH e mm ufiommHOQDH mmmeeammm< encumm ca zumwum> 0H magma mOpMHnOHOfiz pom .mHHmem canon .mowuom MN Hmuoa m>opm ozu mo nw Ocoumdmom mouo< homouom mowm muoemuom pom mmwmam opfim pom pom ouoo opmHnOHOHz mopoanouOflz maamam canon cfiusm v—_-.i.~ IEK»~-_" -__ ... 2923 ANNNV ANN O ANO O aNN O NNNVO aNO O ANNVV ANN O Hz N6 N amv O ON O N N O N N m6>NaM mamNm ANOOO aNONO ANaNO ANONO ANNNO ANO O ANN O ANO O AaaNaam N6 N aav m NN ON O NN O N N aaNaaO OmaaNNoa ANNNO NNamO ANNOV aNOaO aNmmO ANOOO ANNOO ANNaO an ON ONN OON NO NOO am NON OO OO.O “w ANNOO ANNOO ANONO ANOOO aNOOO ANOOO ANOOO ANOOV ANO Om OOm woe NN NNO Om OON mON Oz.mm.m “w mm mom NOO aON OaO NN aON NNN z Nance Oa NNa ONO NO amO as NNN NON a>oaa was N6 “N O ON O N N O N N ma>Nas aumNm Oaaoae O O O O N O O O aaoaaaaom o H H H N o o o mung O O O N a O O N maxNNuaNaam N NN O m N O m N mumamuum aONO H «N w w a» H N m muommuom com a sa mm O Na 8 O O maOaNO OONO O86 use 0 NH OH N H o H m mouou OpmHnONOHz ON NON NNN NN ONN O NO NN aaemNOoauNz O am OON ON OON aN NNN Os aNNaOm aNaaO O OO NNN NO NON ON Na mm aaNuam O.N.O.O ON OOON ONuNN N NOON INN OOOM OOOM ON OOON ON OOON N :Oxe NN OOON mmmeAsmmm< ummuoaumum OH NumHum> RH mHnme 299 Harbour, to which this thesis makes a contribution. In his several preliminary reports, Meldgaard has pointed to the con- tinuous and predictable gradual changes which characterize the Sarqaq (herein called Pre-Dorset) levels at Igloolik (for example, 1962:92). MOre recent communications (personal com- munication to Maxwell, 1972) reinforce this continuity, par- ticularly evident in harpoon tips and in lithic artifacts (see also, notes and photographs of Elmer Harp, Jr., on these Igloo- lik materials). Nowhere in his papers or correspondence does Meldgaard mention significant technological change between 22- 2000 B.C. and 1000 B.C. with a period of more rapid culture change between 1000 B.C. and ca. 900 B.C.. With the possible exception of this later period of more rapid change, the se- quence of Igloolik data is in accordance with the prediction Of the above model. The sequence of Pre-Dorset sites from southern Baffin Island reported by Maxwell (1967, l973a, l973b) and par- tially in this thesis also demonstrate a lack of major changes in technology (Table 17, for example), with several trends continuing throughout the sequence. These trends are in artifact attributes or gradually changing frequen- cies of artifact types, but in no case do they reflect more tflnan the "refinement" of already existing techniques or re- llatively minor oscillations about a mean. Several of these cflianges through the Pre-Dorset period appear significantly (tifferent only when seen in the light of subsequent changes. innese.include the declining proportion of burins and burin Spualls in assemblages, the increasing frequency of ground 300 slate knives (see Table 17), the apparent increasing later- al thinning of burins (from incipient side notched to side notched) (Maxwell 1973a:308,309) and side-notching on knives and microblades. However, it is apparent that these are slow trends of change, certainly evolutionary as they develop from behaviors already present in the earliest assemblages, and are in accordance with a stable rate of technological change predicted by the model. The following aspects of the model were not substan- tiated by this analysis: 1) atrophy trajectory; and by im- plication; 2) growth trajectory. While it may be possible to test such a model on additional data from Eastern Arctic prehistory, the present data on the Pre-Dorset period do not accord with such a model and it is not apprOpriate. However, the portion of the model which demonstrated the impact of a lower threshold for input from trajectories of environmental change was in accordance with the data con- sidered above. In brief, the model predicted that the en- vironmental change 22- 3600-3500 B.P. would be followed by the development of regional variants from the initial Arctic Small Tool horizon, and that each regional variant would demonstrate moderate internal change (if any) for the dur- ation Of the trajectory modeled above. The data substan- tiated these predictions, suggesting that such a model may be appropriately applied to other environmental changes and cultural changes in the Eastern Arctic for further testing. Difficulties in acquiring data suitable for the 301 testing of the above model were encountered because of the failure of the published data to be couched in terms suit- able for the testing of such models. ‘When data from dif— ferent find-spots or even different sites are lumped to- gether for the purposes of analysis and depiction, it pre- cludes the study of variation within that data set. When variations in that data set become significant for the testing of hypotheses, we are forced to infer conclusions regarding variation from the data presented, which is Often presented normatively. Once again, the problem of sample inadequacy and bias was difficult to cope with or avoid. By making explicit possible sources of sampling error and limiting data sets to those with total artifact samples greater than 50 (in which there would be a 90% probability of finding at least one artifact in a random sample with a population propor— tion of 0.04--from Table 9) I attempted to minimize the possibility of inferring significant variation where none exists (a case of Type I error). This approach to modeling aspects of Pre-Dorset be- havioral systems has resolved some Questions which other ap- proaches were unable to explain. The independent deduction of the model from other relevant theoretical formulations allowed the independence of the model from the data--an es- sential requirement if "testing" is to occur. The model led to the recognition of several sub-traditions of the Arctic Small Tool tradition, all of which had at one time 302 or another been lumped into the category Pre-Dorset. More- over, the model explains these sub—traditions as resulting from the reduced social scale caused by significane environ- mental change. In actuality, this model only explains the maintenance of these differences and not their original source (much as the process of natural selection explains the maintenance of certain genetic combinations, but not why the original mutations occurred). For this, we may need to look anew at the ecological diversity in this region (as in Dekin 1974), focusing on the precise ecologic relations between human populations and their environment. In this way, the above model has not only contributed to our understanding of regional variations in the Arctic Smell Tool tradition, but has pointed the way for future research regarding the ultimate causes of such diversity (as noted by Yellen and Harpending 1972:251). By speaking precisely about behavioral processes and by an explicitly scientific approach to model formulation and testing, we have gone beyond the practice of "letting the data speak" and have placed the study Of human beha- vioral processes on a firmer foundation. If there is merit to such an approach, then we will need a revitalization of concern for archaelogical strategy and techniques of analy- sis, in order to provide the data necessary for these kinds of analyses. The model presented here, as substantiated by the available data on Pre-Dorset variation, goes beyond the 303 "what" of the processes of human behavior to the "why", and is only possible through the use of precise and explicit archaeological analyses. This has been but one step in the direction which I see as necessary for Arctic archaeology. The ultimate contribution of this ecological systems model may be to resolve the conceptual muddle regarding the nature of cultural changes in the Eastern Arctic and how to model them. The expansion of this model to subsequent pre- historic dynamics in this region should be a high priority project for future research. The integrity of the sub- traditions of the Arctic Small Tool tradition (Canadian Tundra-Taiga tradition; Pre-Dorset tradition; Refuge tra- dition; Independence II tradition; and Sarqaq tradition) can be easily tested by future field strategies and programs of radiocarbon dating. Taken together, these five tradi- tions may be seen to comprise Taylor's Carlsberg culture (Taylor l968az85). Chapter 8 Discussion: The Conceptual Setting Introduction -- New Archaeology in the Arctic There are several sources of variation in archaeologi- cal interpretations. These include: 1) variations in the data under consideration (including differences in the tech- niques of recovery, depiction and categorization); 2) variations in the methods of analysis applied to the data; 3) variations in the paradigm which each investigator brings to the data and analysis; 4) idiosyncratic phenome- na characteristic of the archaeologist (personality, intel- lect, and experience, for example); and 5) the history of past variations in interpretations. Few archaeologists have studied differences resulting from the personalities of archaeologists or from historic differences in interpretations which gave rise to long- standing disputes, although recent review correspondence suggests that such an approach might be a useful contribu- tion to the history of North American archaeology (see Dekin 1972c, l973c; Lee 1973, 1974; Kuhn 1970:200). At present, there seem to be three sources of variation sub- ject to study. Variations in the data available are basic 304 305 to the study of archaeology, since investigators excavate sites and are, at least initially, sole possessors of sig- nificant data. Unfortunately, the trend in the last few years has been towards the publication of syntheses, sum- maries, and conclusions, leaving the substantive data on which these analyses and interpretations are based unpub- lished and subject to the continued control of the original excavator. Earlier archaeology in the Arctic saw the pub- lication of complete data as a matter of course, with in- terpretations or conclusions somewhat infrequent. In short, variations in interpretations stemming from differ- ences in the data considered are becoming increasingly sig- nificant sources of variation in archaeological interpreta- tion. The last several decades have seen the introduction of numerous methods and techniques of analysis of archaeologi- cal data. While it is only recently that sophisticated sta- tistical techniques have been applied to Arctic data (for example McGhee l972b; Dumond 1974a), trait lists and seria- tions of archaeological data have been commonplace, al- though less so since the advent of extensive reliance on radiocarbon dating for chronology building. Variations in methods of analysis are not as significant as other sources of variation, because the methods used are dependent on the data under consideration and the analytical goals of the investigator. This brings us to a major source of variation which 306 has received little attention from Arctic archaeologists. In spite of the winds of debate over the paradigm.under which archaeological research is conducted, little of this debate has penetrated the Arctic. Few Arctic archaeolo- gists have made explicit the assumptions which guide their research, whether these relate to the definition of pro- blems worthy of study or to field data worthy of recording (see Kuhn 1970:47). This lack of explicitness has resulted in the necessity of searching for evidence of the implicit assumptions which underlie research -- a time-consuming and possibly unreliable process. Much of what has passed for a prevailing paradigm.in Arctic archaeology has assumed that culture, as a system of learned behavioral norms, should be the major focus of our studies, and that most of the variability in archaeological data was caused by cultural variation. To operate under such a paradigm, the entities of archaeological analysis (i.e., attributes, artifacts, structures, components, locales, etc.) were treated com- paratively and normatively (typologically). Hypotheses were presented to explain differences in the data and were evaluated as competing alternatives (Cole and Kleindienst 1974:353), assuming that one of them was "correct". Re- search was overtly inductive, or at least data-centered, rather than deductive. Much effort was expended in trying to develOp a space-time framework based on the assumption that morphological similarity was indicative of 307 contemporaneity (a type-fossil approach). Site-collections were the major unit of analysis, with within-site differences ignored or disregarded. Data cate- gories were collapsed within sites, lumping together data from contiguous structures, find spots, or even beach levels. The "homogenization" of sites precluded the effec- tive study Of the spatial dimension of human activity with- in a settlement. However, evidence of the existence of such a paradigm must be gleaned from scattered statements of research in- tent, methodology or results, as it was invariably implicit and not exposed for examination. Thus, the paradigm under which archaeological research has been conducted in the Arctic has been rarely stated and even more rarely ques- tioned. Archaeologists had been lulled into a state perhaps best described as mindlessness, following Charles Silberman's criticism of American education in general. Mindlessness is .the failure or refusal to think seriously about. . .purpose, the reluctance to question es- tablished practice" (1970:11); "it simply never occurs to more than a handful to ask why_they are doing what they are doing. . ." (1970:11). Such a concern with purpose charac- terizes much of the rhetoric surrounding the "New Archaeology", and it is apprOpriate to examine here the purposes and paradigms of Arctic archaeology. By urging a shift from mindless to mindful and from implicit to 308 explicit, we may place Arctic archaeology on a firm intel- lectual foundation, paving the way for non-trivial contri- butions to scientific method and theory. It is apprOpriate to enunciate a revised strategy for the conduct of archaeological research -- one whose roots are in the distant past and which builds on recent discus- sions of archaeological method and theory. The study of archaeology is first and foremost the study of past human behavior, using essentially artifactual evidence. The fundamental research strategy is the partitioning of varia- tions in human behavior, using techniques analogous to sta- tistical techniques for the partitioning of variance (Alker 1965:73-80). Explicitly, variations in human be- havior are seen as the result of multiple causes (sources of variance), and it is our major goal to understand the processes which contribute to such variations. Methodologi- cally, we must return to the method of multiple working hy- potheses, not as it has recently been utilized, but as Chamberlain originally prOposed (1897, reprinted 1965). We must make explicit notice of the multivariate nature of causes of human behavior, and recognize that multiple hypo- theses are not posed as alternates nor are they necessarily mutually exclusive (Cole and Kleindienst 1974:353). Rather, they are prOposed as sources of variance, each of which may contribute to the variations in the observed behavior which is the focus of our research. Our task is to explore the nature of the variations to determine the behavioral 309 processes which affect them. In partitioning the sources of variation in human be- havior, we do not assume the primacy of any particular pro- cess or set of processes. As anthrOpologists, we should not assume the awkward stance of postulating that which we wish to prove. Those who assume a priori that the varia- tions in human behavior which they have Observed (or which they have data for) result from learned normative systems (i.e., culture) are guilty of the tautology that this ex- plicit paradigm attempts to avoid. We should also avoid the misleading belief that one of the hypotheses we might be considering is "true", while the Others are "false". As recent debates in other aspects of anthrOpology have suggested, previously hotly debated com- peting hypotheses are now seen as complementary, both con- tributing to variations in human behaviors (Drucker 1974: 608). Examples of such debates based on the presumed ex- clusiveness of hypotheses include the nature/nuture contro- versy in biology, psychology and anthrOpology. In the evaluation of methods for the analysis of behavioral data, we find again that the postulate of exclusiveness has been recently discarded. The use of the total morphological pattern as a multivariate approach to the study of human and fossil morphology, and the recent use of multivariate statistical techniques for the description and comparison of human morphological data (Lasker 1970) have replaced the 1000 cc. rubicon and other single criterion approaches T‘E 310 to our understanding of human evolution. It is difficult not to fall into the same error of couching this strategy as the true and best approach to our understanding of human behavior. However, I hope that by demonstrating that this approach is holistic and open we can see how other restricted strategies are not as appro- priate, nor as widely applicable as an umbrella under which archaeological research may be conducted. In particular, I wish to avoid the impression that this is a recipe to be followed mindlessly. Such a cook-book approach has unfor- tunately marred some of the recent advocation of statisti- cal methods and random sampling. A random sample of an in- apprOpriate universe will improve neither the data nor the results. The mindless application of explicit methods will not of itself result in increased understanding of human behavioral processes. It is largely for the above reasons that this study was made as explicit as possible, so that the strategic de- cisions made in the course of the research may be Open to question and evaluation, and so that sources of variation in techniques, methods, and results of archaeological re- search may be elucidated. This study did not produce the definitive understand- ing of human behavior within the Arctic Small Tool tradi- tion. Rather, it made a start and pointed the way in which further research should proceed. Much of the present and past research in this area has not been productive of an 311 increased understanding of human behavior and has suffered from the lack of an explicit paradigm. Much of the data gained archaeologically from the Arctic was collected using imprecise field techniques, lumped into categories which reduced our ability to study variations within each site, and published in an incomplete and normative manner, which also reduced the variability available for study. This study demonstrated the utility of an explicit and precise study by develOping and testing models of human be— havior analyzed at several levels of abstraction. This combination of explicit and precise methods and models should raise more questions than it answers, but such a pro- cedure is important because it has not been done before, particularly at several different levels of abstraction. Moreover, if this approach proves useful for future research, it has serious implications for the organization of such research. We will require different data presented in different forms than has previously been the practice. Explicit and precise raw data must be readily available for all interested researchers, and variety in archaeological data must receive as much attention as do norms. I echo Struever's call for a reorganization of the institutional framework in which we can conduct archaeological research (1968:133) to facilitate the coordination of research which an explicit and precise strategy will require. Programs of archaeological publication will require revision, to facili- tate the exchange of needed data and to coordinate the 312 research strategies. In short, I see these pleas for explicitness and pre- cision as leading to the re-focus of archaeological research on the study of human behavior and to the generation of non- trivial hypotheses and laws. It would be difficult and awk- ward to avoid the hortatory fervor which may appear to characterize this study, and I only hope that clarity has not suffered. The Intellectual Environment It is impossible to consider the development of a field of study without considering the intellectual en* vironment in which such deve10pments occurred. In this in- stance, the major environmental influences on the study of the early prehistory of the Eastern American Arctic came from.the "parent discipline", anthrOpology. It is thus ime possible to divorce a study of archaeological theory and method from that of anthropology, and we will benefit from a brief consideration of the interrelations between these disciplines. Recent discussions in archaeology have been preoccu- pied with the questions of the history and status of scien- tific paradigms (Fitting 1973; Leone l972a; Sterud 1973; Clarke l972a; Binford 1972; etc.), following Kuhn's cate- ggorization of a paradigm as a common body of accepted theory and of appropriate methods and techniques of analy- sis (Kuhn 1970). From the onset of such a discussion, we 313 are forced to assess the evidence for the very existence of such a paradigm by looking at the literature to determine if such agreement on method and theory existed or exists, both in anthropology and in archaeology. Beginning with anthrOpology, the recent assessment of Manners and Kaplan (1968) is a fitting example of the state of theory in anthrOpological research. Scattered through the anthrOpological literature are a number of hunches, insights, hypotheses, and generalizations, some tentative and limited, some of broader sc0pe and more generally accepted. But they tend to remain scattered, inchoate, and unre- lated to one another, so that they often get lost or forgotten; and the tendency has been for each generation of anthrOpologists to start outfresh without any very clear sense of what is known about an area of research. (We are tempted here to para- phrase Santayana: Those who have no memory of the history of anthropology are doomed to repeat it.) Among the consequences of this failure is that theory-building in cultural anthrOpology comes to resemble slash-and-burn agriculture as Anthony Wallace has recently noted: "After cultivating a field for a while, the natives move on to a new one and let the bush take over; then they return, slash and burn and raise crops in the old field again" (1966, p.1254). This "slash-and-burn" character of anthropo- logical theory-building also stems from the fail— ure of anthrOpologists to be more self-conscious about the logical properties of theories and about what it means to assert that a theory "explains" a set of phenomena. A more explicit awareness of such issues would, if nothing else, greatly reduce the output of what often passes for explanation in anthrOpology. Here, we believe, anthrOpolgists may learn a great deal from philosophers of science, provided that their learning is somewhat tempered by the knowledge they have of their own discipline. After all, it is the anthrOpologist who knows the significant empirical problems of his discipline and not the logician or the philoSOpher. But there is always the danger that anthropo- logists can become so over-awed by the impressive technical arsenal of the mathematicians, logicians, and philOSOphers that, in an attempt to achieve 314 greater methodological rigor and to do what philo- sophers and mathematical logicians say they ought to be doing, they may unnecessarily constrict the discipline rather than liberate it. We are all for greater logical and methodological sophistica- tion. But if we permit methodology to suggest the problems we deal with rather than allowing problems to determine the methodology, we clearly run the risk of being more precise about a continually narrowing range of cultural issues or phenomena (Manners and Kaplan 1968:11). In their view, there seems to be no generally accepted body of theory and methods that could be called the paradigm of anthrOpology. Theoretical statements are rarely explicit and the recent emphasis on method appears to put the horse before the cart. The recent review of the concept of cu1~ ture by Singer (1968) also points to the lack of agreement on the conception and use of the concept of culture which is generally accepted as the focus of anthrOpological re- search. It is widely accepted that the major contribution of anthropologists to western intellectualism has been this very concept of culture. Because to a great degree, the changing interests and interpretations of archaeologists re- flect the changing perceptions of the concept of culture, it is useful to review briefly trends in this perception. Kroeber and Kluckhohn (1952) reviewed the variety of con- ceptualizations and definitions associated with culture, producing a monumental work which is frequently as con- fusing as it is enlightening. Trends in the use of the concept were seen as extending from.Tyler's definition in behavioral terms to abstractions regarding symbol systems 315 (see also Singer 1968:540). While this process of change in the use of the concept of culture was generally accepted, the value of these changes was debated. Despite Kroeber and Kluckhohn's feeling that 'the greatest advance in contemporary anthropological theory' is the shift from 'concrete-mindedness' to traffic in abstractions, I venture to predict that anthropology will again revert to defining culture in terms of concrete, Objective, observ- able things and events in the external world. I make this prediction with some confidence because it is the procedure in every other science--in all of the more mature sciences, at least--and we believe that cultural anthrOpology will mature some day. We believe that they have expressed a prominent --perhaps the dominant--trend in conceptions of culture during the past twenty-five years, and they have done it effectively and well. This trend is away from the conception of culture as objective, observable things and events in the external world and toward the conception of cul- ture as intangible abstractions. We deplore this trend because we believe that it is a veering away from a point of view, a theoretical stand- point, that has become well-established and has proved itself to be fruitful in the tradition that is science; the subject matter of any science is a class of objectively observable things and events, not abstractions. This shift in concep- tions of culture, will, therefore, only make the achievement of a science of culture more diffi- cult (White 1954:465,467). The picture of past and present anthropological theory does not support the existence of any set of accepted theory and methods, thus it would be hard to reach any sort of agree- ment on the existence of an anthropological paradigm. In the sense of Kuhn, we find ourselves in a preparadigmatic stage in which there are competing theories and concepts, diverse Opinions on what constitute legitimate research 316 endeavors, disagreements on what and how to measure, and no agreement on what constitutes an explanation of anthropolo- gical phenomena. Fitting does not believe that the present state of archaeology is preparadigmatic (l973:6) but I be- lieve his view to be misled by the fragmentation of the disciplines of anthropology and archaeology. Overall com- munications have been reduced and special interest groups have formed in which we are beginning to learn more and more about less and less. In certain areas of anthropolo- gical or archaeological research, there existed sets of ac- cepted theories and possibly paradigms, but these were fre- quently restricted to small groups of researchers working on similar problems or in similar regions. The dissatis- faction with attempts to synthesize the results of archaeo- logical research in North America (Willey 1966; Jennings 1968; see Fitting 1973:2) is perhaps indicative of the ex- tent of the lack of communication among such informal re- search networks. I suggest that the existence of previous paradigms in anthropology and archaeology was more apparent than real, and that our view of the history of anthropology is colored by changes in the patterns of communication and of research endeavors. The recent broadening of communica- tions among these informal research networks has increased the competition among methods and theories and may be seen as either the escalation of the competitions inherent in a preparadigmatic stage or as an attempt at paradigm revolu- tion. Because many discussants of these trends see the 317 explicit use of scientific methods as progress (see Watson et a1 1971; Martin 1971; Johnson 1972; Thompson 1972; Clarke l972a; Sterud 1973; Fritz and Plog 1970; Stickel and Chartkoff 1973; etc.) and because there is essential agreement that we are going through (or have gone through) a major revolution in archaeology involving the explicit use of a scientific approach, I suggest that what we are seeing is the culmination of a competition among theories and methods which may result in the final acceptance of a paradigm for archaeology in North America at least. This acceptance is made more difficult because anthrOpology does not itself have a paradigm (see Chaney 1972:996-997; Leone l972a:24). The current debate in archaeology and anthrOpology was anticipated by White (see above) as was the difficulty in achieving a science of culture. By and large, the "New Archaeology" is a concern for technical rigor in data gathering, for clear statement Of methods and assumptions, for the explicit utilization of inductive and deductive hypothesizing, for the utilization of methods of systems and ecological analyses, and, perhaps most importantly, for the focus on the study of human be- havior (Binford 1972:132; Taylor 1968:108: Deetz 1968; Leone l972a:23; MacNeish 1974:463; etc.). Viewing our conceptual system as a system of con- straints on our scientific research, it becomes apparent that the concept of culture as some sort of abstraction from.behavior constrains our use of the terms culture, T‘: 318 cultural system, and cultural sub-system (see for example, Taylor 1968:108; Binford 1962; White 1954). While Binford's early statements on method and theory were significant ap- praisals of archaeological and anthrOpological concepts, he has yet to articulate a well-integrated paradigm to compete with his version of traditional archaeology. To Binford, the traditional archaeologist is concerned with the study of culture, first and foremost. Variations in artifactual data are seen as stemming from variations in culture; or more simply, artifact sets are cultures. It is this aspect of traditional archaeology that has come into question. As anthrOpologists are undecided re- garding the definition and use of the concept of culture and of cultures, so their constituent disciplines that seek to operationalize their research based on an uncertain paradigm create similar uncertainty (see, for example, Wax 1973:167-168 regarding the inapprOpriateness of the concept of cultural pluralism). It is this uncertainty regarding the use of the con- cepts of culture presently under discussion and debate that is symptomatic of the present paradigmatic crisis. The present criticism of the use of the term culture as an ab- straction is much the same as the earlier criticism of the use of the concept of "artifact type" and the entire norma- tive typological approach as it has been practiced in archaeology (see Hill and Evans 1973; Kehoe and Kehoe 1973; Binford 1965). 319 The conjunctive force uniting the "new archaeologists" has been largely their criticism of the "Old archaeology." During the middle 1960's, many archaeologists adopted a "show me" stance, in which they held judgment of the methods of the proponents of the new archaeology in abey- ance until they had produced something concrete to evaluate (Taylor 1972:30). The studies of Martin §E_§l. (1962,1964), Hill (1970), Longacre (1970) and others provided grist for the mill of criticism while the delay in formidable criti- cism increased the influence of their ideas on that genera- tion of archaeologists trained in the 1960's and early seventies. There remains one remarkable failure of the new archaeologists. This is the failure to develOp an explicit statement of a paradigm to compete with that of the tradi- tionalists. This failure is remarkable for two reasons. First, Binford and his colleagues have had at least ten years of intellectual growth and exposure in which to formu- late such a paradigm. Second, the competition of theories and models necessary for a scientific revolution demands the formulation of some sort of paradigm, even as a straw man, yet it has not been forthcoming. Instead, what we have seen is the suggestion of ex- plicit methodologies (Watson et 31. 1971), utilization of results of other models such as systems theory or ecology (Leone l972a:23; Clarke l972a) and most recently the focus on the study of behavior (Deetz 1967:105; Binford 1972:132) 320 <31: activities (Redman l973a:717), but no one has yet pre- semnted a consistent and explicit paradigm. The journals lLave presented numerous discussions of methodology, most of Vehich.appear to offer the hope of impressive theoretical «contributions. Is methodology all there is to the new zirchaeology? I think not, but the reviewer of the litera- ‘ture would be hard pressed to come to a different conclu- sion, especially given the composition of recent monumental ‘works in this area (Clarke 1972b; Leone 1972b; Renfrew 1973). If such a paradigm of the new archaeologists exists, *we might expect that these volumes would contain it, but they do not. I believe the message from the rhetoric surrounding -the present paradigmatic crisis to be simple. We should ‘take as our goal the study of human behavior, perhaps fo- cnising on that portion of human behavior which is learned (i.e., culture). Our general approach should be scientific, 1lsing inductive methods where our paramount concern is with the data available, and using deductive methods from general systems theory, from ecological theory, from anthropological theory, or from other logically relevant bodies of theory. (hir research techniques should be as SOphisticated as Exassible, consistent with the nature of our data, their \nariety, diversity and frequency, and should be guided by tflne failures and successes in the techniques of other sciences. To accomplish this, we should make our assump- ‘tions, biases, premises, postulates, theories and data 321 collection methods and techniques as explicit and precise as necessary to allow the replication of our results in so far as possible. In a real sense, the strategy which I prOpose forces us to study Observed variations and patterns in artifacts without a priori assumptions about the nature of such patterns resulting from "cultural" behavior, as it forces the consideration of alternatives (multiple working hypo- theses) and facilitates the partitioning of numerous sources of archaeological variation. If we wish to focus our attention on cultural variation, then we must have con- sidered and controlled for other sources of variation. To do this, we must have a strategy which facilitates this re- search process, and it is just such a strategy which I propose. Chapter 9 Recapitulation Variations in models of archaeological data have re- sulted from differences in the research activities of archaeologists as well as from differences in archaeologi- cal data. Differential archaeological field techniques and research strategies have contributed to sample bias (Knuth 1967; McGhee l973a; and Taylor 1968a) which we have not al- ways considered. Perhaps a more pervasive source of varia- tion has been the normative approach to data sets which have reduced the study of the within-sample variety (Anderson 1970; Knuth 1967; Maxwell l973a; and Taylor 1968a). While this approach may be appropriate with some data-sets and some problems, it should not be routinely ap- plied to all problems, nor should it be allowed to dominate the selection of field techniques or data categorization and depiction processes. The paradigm of archaeological interpretation which assumes that the variation in archaeological data stems largely from cultural variation has led to the interpreta- tion of heterogeneous samples as the result of mixtures from two normative (cultural) systems (Harp 1958; Larsen 322 323 and Meldgaard 1958; McGhee 1973, 1974), and a tendency to minimize variation within cultural units. By focusing on the study of normative systems, we have largely ignored the study of behavioral variability within such systems. Our basic strategy has involved the generation of com- peting models (Often implicit), when it is becoming in- creasingly apparent that there are a number of sources of variation to be modelled, probably initially separately, and ultimately collectively in a very complex system Of re- lated models. The process Of model building affects the selection of methods and techniques, especially within a deductive or a problem-oriented approach to archaeological research. These selective processes will yield data amenable to certain analytical techniques and not to others. Since we cannot avoid such selection, it is essen- tial that we specify the techniques and methods used and make them as explicit and precise as possible to allow the fullest reconstruction of the data base from its descrip- tive parameters. This necessitates a change in approach to emphasize the dispersion parameters as well as the modal (normative) ones, particularly in "site reports" and recon- naissances in which the data assume paramount importance. As Arctic archaeology has developed, it has passed through a number of stages of problem orientation. These include an initial emphasis on ”what" and "where" kinds of questions, with passing interest in the "who" (both bio- logical, or racial, and ethnic). As the answers to these 324 questions became available, increased emphasis was placed on the "when", with numerous techniques being applied (cross-dating by sequences from other disciplines, seria- tions, etc.). With the advent of radiocarbon dating, the answers to these questions become more easily obtained and the space-time framework of human occupations could be generally established. Prior to this time, one major tactic was to postulate away the complexities in the data, or to Operate at such a high level of abstraction that the control of time was only generally a problem. With the answers to what, where, and when within reach, our attention has turned to strategies aimed at answering questions of "why". This search for ex- planations had been largely speculation based on implicit assumptions, but this is changing in archaeology in general, and in the Arctic in particular. To utilize a scientific approach to explanation, we re- quire explicit and precise answers to the questions of what, where and when. For this reason, our radiocarbon chronolo- gy requires careful attention and we can no longer play it "fast and loose" if we wish our search for explanations to be productive. For this reason, the chronology develOped explicitly in Chapter 2 is as precise as possible at this time and can serve as a framework for this research. Those areas of imprecision have been isolated and discussed and are thus Open to the scrutiny of others who might wish to criticize the use to which such a framework has been put. 325 It is readily apparent that the prehistoric human Occu— pation of Arctic North America occurred in an environmental setting marked by a relatively simple ecosystem of great specific variability, through both Space and time. As we wish to make our study of the adaptive relations between the behavioral systems of human foragers and their environ— ment mbre specific, we will need more specific and detailed reconstructions of paleoenvironmental change. However, the complex ecological relations in the Eastern Arctic are only generally known and are subject to differential selection and interpretation. In order to avoid the selection of in- terpretations and data which suit our predelictions, we must treat our selection process precisely and explicitly. Such treatment has resulted in the definition of an ecological core area and variegations in significant en- vironmental variables through space and time. At the risk of over-generalizing, the core area has been realtively un— changed by environmental fluctuations which have had greater impact in peripheral regions. Through time, its occupants have had consistent access to a variety of re- sources, subject to greater variability in peripheral areas. Peripheral regions have had fewer alternate resources available and may have had subsistence patterns focused on one or more resources. Thus, the occupants of the western tundra may have exploited land mammals of the tundra to a greater extent than the eastern sea mammal hunters. The northern islands and North Greenland may have shown greater 326 dependence on musk ox and fishing, and so forth. These regional variations on an overall subsistence capability may have been forced adaptations to environmental changes or may represent attempts to optimize subsistence strate- gies in areas with different resources. In order to speci- fy the adaptive changes in subsistence, we need to pay greater attention to the specific evidence for subsistence activities (i.e., faunal remains) and changes in subsis- tence strategies. The dynamics of Eastern Arctic environments through time have been the subject of considerable attention, and the sequence of changes which I have specified elsewhere continues to be substantiated by additional research. During the Pre-Dorset occupation of the Eastern Arctic, we can reconstruct general environmental changes with some reliability (see Table 8), while the specific local impacts are only occasionally assessable. However, if we wish to explain the variable adaptive relations between human be- haviors and their environment, we must treat the environ- mental data as precisely and explicitly as possible. Their use as input into the ecological systems model is ample demonstration of their utility and for the need for care in their depiction and utilization. Partitioning the sources of variation in the data for prehistoric human behaviors is a complex and difficult task. When we were asking relatively simple questions of these data, such variations were of less consequence than they 327 are in our present efforts at explanation. I have attemp- ted to list the potentially significant sources of varia- tion as a first step in sensitizing us to their frequency and complexity. I have suggested several specific strate— gies aimed at controlling some of these sources of variance on the way to studying others. The first step in controlling these variables is to recognize their potential as sources of variance. The second step is to devise research strategies which allow the study Of the sources of variance in which we are inter- ested. It is essential that our efforts at studying cultural processes and explaining human behavior be based on ex- plicit and precise research strategies, techniques, and methods. Because of the complexity of these sources of variance, our first steps will be fundamental, tentative, and possibly even faulty, but it is essential that we begin our research on a scientifically valid series of strategies which some may see as a paradigm of scientific archaeology. Variations in the evidence for Arctic Small Tool tradi- tion structures have posed a considerable interpretational problem in the Eastern Arctic. While some structures have clearly delineated outlines and patterned internal features, others were rock and artifact clusters (Occasionally with central features, such as hearths and mid-passages) without defined outlines. These had been variously interpreted as some kind of dwelling (Taylor 1968a:14), possible of snow 328 houses (Maxwell 1967; 1973a:303-304,310). The data avail- able were not subject to testing because specific data re- garding the relations among the artifacts and rocks had not been generally recorded or depicted. By utilizing precise field techniques, we were able to study the relations among the artifacts and rocks in such clusters and to utilize a novel and innovative data struc- ture for the testing of deductively generated hypotheses. This previously untried combination of precise data and ex- plicit scientific methodology led to the ability to resolve the engima of these rock and artifact clusters and de- velOped a tentative model of tent structures and behaviors at the Closure site. This model may now be tested on rele- vant data sets in the Eastern Arctic and the techniques de- veloped here may be applied in the resolution of similar interpretational problems wherever they occur. The early sites in the Eastern Arctic comprise a data mosaic which has been subject to different interpretations. The diversity in this data mosaic has several sources of variance, which include: sampling error; sampling bias; differential data depiction and lumping; different environ- mental settings with differential resources; and excavation and preliminary interpretation by different archaeologists. Our perception of this mosaic is also influenced by a reliance on relatively rare artifacts (curated) as diagnos- tic of particular cultural systems. Such a normative ap- proach focuses our attention on culture as a modal system 329 and inhibits our ability to study diversity within the sys- tem. Again, we find ourselves trying to understand a come plex interactive system by studying the differences between its nodes, seen normatively (through the use of diagnostic lists, guide fossils, means, etc.). By making explicit our awareness of culture as an adaptive system of potential behaviors which may be uti-. lized differentially in response to particular characteris- tics of the environment, we are better able to examine the diversity in this mosaic. By modeling the behavioral pro- cesses which contribute to the observed variance, we can study the complex dynamics of human populations, recog- nizing that the evidence we have for changes in behaviors may result from the adaptive competition among potential be- haviors within the same cultural system (adapting to en- vironments changing through space and time--evolving, mi- grating, fluctuating through seasons, etc.). For this reason, we may need to focus our study on discontinuities in the data mosaic, partitioning sources of variance as they reflect processes of our acquisition of archaeological knowledge or processes of the behavior of the prehistoric inhabitants of this region. The use of a deductively derived model of the dynamics of human popula- tions which focuses on human behavioral processes as they affect human technology is one step in this strategy. The model of the Arctic Small Tool horizon accurately predicts variation in several dimensions of technology by 330 explaining this variation as resulting from the dispersal vector of an adaptive and diversified cultural system. There is clinal variation in expedient artifacts only loosely constrained by functional requirements. There is increased variety in curated artifacts within the horizon, with similar artifacts found across large areas. Such homogeneity among sites in curated artifacts is also pre- dicted by the model. Such an interpretation which in- volves the mixing of behaviors (and not "cultures") is more parsimonious than one which views these similarities as the mixture of the products of two separate cultural systems, in which the failure to find "pure" manifestations is the result of sampling error and negative evidence. The search for such cultural purity is itself reflective of a norma— tive bias, and of a paradigm which postulates cultural dif- ference as the major source of technological variegation. The strategy reflected by the development of the above model will necessitate numerous complementary models if we are to be able to predict and explain the variety in the Arctic Small Tool horizon. By making explicit our focus on the behavioral processes which contribute to such variance and by modeling the relations among the nodes in this data mosaic (as well as the nodes themselves), we will move be- yond the use of implicit models and imprecise concepts such as migration and diffusion to a greater understanding of complex human behaviors. There have been a number of conceptual frameworks 331 presented as models of the Pre-Dorset deve10pment in the Eastern Arctic, none of which provides an adequate explana- tion. This model depicts the macro-fragmentation of an Arctic Small Tool horizon base into regionally distinct variants as resulting from a reduction in social scale caused by a major environmental change. It is apparent that the alternate sources of variation were differential invention and differential adaptation to environmental dif- ferences (in the macro-scale), and that the model explains the maintenance of these behavioral variations through. time. Such a model is markedly similar in structure and con- tent to those demonstrated by McKennan (1969:99) for Athapaskan linguistic variation and by Meiklejohn (1974: 135,138) for population genetics variation in band socie- ties. Both of these have focused on the changing relations among spatially distributed nodes (or foci). All have been the result of the study Of human behaviors in band socie- ties. Inasmuch as all Of these phenomena being modeled (technology, language, and genetics) are both evidence of and are influenced by human behavioral changes and changes in communications patterns among nodes, a model which re- lates evidence of behavioral change to changing patterns of communications and changing environments can have wide- spread utility. This explicit focus on relations among social entities within a widespread differentiated cultural system will require explicit and precise archaeological 332 data in order to facilitate the develOpment of additional models Of human behavioral processes. By proceeding with such a strategy as outlined and demonstrated in this thesis, we will place the study of human behavioral processes in a changing environmental setting on a level of sophistication which will allow the integration of models from archaeolo- gical studies with models of human behavioral processes from ethnographic studies (such as Hill 1970 and Longacre 1970) and with models of develOpmental change (Chance 1968 and Pothier 1968). We will then be in a position to pro- vide the time depth and ecological dimension so lacking in present efforts at modeling behavioral and cultural change, and will demonstrate the utility of archaeological research in develOping a more SOphisticated and widely applicable understanding of human behavioral processes. Such models must be based on explicit and precise methodologies such as I have prOposed herein. Recommendations for Further Research Perhaps the most pressing need for research in the Eastern Arctic is the establishment of seasonal variegation in any one archaeologically-known behavioral system at any one point in time. This will necessitate a well established chronology with established contemporaneity, as well as analysable collections of faunal and floral materials to establish the sequence of seasonal occupation. The variations in the data depicted by such a strategy could 333 then be used as a model for studying the variations in other data less well—controlled and could be used for the further study of within-site behaviors and division of labor. While these studies could be conducted in several regions, I know of no concerted effort to study such season- al variation within one locale or region. This would in- volve the archaeological survey of a variety of potential land-use patterns in a variety of accessible habitats. From studies such as these, we could refine models of behavioral change and specify the adaptive responses to changes in environments. Changes in the scheduling of sub- sistence activities could be specified, and models of adap- tive responses to environmental changes could be tested. It is also time for increased theoretical attention to population dispersals (migrations), in order to develop models comparable in specificity to cultural traditions. Such horizons of archaeological data have received little attention, in spite of their frequency. 0f the three most extensive and rapid archaeological horizons presently known in North America, two are found in the Arctic. These are the Arctic Small Tool horizon and the Thule horizon. The above model could be tested on the Thule horizon, as well as on the Paleo-Indian horizon in southern North America. Ecological systems models such as I developed above could be tested on additional data in which environmental changes may be significant inputs. The most appropriate data for further testing would be the subsequent prehistory 334 of the Eastern Arctic, in which these same processes could be expected to be significant sources of variation. Because much of the above discussion has considered (albeit briefly and with a certain lack of specificity) the partitioning of sources of variance in human behaviors known archaeologically to be the major research strategy of archaeology, it would be useful if an explicit attempt were made to study both the process of archaeological research and the interpretation of archaeological data with the ex- press intent of partitioning the variance. This would re- quire a region in which the archaeology was well known and in which the archaeological data were readily available. It would also require the use of statistical and mathemati- cal models, as well as specific knowledge of paleoenviron- ments and ecological relations. This seems a large study, utilizing various specialists comparable in sc0pe to MaCNeish's research efforts in the Tehuacan Valley. It would also be appropriate to study the deve10pments in the Western Arctic following the Arctic Small Tool hori- zon, in particular to see if the models developed above ap- ply there. The deve10pment of the late Arctic Small Tool tradition at Punyik Point and Walakpa could be tested to see if growth (as defined above) occurred. Preliminary in- terpretations by Irving regarding lithic specialization and the Arctic Small Tool technique of lithic flaking suggest that such research might be very fruitful in this area. There are undoubtedly other adaptive, ecological, 335 cultural or behavioral processes which contribute to the variations in the Arctic Small Tool horizon and in the sub- sequent Pre-Dorset period. With increased attention to precise field techniques and ecological analyses, we may be able to specify and model such additional sources of variance. Alternatives to be considered include: the im- portance of raw material size for lithics manufacture (par- ticularly, whether variation in size of finished products are related to raw material size); the variations in flint tools necessitated by differences in hafting materials or raw materials on which they are used (particularly, the differences between the use of ivory and the use of antler, both as hafting material and as raw material); and differ- ences in projectile points and end blades (including har- poons) resulting from hunting practices and other subsis- tence activities. I hope that the methodological contributions of this thesis and the models presented herein may result in the resolution of previously unresolved problems and may place Arctic archaeology in a firm position to contribute to our developing understanding of human behavior. B IBLIOGRAPHY BIBLIOGRAPHY Ahlmann, H. W. 1953 Glacier Variations and Climatic Fluctuations. American Geographical Society Series No. 3. Alexander, Herbert L., Jr. 1969 Prehistory of the Central Brooks Range--An Archaeological Analysis. Ph.D. dissertation. University of Oregon. Eugene. Alker, Hayward R., Jr. 1965 Mathematics and Politics. New York: The Macmillan Company. Allen, Francis R. 1971 Socio-Cultural Dynamics. New York: The Macmillan Company. Alverson, Dayton L. and Norman J. Wilimovsky 1966 Fisheries Investigations of the Southeastern Chukchi Sea. In Environment of the Cape Thompson Region, Alaska. Norman J. Wilimovsky and John N. Wolfe, eds. U.S. Atomic Energy Commission Publication PNE-48l. Washington, D.C.: U. S. Govt. Printing. pp. 843-860. Anderson, Douglas D. l968a A Stone Age Campsite at the Gateway to America. Scientific American 218(6):24-33. l968b The American Paleo-Arctic Tradition: New Evidence for Early Man in Alaska. Paper for the Society for American Archaeology. Santa Fe. 1970 Microblade Traditions in Northwestern Alaska. Arctic Anthropology 7(2):2-16. 1972 An Archaeological Survey of the Noatak Drainage, Alaska. Arctic Anthropology 9(1):66-117. 336 337 Andrews, J. T. 1972 Recent and Fossil Growth Rates of Marine Bi- valves, Canadian Arctic, and Late Quaternary Arctic Marine Environments. Palaeogeography, Palaeoclimatology, Palaeoecology 11(1972): 157-176. Andrews, J. T., RObert McGhee and Lorna McKenzie-Pollock 1971 Comparisons of Elevations of Archaeological Sites and Calculated Sea Levels in Arctic Canada. Arctic 24(3):210-228. Andrews, John T. and Jack D. Ives 1972 Late— and Postglacial Events (10,000 B.P.) in the Eastern Canadian Arctic and Particular Reference to the Cockburn Moraine and Break-up of the Laurentide Ice Sheet. In Climatic Changes in Arctic Areas During the Last Ten Thousand Years. Y. Vasari, H. Hyvarinen and S. Hicks, eds. University of Oulu Symposium ‘Series A 3(1):l49-l76. Oulu, Finland. Arutiunov, S. and D. Sergeev 1973 Stability and Adaptability in the Evolution of Ancient Eskimo Hunting Tools. Paper Presented. to the IXth International Congress of Anthro- pological and Ethnological Sciences. Chicago. Ashby, W. Ross 1963 An Introduction to Cybernetics. New York: John Wiley. Ascher, Robert 1968 Time's Arrow and the Archaeology of a Contem— porary Community. In, Settlement Archaeology. K. C. Chang, ed. Palo Alto: National Press. pp. 43-52. Balikci, Asen 1970 The Netsilik Eskimo. Garden City: The Natural History Press. Bandi, Hans-Georg 1963 The Burins in the Eskimo Area. In_ Early Man in the Western Arctic, A Symposium. Anthropo- logical Papers of the University of Alaska 10 (3):l9-28. Barry, Roger G. 1967 Seasonal Location of the Arctic Front Over North America. Geographical Bulletin 9(2): 79-95. 338 Barry, Roger G. and R. J. Chorley 1971 Atmosphere, Weather and Climate. London: Methuen & CO., Ltd. Barry, Roger G. and Allen H. Perry 1973 Synoptic Climatology: Methods and Applications. London: Methuen & CO., Ltd. Bartley, D. D. and Barry Matthews 1969 A Palaeobotanical Investigation of Post Glacial Deposits in the Sugluk Area of Northern Ungava (Quebec, Canada). Review of Palaeobotany and Palynology 9:45-61. Bee, James W. and E. Raymond Hall 1956 Mammals of Northern Alaska. University of Kansas Museum.of Natural History Miscellaneous Publication No. 8. Lawrence: Allen Press. Bee, Robert L. 1974 Patterns and Processes: An Introduction to AnthrOpological Strategies for the Study of Sociocultural Change. New York: The Free Press. Berry, Brian J. L. 1962 Sampling, Coding, and Storing Flood Plain Data. U. S. Dept. of Agriculture Handbook No. 237. Washington, D.C.: U. S. Govt. Printing. Binford, Lewis R. 1962 Archaeology as Anthropology. American Antiquity 28:217-225. 1965 Archaeological Systematics and the Study of Culture Process. American Antiquity 31:203-210. 1972 Contemporary Model Building: Paradigms and the Current State of Palaeolithic Research. In MOdels in Archaeology. David L. Clarke, Ed. London: Methuen & CO., Ltd. pp. 227-254. 1973 Interassemblage variability--the MOusterian and the 'functional' argument. In The Explanation of Culture Change: Models in Prehistory. Colin Zenfrew, ed. London: Duckworth. pp. 227-25 . 339 Binford, Sally and Lewis R. Binford 1966 A Preliminary Analysis of Functional Varia- bility in the Mousterian of Levallois Facies. American AnthrOpologist 68(2):138-195. 1969 Stone Tools and Human Behavior. Scientific American 220(4):70-84. Birket-Smith, Kaj 1928 The Greenlanders of the Present Day. In_ Greenland, Volume II--The Past and Present Populations of Greenland. ‘M. Vahl, et. al., eds. Copenhagen: C. A. Reitzel, Publishers. pp. 1-208. Bissett, Don 1970 Northern Baffin Island: An Area Economic Sur- vey. Department of Indian Affairs and Northern Development. Information Canada Publication. Ottawa: Queen's Printer. Blake, Weston, Jr. 1970 Studies of Glacial History of Arctic Canada. I. Pumice, Radiocarbon Dates and Differential Postglacial Uplift in the Eastern Queen — Elizabeth Islands. Canadian Journal of Earth i Sciences 7(2):634-664. ' 1972 Climatic Implications of Radiocarbon Dated Driftwood in the Queen Elizabeth Islands, 1 Arctic Canada. In. Climatic Changes in Arctic ' Areas During the Last Ten Thousand Years. Y. Vasari, H. Hyvarinen and S. Hicks, eds. Uni- versity of Oulu Symposium Series A 3(1):77- 104. Oulu, Finland. Boas, Franz 1964 The Central Eskimos. Lincoln: University of Nebraska Press. Bockstoce, John 1973 A Prehistoric Population Change in the Bering Strait Region. Polar Record 16(105):793-803. Bray, J. R. 1965 Forest Growth and Glacier Chronology in Northwest North America in Relation to Solar Activity. Nature 205:440-443. 1966 Similarity of Tree Growth in Northern Scandi- navia, Polar Urals and the Canadian Rockies. Journal of Glaciology 6(44):321-322. 1968 Glaciation and Solar Activity Since the Fifth 340 Century BC and the Solar Cycle. Nature 200: 672-674. Bray, J. R. and G. J. Struick 1963 Forest Growth and Glacial Chronology in Eastern British Columbia, and Their Relation to Recent Climatic Trends. Canadian Journal of Botany 41(8):1245-127l. Breverton, R. J. H. and A. J. Lee 1965 Hydrographic Fluctuations in the North At- lantic Ocean and Some Biological Consequences. In, The Biological Significance of Climatic Changes in Britain. C. G. Johnson and L. P. Smith, eds. New York & London: Academic Press. pp. 79-109. Bryson, Reid A. 1966 Air Masses, Streamlines, and the Boreal Forest. Geographical Bulletin 8(3):228-269. 1974 A Perspective on Climatic Change. Science 184 (4138):753-760. Bryson, Reid A. and W. M. Wendland 1967a Tentative Climatic Patterns for Some Late- Glacial and Post-Glacial Episodes in Central North America. University of Wisconsin De- partment of Meteorology Technical Report No. 34. Madison. l967b Tentative Climatic Patterns for Some Late- Glacial and Post-Glacial Episodes in Central North America. In Life, Land and Water. ‘William L. Mayer:0akes, ed. Proceedings of the 1966 Conference on the Environmental Studies of the Glacial Lake Agassiz Region. Occasional Paper No. l of the University of Manitoba. Winnipeg. pp. 271-298. Bryson, Reid A., W. M. Wendland, J. D. Ives and J. T. Andrews 1969 Radiocarbon Isochrones on the Disintegration of the Laurentide Ice Sheet. Arctic and Alpine Research l(l):l-14. Bryson, Reid.A., David A. Baerreis and Wayne Wendland 1970 The Character of Late-Glacial and Post-Glacial Climatic Changes. In] Pleistocene and Recent Environments of the Central Great Plains. W. Dort and J. K. Jones, eds. University of Kansas Department of Geology Special Publica- tion No. 3. Lawrence: Allen Press. pp. 53— 74. 341 Buckley, Walter 1968 Society as a Complex Adaptive System. In Modern Systems Research for the Behavioral Scientist. Walter Buckley, ed. Chicago: Aldine. pp. 490-513. . Campbell, John M. 1962a Cultural Succession at Anaktuvuk Pass, Arctic Alaska. In Prehistoric Cultural Relations Between the Arctic and Temperate Zones of North America. John M. Campbell, ed. pp. 39- 54. Arctic Institute of North America, Tech- nical Paper No. 11. Montreal. Campbell, John M., Ed. 1962b Prehistoric Cultural Relations Between the Arctic and Temperate Zones of North America. Arctic Institute of North America, Technical Paper NO. 11. Montreal. Campbell, John M. 1968 Territoriality Among Ancient Hunters: Inter- pretations from.Ethnography and Nature. Ig_ Anthropological Archaeology in the Americas. Anthropological Society of Washington Lecture Series, 1966-67. Washington, D. C.: Anthro- pological Society of Washington. pp. 1-21. Chamberlain, Thomas C. 1965 The Method of Multiple Working Hypotheses. Science l48(367l):754-759. Chance, Norman A. 1968 The Cree Developmental Change Project: An In- troduction. In Conflict in Culture: Problems of Developmental Change Among the Cree. Norman A. Chance, ed. pp. 1-9. Canadian Research Centre for Anthropology. Ottawa: Saint Paul University. Chaney, R. P. 1972 Scientific Inquiry and MOdels of Socio— Cultural Data Patterning: An Epilogue. In MOdels in Archaeology. David L. Clarke, —d. London: Methuen & CO., Ltd. pp. 991-1031. Chang, K. 0., Ed. 1968 Settlement Archaeology. Palo Alto: National Press Books. 342 Chorley, Richard J. 1962 Geomorphology and General Systems Theory. Govt. Theoretical Papers in the Hydrologic and Geomorphic Sciences. Geological Survey Pro- fessional Paper SOO-B. Washington, D.C.: U. S. Govt. Printing. Clarke, David L. 1968 Analytical Archaeology. London: Methuen & CO., Ltd. 1972a MOdels and Paradigms in Contemporary Archaeo- logy. l2 ‘Models in Archaeology. David L. Clarke, ed. London: ‘Methuen & CO., Ltd., pp. 1-60. Clarke, David L., Ed. l972b Models in Archaeology. London: Methuen & CO., Ltd. Cleland, Charles E. 1966 The Prehistoric Animal Ecology and Ethnozoolo- by of the Upper Great Lakes Region. Museum of Anthropology Paper No. 29. Ann Arbor:: University of Michigan. Cochran, William G. 1953 Sampling Techniques. New York: Wiley & Sons, Inc. Cole, G. H. and M. R. Kleindienst 1974 Further Reflections on the Isimila Acheulian. Quaternary Research 4(3):346-365. Collins, Henry B., Jr. 1951 The Origin and Antiquity of the Eskimo. Smithsonian Annual Report for 1950. pp. 423- 467. Washington, D. C.: Smithsonian Press. 1953a Recent Developments in the Dorset Culture Area. Memoirs of the Society of American Archaeology 9:32-39. 1953b Radiocarbon Dating in the Arctic. American Antiquity 18:197-213. 1954a "Arctic Area". Program.of the History of America. Comission de Historia l(2):152. Mexico City. 1954b Archaeological Research in the North American Arctic. Arctic 7(3 and 4):296-306. 1962 Corte, Arturo 1963 Daniels, S. G. 1972 Dansgaard, W., 1969 Dansgaard, W., 1970 1971 Deetz, James 1965 1967 1968 Dekin, Albert 1968 343 Bering Strait to Greenland. In, Prehistoric Cultural Relations Between the Arctic and Temperate Zones of North America. John M. Campbell, ed. pp. 126-139. Arctic Institute of North America, Technical Paper No. 11. MOntreal. E. Particle Sorting by Repeated Freezing and Thawing. Science l42(359l):499-501. H. Research Design Models. 13. Models in Archaeology. David L. Clarke, ed. London: Methuen & CO., Ltd. pp. 201-230. S. Johnsen, J. Mo1ler and C. Langway Jr. One Thousand Centuries of Climatic Record from.Camp Century on the Greenland Ice Sheet. Science 166(3903):377-38l. S. Johnsen, H. Clausen and C. Langway, Jr. Ice Cores and Paleoclimatology. In_ Radio- carbon Variations and Absolute Chronology. Ingrid Olsson, ed. New York: Wiley- Interscience. pp. 337-348. Climatic Record Revealed by the Camp Century Ice Core. 12. The Late Cenozoic Glacial Ages. Karl Turekian, ed. New Haven: Yale Univer- sity Press. pp. 37-56. The Dynamics of Stylistic Change in Arikara Ceramics. Urbana: University of Illinois Press. Invitation to Archaeology. Garden City, N.Y.: Natural History Press. Cultural Patterning of Behavior as Reflected by Archaeological Materials. l2. Settlement Archaeology. K. C. Chang, ed. Palo Alto: National Press Books. pp. 31-42. A., Jr. Hypotheses on Cultural Interaction in the Eastern American Arctic. Paper Presented to the Michigan Academy of Science, Arts and Letters, 22 March 1968. Grand Valley State College, Grand Valley, Michigan. 1969 1970 l972a 1972b l972c l973a 1973b 1973c 1974 344 Paleo-Climate and Prehistoric Cultural Inter- action in the Eastern Arctic. Paper Presented at the 34th Annual Meeting of the Society for American Archaeology, May 1969. Milwaukee. Paleo-Climate and Paleo-Ecology Of the Eastern North American Arctic During Its Occupancy by Man (2500 BC to date). Paper Presented at the 3rd Annual Meeting of the Canadian Archaeolo- gical Association, March 1970. Ottawa. Climate and Culture in Greenlandic Prehistory: An Explicitly Scientific Approach. Paper Pre- sented at the Fifth Annual Meeting of the Canadian Archaeological Association. Sympo- sium on Prehistoric Climatic and Cultural Change in Eastern Canada and Adjacent Regions. St. John's: Memorial University Of Newfound- land. Mimeographed in three parts. Climatic Change and Cultural Change: A Correlative Study from Eastern Arctic Pre- history. Polar Notes 12:11-31. Review of Fort Chimo and Payne Lake, Ungava, ArchaeoIOgy, 1965.; Archaeological Discover- ies, Payne Bay Region, Ungava, 1966.; Archaeological Findings, Gyrfalcon to Eider Islands, Ungava, 1968., all by Thomas E. Lee. American Anthropologist 74(6):1501-1504. The Arctic. In. The DeveIOpment of North American Archaeology. James E. Fitting, ed. Garden City, New York: Anchor Press/ Doubleday. pp. 14-48. Preliminary Paper and Discussion. School of American Research Advanced Seminar on Pre- Dorset--Dorset Problems. Santa Fe. Reply to Lee. American Anthropologist 75(6): 2040-2041. The Walrus and the Polished Burin: A Possible Dynamic Duo from Eastern Arctic Prehistory. Paper Presented at the 39th Annual Meeting of the Society for American Archaeology, 3 May. Washington, D.C. Denton, George H. and Wibjorn Karlen 1973 Holocene Climatic Variations--Their Pattern and Possible Cause. Quaternary Research 3(2): 155-205. 345 Dice, Lee R. 1943 Biotic Provinces of North America. Ann Arbor: University of Michigan Press Dillon, L. S. 1970 Speciation and Changing Environment. American Zoologist 10:27-39. Drucker, Charles B. 1974 Review of. Social Relations in a Philippine Market: Self Interest and Subjectivity, by William G. Davis. American Anthropologist 76(3):608-609. Dumond, Don E. 1964 A Note on the Prehistory of Southwestern Alaska. AnthrOpological Papers of the Univer- sity of Alaska 12(1):33-45. 1965 On Eskaleutian Linguistics, Archaeology and Prehistory. American Anthropologist 67:1231- 1257. 1969 Prehistoric Cultural Contacts in Southwestern Alaska. Science 166:1108-1115. 1971 A Summary of Archaeology in the Katmai Region, Southwestern Alaska. Anthropological Papers of the University of Oregon NO. 2. 1974a Some uses of R-MOde Analysis in Archaeology. American Antiquity 39(2):253-270. l974b Remarks on the Prehistory of the North Pacific: To Lump or Not to Lump. IE. Inter- national Conference on the Prehistory and Paleoecology of Western North American Arctic and Subarctic. Scott Raymond and Peter Schledermann, eds. pp. 47-56. Calgary, Alberta: University of Calgary Archaeological Association. Dunbar, Maxwell J. 1951 Eastern Arctic Waters. The Fisheries Research Board of Canada Bulletin No. 88. Ottawa: King's Printer. 1968 Ecological DevelOpment in Polar Regions: A Study in Evolution. Englewood Cliffs: Prentice Hall, Inc. 346 Durkheim, Emile 1933 On the Division Of Labor in Society. New York: The Macmillan Company. Trans. by George Simpson. (Original publication 1893). Dzerdzeevskii, B. L. and V. S. Sergin 1972 A Procedure for Studying Climatic Fluctua- tions on Different Time Scales. In, Climatic Changes in Arctic Areas During the Last Ten Thousand Years. Y. Vasari, H. Hyvarinen and S. Hicks, eds. University of Oulu Symposium Series A 3(1):427-454. Oulu, Finland. Fitting, James E. 1965 1973 Observations on Paleo-Indian Adaptive and Settlement Patterns. The Michigan Archae- ologist 11(3-4):103-110. History and Crisis in Archaeology. In The Development of North American ArchanIOgy. James E. Fitting, ed. Garden City, N.Y.: Anchor Press/Doubleday. pp. 1-13. Fitzhugh, William W., IV 1970 1972 1973 Environmental Archaeology and Culture Sys- tems in Hamilton Inlet, Labrador: A Survey of the Central Labrador Coast from 3,000 B.C. to the Present. A Thesis Presented to the Department of Anthropology. Cambridge: Harvard University. Environmental Archaeology and Cultural Sys- tems in Hamilton Inlet, Labrador: A Survey of the Central Labrador Coast from 3,000 B.C. to the Present. Smithsonian Contributions to Anthropology No. 16. Washington, D.C.: Smithsonian Institution Press. Preliminary Paper and Discussion. School of American Research Advanced Seminar on Pre- Dorset--Dorset Problems. Santa Fe. Flannery, Kent V. 1968 Archaeological Systems Theory and Early Meso- america. In_ Anthropological Archaeology in the Americas. Anthropological Society of Washington Lecture Series, 1966-67. Washing- ton, D.C.: Anthropological Society of Washington. pp. 67-87. 347 1973 Archaeology with a Capital S. In, Research and Theory in Current Archaeology. Charles L. Redman, ed. New York: John Wiley and Sons, Inc. pp. 47-58. Flerow, C. C. 1967 On the Origin of the Mammalian Fauna of Canada. In. The Bering Land Bridge. David M. Hopkins, ed. Stanford: Stanford Univer- sity Press. pp. 271-280. Foote, Don C. and Bryn Greer-Wootten 1966 Man-Environment Interactions in an Eskimo Hunting System. Paper Prepared for the 133rd Annual Meeting of the American Associa- tion for the Advancement of Science, Section H, AnthrOpology. Washington, D. C. Fredskild, Bent 1967 Paleobotanical Investigation at Sermermiut, Jakobshavn, West Greenland. Meddelelser om Gr¢nland 178(4):45. 1969 A Postglacial Standard Pollendiagram from Peary Land, North Greenland. Pollen et Spores 11:573-583. 1972 Palynological Evidence for Holocene Climatic Changes in Greenland. In Climatic Changes in Arctic Areas During Efie Last Ten Thousand Years. Y. Vasari, H. Hyvarinen and S. Hicks, eds. University of Oulu Symposium Series A 3(1):277-306. Oulu, Finland. 1973 Studies in the Vegetational History of Green- land. Meddelelser on Gr¢n1and l98(4):1-24l. Freeman, Milton, M. R. 1971 A Social and Ecologic Analysis of Systematic Female Infanticide Among the Netsilik Eskimo. American Anthropologist 73(5):1011-1018. French, D. H. 1973 Theoretical and Methodological Aspects of Data Recovery. In The Explanation of Culture Change: Models in Prehistory. Colin Renfrew, ed. London: Duckworth. pp. 105-108. Fritz, J. M. and F. T. Plog 1970 The Nature of Archaeological Explanation. American Antiquity 35:405-412. 1971 1949 1950 1951 1954 1955 1956 1960 1961 1962 1963 1964 1965 1966 348 Geist, Valerius The Relation of Social Evolution and Dis- persal in Ungulates During the Pleistocene, with Emphasis on the Old World Deer and the Genus Bison. Quaternary Research 1(3): 283-315. Giddings, J. Louis, Jr. Early Flint Horizons on the North Bering Sea Coast. Journal of the Washington Academy of Sciences 39(3):85-89. Annals of Series II, Early Man on the Bering Sea Coast. the New York Academy of Science. l3(1):18-21. The Denbigh Flint Complex. American Antiquity 16(3):l93-203. Early Man in the Arctic. Scientific American 190(6):82-88. The Denbigh Flint Complex is not Yet Dated. American Antiquity 20(4):375-376. A Flint Site in Northernmost Manitoba. American Antiquity 21(3):255-268. The Archaeology of Bering Strait. Current Anthropology l(2):121-138. Cultural Continuities of Eskimos. American Antiquity 27(2) 155-173. Onion Portage and Other Flint Sites of the Kobuk River. Arctic Anthropology l(l):6-27. Some Arctic Spear Points and Their Counter- parts. Anthropological Papers of the Univer- sity of Alaska 10(2):l-12. The Archaeology of Cape Denbigh. Providence: Brown University Press. A Long Record of Eskimos and Indians at the Forest Edge. In, Context and Meaning in Cultural AnthrOpology. M. E. Spiro, ed. New York: The Free Press. pp. 89-205. Cross—Dating the Archaeology of Northwestern Alaska. Science 153(3732):127-135. 349 1967 Ancient Men of the Arctic. New York: Alfred A. KnOpf. Goldschmidt, Walter 1959 Man's Way. New York: Holt-Dryden. Goodenough, W. 1971 Culture, Language and Society. Reading, Massachusetts: Addison-Wesley. Gordan, Bryan C. 1972 Interim Report, Upper Thelon River Archaeo- logical-Botanical Project. University of Calgary Department of Archaeology. Mimeo- graphed. Gosselin, A., P. Plumet, P. Richard and J. P. Salaun 1974 Recherches Archéologiques et Paléoécologiques au Nouveau Quebec. Paléo-Québec No. 1x Trois-Rivieres: Universite du Québec a Trois-Riviéres. Gould, Peter R. 1969 Spatial Diffusion. Association of American Geographers Resource Paper No. 4. Graburn, Nelson H. H. 1969 Eskimos without Igloos. Boston: Little, Brown and CO. Griffin, James B. 1943 The Fort Ancient Aspect. Ann Arbor: Uni? versity of Michigan Press. Hadleigh-West, Frederick, Ed. 1963 Early Man in the Western American Arctic: A Symposium, Anthropological Papers of the University of Alaska 10:2. Hagerstrand, Torsten 1967 Innovation Diffusion as a Spatial Process. Chicago: University of Chicago Press. Trans. by A. Pred. (With a Postscript by A. Pred.) Haggett, Peter 1966 Locational Analysis in Human Geography. New York: St. Martin's Press. Haggett, Peter, R. J. Chorley and D. R. Stoddart 1965 Scale Standards in Geographical Research: A New Measure of Areal Magnitude. Nature 206(4974):844-847. 350 Haggett, Peter and Richard J. Chorley 1970 Network Analysis in Geography. New York: St. Martin's Press. Hall, Edwin S. 1970 Excavations at Tukuto Lake: The Late Pre- historic/Early Historic Eskimos of Interior Northwest Alaska. Mimeographed. 1973 Archaeological Investigations in the Noatak River Valley--Summer 1973. Mimeographed. Hansell, R.I.C., D. A. Chant and J. Weintraub 1971 Changes in the Northern Limit of Spruce at Dubawnt Lake, Northwest Territories. Arctic 24(3):233-234. Hare, F. Kenneth 1968 The Arctic. Quarterly Journal of the Royal Meteorological Society 94(302):439-459. Harington, C. R. 1970 A Postglacial Muskox (Ovibos MOschatus) from Grandview, Manitoba, and Comments on the Zoogeography of Ovibos. National Museum of Natural Sciences Publication in Palaeontology No. 2. Ottawa: Queen's Printer. Harp, Elmer E., Jr. 1953 New World Affinities of Cape Dorset Culture. AnthrOpological Papers of the University of Alaska l(2):37-54. 1958 Prehistory in the Dismal Lake Area. Arctic ll(4):2l8-249. 1959 Ecological Continuity on the Barren Grounds. Polar Notes 1:48-56. 1961 The Archaeology of the Lower and Middle Thelon, N.W.T. Arctic Institute of North America, Technical Paper NO. 3. 1962 The Culture History of the Central Barren Grounds. In, Prehistoric Cultural Relations Between the Arctic and Temperate Zones of North America. John M. Campbell, ed. pp. 69- 75. Arctic Institute of North America, Tech- nical Paper NO. 11. Montreal. 1964 The Cultural Affinities of the Newfoundland Dorset Eskimo. National Museum of Canada Bulletin No. 200. Ottawa: Queen's Printer. 351 1973 Preliminary Paper and Discussion. School of American Research Advanced Seminar on Pre- Dorset--Dorset Problems. Santa Fe. Harvey, David 1969 Explanation in Geography. London: Edward Arnold. Haynes, E. Vance 1974 Paleoenvironments and-Cultural Diversity in Late Pleistocene South American: A Reply to A. L. Bryan. Quaternary Research 4(3):378- 382. Hickey, Clifford G. 1974 The Effects of Treeline Shifts on Human Societies: Crazy Quilt Variability vs. Macrozonal Adaptation. In International Conference on the Prehistory and Paleoecology of Western North American Arctic and Sub- arctic. Scott Raymond and Peter Schledermann, eds. pp. 87-100. Calgary, Alberta: Univer- sity of Calgary Archaeological Association. Higgins, G. M. 1968 South Coast-Baffin Island: An Area Economic Survey. Department of Indian Affairs and Northern Development. Information Canada Publications. Ottawa: Queen's Printer. Hill, James N. 1970 Broken K Pueblo: Prehistoric Social Organiza- tion in the American Southwest. Anthropologi- cal Papers of the University of Arizona NO. 18. 1972 The Methodological Debate in Contemporary Archaeology: a Model. In Models in Archaeology. David L. Clarke, ed. London: Methuen & CO., Ltd. pp. 61-108. Hill, James N. and R. K. Evans 1972 A Mbdel for Classification and Typology. In Models in Archaeology. David L. Clarke, ed? London: Methuen & CO., Ltd. pp. 231-273. Hodgman, Charles D., Ed. 1958 Handbook of Chemistry and Physics. Fortieth Edition. Cleveland: Chemical Rubber Publishing Company. 352 Hole, Frank 1973 Questions of Theory in the Explanation of Culture Change in Prehistory. In_ The Explanation of Culture Change: Models in Prehistory. Colin Renfrew, ed. London: Duckworth. pp. 19-34. Hopkins, David M., Ed. 1967 The Bering Land Bridge. Stanford: Stanford University Press. Humphrey, Robert L., Jr. 1970 The Prehistory of the Arctic Slope of Alaska: Pleistocene Cultural Relationships Between Eurasia and North America. Ph.D. Disserta- tion. Department of AnthrOpologY. Univer- sity of New Mexico. Irving, Laurence 1972 Arctic Life of Birds and Mammals--Including Man. New York: Springer-Verlag of Berlin. Irving, William N. 1951 Archaeology in the Brooks Range of Alaska. American Antiquity 17(1):52-53. 1953 Evidence of Early Tundra Cultures in Northern Alaska. Anthropological Papers of the Uni- versity of Alaska 1(2):55-85. 1954 Preliminary Report of an Archaeological Reconnaissance in the Western Brooks Range, Alaska. Mimeographed. 1955 Burins from Central Alaska. American Antiquity 20(4):380-383. 1957 An Archaeological Survey of the Susitna Valley. AnthrOpological Papers of the Uni- versity of Alaska 6(1):37-52. 1962 A Provisional Comparison of Some Alaska and Asian Stone Industries. In Prehistoric Culture Relations Between the Arctic and Temperate Zones of North America. John M. Campbell, ed. pp. 55-68. Arctic Institute of North America, Technical Paper No. 11. Montreal. 1964 Punyik Point and the Arctic Small Tool Tradi- tion. Ph.D. Dissertation, Department of Anthropology, University of Wisconsin, Madison. 353 1968 The Barren Grounds. In Science, History, and Hudson Bay, Vol. I. C. S. Beals, ed. Ottawa: Queen's Printer. pp. 26-54. Isaac, G. L. 1972 Early Phases of Human Behaviour: Models in Lower Palaeolithic Archaeology. 13_ Models in Archaeology. David L. Clarke, ed. London: Methuen & CO., Ltd. pp. 167-200. Jennings, Jesse D. 1968 Prehistory of North America. New York: McGraw-Hill Book Company. Jensen, Adolf S. 1939 Concerning a Change of Climate During Recent Decades in the Arctic and Sub-Arctic Regions, from Greenland in the West to Eurasia in the East, and Contemporary Biological and Geo- graphical Changes. Danske Videnskabernes Sels- kab, Biolegiske Meddelelser l4(8):1-75. Johnsen, S. J., W. Dansgaard, H. B. Clausen and C. C. Langway, Jr. 1970 Climatic Oscillations 1200-2000 A.D. Nature 227:482-483. Johnson, Albert, Leslie Viereck, Ross Johnson and Herbert Melchior 1966 Vegetation and Flora. In. Environment of the Cape Thompson Region, Alaska. Norman J. Wilimovsky and John N. Wolfe, eds. U. S. Atomic Energy Commission Publication PNE-481. Washington, D. C.: U. S. Government Printing. pp. 277-354. Johnson, Leroy, Jr. 1972 Problems in "Avant-Garde" Archaeology. American Anthropologist 74:366-377. Jungerius, P. D. 1969 Soil Evidence of Post-Glacial Tree Line Fluctuations in the Cypress Hills Area, Alberta, Canada. Arctic and Alpine Research 1(4):235-246. Kehoe, Alice B. and Thomas F. Kehoe 1973 Cognitive Models for Archaeological Interpre- tation. American Antiquity 38(2):150-154. Kelsall, John P. 1968 The Migratory Barren-Ground Caribou of Canada. Canadian Wildlife Service Monograph No. 3. Ottawa: Queen's Printer. 354 Kemeny, John G. 1959 A PhilOSOpher Looks at Science. New York: Van Nostrand. Kennewell, J. A. and C. D. Ellyett 1974 Volcanic Dust and Meteor Rates. Science 186:355-357. King-Hele, D. G. 1964 Variations in Upper-Atmosphere Density Be- tween Sunspot Maximum and Minimum. Nature 203:959-960. King-Hele, D. G. and D. M. C. Walker 1960 Variation of Upper—Atmosphere Density with Latitude and Season: Further Evidence from Satellite Orbits. Nature 185(4715):727-729. Knuth, Eigil 1952 An Outline of the Archaeology of Pearyland. Arctic 5(1):l7-33. 1958 Archaeology of the Farthest North. Pro- ceedings of the 32nd International Congress of Americanists in 1956. Copenhagen. 1967 Archaeology of the Musk Ox Way. Contribu- tions du Centre d'Etudes Arctiques et Finno Scandinaves No. 5. Paris. 1968 The Independence II bone Artifacts and the Dorset-evidence in North Greenland. Folk 10:61-80. Krader, L. 1952 Neolithic Find in the Chukchi Peninsula. American Antiquity 17(3):26l-262. Kroeber, Alfred L. and Clyde Kluckhohn 1952 Culture: A Critical Review of Concepts and Definitions. Peabody Museum.of American Archaeology and Ethnology Papers 47(1). Cambridge: Harvard University Press. Kuhn, Thomas S. 1970 The Structure of Scientific Revolutions. Enlarged Edition. Chicago: University of Chicago Press. 355 Kushner, Gilbert, Michey Gibson, John Gulick, John T. Honingmann and Richard Nonas 1962 What Accounts for Sociocultural Change?: A Propositional Inventory. Report of the In- stitute for Research in Social Science. University of North Carolina. Chapel Hill. Ladurie, LeRoy 1971 Times of Feast, Times of Famine: A History of Climate Since the Year 1000. Garden City, New York: Doubleday & CO., Inc. Lamb, H. H. 1964 Trees and Climate History in Scotland. Quaternary Journal of the Royal Meteorologi- cal Society 90:382-394. 1966 The Changing Climate. London: Methuen & CO., Ltd. 1973 Whither Climate Now? Nature 244(5416):395- 397. Langway, Chester 1970 Stratigraphic Analysis of a Deep Ice Core from Greenland. Geographical Society of America Special Publication No. 125. Larsen, Helge 1950 Archaeological Investigations in Southwestern Alaska. American Antiquity 15(3):l77-196. 1968 Trail Creek, Final Report on the Excavation of Two Caves on Seward Peninsula, Alaska. Acta Arctica 15:7-79. Larsen, Helge and Jorgen Meldgaard 1958 Paleo-Eskimo Cultures in Disko Bugt, West Greenland. Meddelelser Om Gr¢nland 161(2): 1-75. Larsen, James A. 1971 Vegetational Relationships with Air Mass Fre- quencies: Boreal Forest and Tundra. Arctic 24(3):177-194. Lasker, Gabriel W. 1970 Physical Anthropology: The Search for General Processes and Principles. American AnthrOpologist 72(1):1-8. Leach, Edmund 1973 Lee, Thomas E. 1973 1974 Leone, Mark P. 1972a Leone, Mark P. 1972b 356 Concluding Address. In The Explanation of Culture Change: Models in Prehistory. Colin Renfrew, ed. London: Duckworth. pp. 761- 771. Rebuttal to Dekin's Review. American Anthropologist 75(6):2037-2040. Comment on the Dekin Affair. Anthropological Journal of Canada 12(3):30-32. Issues in AnthrOpological Archaeology. Ig_ Contemporary Archaeology: A Guide to Theory and Contributions. Mark P. Leone, ed. Carbondale: Southern Illinois University Press. pp. 14-27. , Ed. Contemporary Archaeology: A Guide to Theory and Contributions. Carbondale: Southern Illinois University Press. Levins, Richard 1966 The Strategy of MOdel Building in Population Biology. American Scientist 54(4):421-431. Longacre, William A. 1970 Lowther, G. R. 1962 Lynch, Thomas 1974 Lyons, John B. 1973 Archaeology as Anthropology: A Case Study. Anthropological Papers of the University of Arizona No. 17. An Account of an Archaeological Site on Cape Sparbo, Devon Island. Contributions to Anthropology, 1960. National Museum Of Canada Bulletin, l80(l):l-19. Ottawa: Queen's Printer. F. The Antiquity of Man in South America. Quaternary Research 4(3):356-377. and James E. Mielke Holocene History of a Portion of Northernmost Ellesmere Island. Arctic 26(4):314-323. MacArthur, Robert H. 1972 Geographical Ecology: Patterns in the Dis- tribution of Species. New York: Harper & Row, Publishers, Inc. 357 MacDonald, George F. 1968 Debert: A PaleO-Indian Site in Central Nova Scotia. National Museum of Canada Papers in AnthrOpology NO. 16. Ottawa: Queen's Printer. MacNeish, Richard S. 1954 The Pointed Mountain Site Near Fort Liard, N.W.T., Canada. American Antiquity 19(3): 234-253. 1956 The Engigstciak Site on the Yukon Arctic Coast. AnthrOpological Papers of the Univer- sity of Alaska 4(2):91-111. 1959 A Speculative Framework of Northern North American Prehistory as of April 1959. AnthrOpologica 1:1-17. 1974 Review of The DevelOpment of North American Archaeolagy: Essays in the History of Regional Traditions, by James E. Fitting, ed. American Anthropologist 76:462-463. Malaurie, Jean, Yrjo Vasari, Hannu Hyvarinen, Georgette Delibrias and Jacques Labeyrie 1972 Preliminary Remarks on Holocene Paleoclimates in the Regions of Thule and Inglefield Land, above since the Beginning of Our Era. In Climatic Changes in Arctic Areas During the Last Ten Thousand Years. Y. Vasari, H. Hyvarinen and S. Hicks, eds. University of Oulu Symposium Series A 3(1):105-136. Oulu, Finland. Manners, Robert A. and David Kaplan 1968 Notes on Theory and Non-Theory in AnthrOpology. In Theory in AnthrOpology: A sourcebook. Robert A. Manners and David Kaplan, eds. Chicago: Aldine/Atherton. pp. l-12.- Manning, T. H. 1960 The Relationship of the Peary and Barren Ground Caribou. Arctic Institute of North America Technical Paper No. 4. MOntreal. Manning, T. H. and A. H. MacPherson 1958 The Mammals of Banks Island. Arctic Institute of North America Technical Paper No. 2. Montreal. 358 Mansfield, A. W. 1959 The Walrus in the Canadian Arctic. Fisheries Research Board of Canada, Arctic Unit. Circular No. 2. Montreal: Queen's Printer. 1964 Seals of Arctic and Eastern Canada. Fisheries Research Board of Canada, Arctic Unit. Bulletin No. 137. Ottawa: Queen's Printer. 1966 The Walrus in Canada's Arctic Canadian Geographical Journal 72(3):88-95. Margalef, Ramon 1968 Perspectives in Ecological Theory. Chicago: University of Chicago Press. Martin, Paul S. 1971 The Revolution in Archaeology. American Antiquity 36:1-8. Martin, Paul S. n5. nI. 1962 Chapters in the Prehistory of Eastern Arizona, 1. Fieldiana: Anthropology Vol. 53. Chicago: Chicago Natural History Museum, 1964 Chapters in the Prehistory of Eastern Arizona, 2. Fieldiana: AnthrOpology Vol. 55. Chicago: Chicago Natural History Museum, Mathiassen, Therkel 1925 Preliminary Report of the Fifth Thule Expe- dition: Archaeology. Proceedings of the let International Congress of Americanists. Goteborg, 1924. pp. 206-215. 1927 Archaeology of the Central Eskimos, the Thule Culture and its Position within the Eskimo Culture. Report of the Fifth Thule Expedition, 1921-1924, 4:182. Copenhagen. 1958 The Sermermiut Excavations, 1955. Meddelelser om Grpnland l61(3):52. Matthews, Barry l967a Late Quaternary Land Emergence in Northern Ungava, Quebec. Arctic 20(3):167-202. 1967b 359 Late Quaternary Marine Fossils from Frobisher Bay (Baffin Island, N.W.T., Canada). Palaeogeography, Palaeoclimatology, Palaeoecology 3(1967):243-263. Maxwell, Moreau S. 1960 1962 1967 1973a 1973b May, Robert M. 1973 McConnell, H. 1972 McGhee, Robert 1970a l970b l970c The MOvement of Cultures in the Canadian High Arctic. AnthrOpologica, N.S.:2:2:177-189. Pre-Dorset and Dorset Sites in the Vicinity of Lake Harbour, Baffin Island, N.W.T. Contributions to Anthropology, 1960. National Musewm of Canada Bulletin 180(1): 20-55. The Origin and Development of Dorset Culture. Paper presented at the 32nd Annual Meeting of the Society for American Archaeology, 4 May. Ann Arbor. Archaeology of the Lake Harbour District, Baffin Island. Mercury Series, National Museums of Canada. Archaeological Survey of Canada Paper No. 6. Ottawa: Queen's Printer. Preliminary Paper and Discussion. School of American Research Advanced Seminar on Pre- Dorset--Dorset Problems. Santa Fe. Stability and Complexity in Model Ecosystems. Monographs in Population Biology No. 6. Princeton, N.J.: Princeton University Press. and J. M. Horn Probabilities of Surface Karst. In_ Spatial Analysis in Geomorphology. Richard J. Chorley, ed. New York: Harper & Row, Publishers. pp. 11-133. Excavations at Bloody Falls, N.W.T., Canada. Arctic Anthropology 6(2):53-73. A Quantitative Comparison of Dorset Culture Microblade Samples. Arctic AnthrOpology 7(2):89-96. Speculations on Climatic Change and Thule Culture DevelOpment. Folk 11-12:l73-184. 360 McGhee, Robert, Ed. 1971a Current Research: Far North. American Antiquity 36(4):489-493. McGhee, Robert 1971b An Archaeological Survey Of Western Victoria Island, N.W.T., Canada. Contributions to AnthrOpology VII: Archaeology and Physical AnthrOpology. National Museums of Canada Bulletin No. 232. Ottawa: Queen's Printer. pp. 157-191. l972a Climatic Change and the DevelOpment of Canadian Arctic Cultural Traditions. In Climatic Changes in Arctic Areas During the Last Ten Thousand Years. Y. Vasari, H. Hyvarinen and S. Hicks, eds. University of Oulu Symposium Series A 3(1):39-60. l972b Copper Eskimo Prehistory. Publications in Archaeology No. 2. National Museum of Man. Ottawa: National Museums of Canada. 1973 Preliminary Paper and Discussion. School of American Research Advanced Seminar on Pre- Dorset--Dorset Problems. Santa Fe. 1974 The Early Arctic Small Tool Tradition: A Prediction from Arctic Canada. In_ Inter- national Conference on the Prehistory and Paleoecology of Western North American Arctic and Subarctic. Scott Raymond and Peter Schledermann, eds. pp. 127-132. Calgary, Alberta: University of Calgary Archaeological Association. McGhee, Robert and James Tuck 1973 Preliminary Paper and Discussion. School of American Research Advanced Seminar on Pre- Dorset--Dorset Problems. Santa Fe. McKennan, Robert A. 1969 Athapaskan Groupings and Social Organization in Central Alaska. Contributions to Anthro- pology: Band Societies. National Museums of Canada Bulletin NO. 228. Ottawa: Queen's Printer. pp. 93-115. McLaren, Ian A. 1961 Methods of Determining the Numbers and Availability of Ringed Seals in the Eastern Arctic. Arctic 14(3):162-195. 361 McPhail, J. D. and C. C. Lindsay 1970 Freshwater Fishes of Northwestern Canada and Alaska. Fisheries Research Board of Canada. Bulletin No. 173. Ottawa: Queen's Printer. MacPherson, A. J. 1965 The Origin of Diversity in Mammals of the Canadian Arctic Tundra. Systematic Zoology l4(3):153-173. Meldgaard, Jorgen 1952 1955 1960 1962 A Paleo-Eskimo Culture in West Greenland. American Antiquity 17(3):222-230. Dorset Culture. The Danish-American Expedi— tion to Arctic Canada, 1954. Kmul, Arbog Jysk Arkaeol. Selskab. pp. 158-177. Prehistoric Culture Sequences in the Eastern Arctic as Elucidated by Stratified Sites at Igloolik. Selected Papers of the Fifth International Congress of Anthropological and Ethnological Sciences, 1956. Philadel- phia: University of Pennsylvania Press. pp. 588-595. On the Formative Period of the Dorset Cul- l ture. In, Prehistoric Cultural Relations Between the Arctic and Temperate Zones of North America. John M. Campbell, ed. pp. 92-95. Arctic Institute of North America Technical Paper No. 11. Montreal. Meiklejohn, Christopher 1974 Meyer, D. 1970 Biological Concomitants of a Model of Band Society. In International Conference on the Prehistory and Paleoecology of Western North American Arctic and Subarctic. Scott Raymond and Peter Schledermann, eds. pp. 133-141. Calgary, Alberta: University of Calgary Archaeological Association. Pre-Dorset Settlements at the Seahorse Gully Site. Unpublished master's thesis, Depart- ment of AnthrOpology, University of Manitoba. Miller, Gifford H. 1973 Late Quaternary Glacial and Climatic History of Northern Cumberland Peninsula, Baffin Island, N.W.T., Canada. Quaternary Research 3(4):561-583. 362 Moroney, M. J. 1965 Facts From Figures. Baltimore, Md: Penguin Books, Inc. (Third Revised Edition, Reprinted) Morrill, Richard L. 1970 The Spatial Organization of Society. Belmont, California: Wadsworth Publishing Company, Inc. Munton, R. J. C. 1973 Systems Analysis: A Comment. In The Ex- planation of Culture Change: Madels in Pre- history. Colin Renfrew, ed. London: Duckworth. pp. 685-690. Nash, Ronald J. 1969 The Arctic Small Tool Tradition in Manitoba. Occasional Paper of the Department of Anthropology No. 2. Winnipeg: University of Manitoba. 1973 Preliminary Paper and Discussion. School of American Research Advanced Seminar on Pre- Dorset--Dorset Problems. Santa Fe. Nelson, Richard K. 1969 Hunters of the Northern Ice. Chicago: The University of Chicago Press. ’ Nichols, Harvey l967a The Post Glacial History of Vegetation and Climate at Ennadai Lake, Keewatin, and Lynn Lake, Manitoba (Canada). Eiszeitalter and Gegenwart 18:176-197. l967b Central Canadian Palynology and Its Relevance to Northwestern Europe in the Late Quaternary Period. REVIGW'Of Palaeobotany and Palynology 2:231-243. l967c Pollen Diagrams from Sub-Arctic Central Canada. Science 55(3770):1665-1668. 1970 Late Quaternary Pollen Diagrams from the Canadian Arctic Barren Grounds at Pelly Lake, Northern Keewatin, N.W.T. Arctic and Alpine Research 2(1):43-61. 1972 Summary of the Palynological Evidence for Late-Quarternary Vegetational and Climatic Change in the Central and Eastern Canadian Arctic. In, Climatic Changes in Arctic 363 Areas During the Last Ten Thousand Years. Y. Vasari, H. Hyvarinen and S. Hicks, eds. University of Oulu Symposium Series A 3(1): 309-340. Oulu, Finland. Noble, William C. 1971 Archaeological Surveys and Sequences in Central District of Mackenzie, N.W.T. Arctic Anthropology 8(1):102-135. 1974 The Tunda-Taiga Ecotone: Contributions from the Great Slave-Great Bear Lake region. In International Conference on the Prehistory and Paleoecology of Western North American Arctic and Subarctic. Scott Raymond and Peter Schledermann, eds. pp. 153-172. Calgary, Alberta: University of Calgary Archaeological Association. Odum, Eugene P. 1959 Fundamentals of Ecology. Philadelphia: W. B. Saunders CO. Okladnikov, A. P. 1955 Yakutia Prior to its Merger with the Russian State. MOscow. Olsson, Ingrid U., Ed. 1970 Radiocarbon Variations and Absolute Chrono- logy. Proceedings of the Twelfth Nobel Symposium. New York: John Wiley & Sons, Inc. Pelto, Pertti J. 1970 Anthropological Research: The Structure of Inquiry. New York: Harper & Row, Publi- shers. Plog, Fred 1973a Laws, Systems of Law,.and the Explanation of Observed Variation. In The Explanation of Culture Change: MOdeIs in Prehistory. Colin ggnfrew, ed. London: Duckworth. pp. 649- l973b Diachronic Anthropology. In, Research and Theory in Current Archaeology. Charles L. Redman, ed. New York: John Wiley & Sons, Inc. pp. 181-198. 1974 The Study of Prehistoric Change. (Studies in Archaeology Series) New York: Academic Press. 364 Polach, H. A. and J. Golson 1966 Collection of Specimens for Radiocarbon Dating and Interpretation of Results. Australian Institute of Aboriginal Studies Manual No. 2. Canberra: Australian National University Press. Polunin, Nicholas 1948 Botany of the Canadian Eastern Arctic: Part III; Vegetation & Ecology. National Museum of Canada Bulletin No. 104. Ottawa: King's Printer. Pothier, Roger 1968 Community Complexity and Indian Isolation. In’ Conflict in Culture: Problems of De- velOpmental Change Among the Cree. Norman A. Chance, ed. pp. 33-45. Canadian Research Centre for Anthropology. Ottawa: Saint Paul University. Pred, Allan 1967 Postscript. In. Innovation Diffusion as a Spatial Process. Torsten Hagerstrand. Chicago: University of Chicago Press. Trans. by A. Pred. pp. 299-324. Pruitt, William 0., Jr. 1966 Ecology of Terrestrial Mammals. In Environ- ment of the Cape Thompson Region,_Alaska. Norman J. WilimOvsky and John N. Wolfe, eds. U. S. Atomic Energy Commission Publication PNE-581. Washington, D.C.: U. S. Govt. Printing. pp. 519-564. Rainey, Froelich G. and Elizabeth Ralph 1959 Radiocarbon Dating in the Arctic. American Antiquity 24(4):365-374. Ralph, Elizabeth and Robert Ackerman 1961 University of Pennsylvania Radiocarbon Dates IV. Radiocarbon 3:4-14. Redfield, Robert 1950 A Village that Chose Progress; Chan Kom Revisited. Chicago: University of Chicago Press. Redman, Charles L. l973a Multivariate Approach to Understanding Changes in an Early Farming Community in Southeast Anatolia. In The Explanation of Culture Change: Models in Prehistory. Colin Renfrew, ed. London: Duckworth. pp. 717-724. 1973b 365 Research and Theory in Current Archaeology: An Introduction. In Research and Theory in Current Archaeology. Charles L. Redman, ed. New York: John Wiley & Sons, Inc. pp. 5-20. Redman, Charles L., Ed. 1973c Research and Theory in Current Archaeology. New York: John Wiley & Sons, Inc. Renfrew, Colin 1971 Carbon 14 and the Prehistory of Europe. Scientific American 225(4):63-72. Renfrew, Colin, Ed. 1973 Ritchie, J. C. 1972 Ritchie, J. C. 1971 Rosen, Robert 1972 Rostlund, E. 1952 Rouse, Irving 1953 1960 The Explanation of Culture Change: Models in Prehistory. London: Duckworth. Pollen Analysis of Late-Quaternary Sediments from the Arctic Treeline Of the Mackenzie River Delta Region, Northwest Territories, Canada. In Climatic Changes in Arctic Areas During the Last Ten Thousand Years. Y. Vasari, H. Hyvarinen and S. Hicks, eds. University of Oulu Symposium.Seris A 3(1): 251-271. Oulu, Finland. and F. K. Hare Late-Quaternary Vegetation and Climate Near the Arctic Tree Line of Northwestern North America. Quaternary Research 1(3):331-342. Some Systems Theoretical Problems in Biology. In The Relevance of General Systems Theory. Ervin Laszlo, ed. New York: George Braziller, pp. 43-66. Freshwater Fish and Fishing in Native North America. University of California Publica- tions in Geography No. 9. The Strategy of Culture History. In Anthropology Today. A. L. Kroeber, ed. Chicago: University of Chicago Press. pp. 57-76. The Classification of Artifacts in Archaeology. American Antiquity 25(3):313- 323. 366 Rousseliére, G. M. 1964 Palaeo-Eskimo Remains in the Pelly Bay Region, N.W.T. Contributions to AnthrOpology 1961-62, Part I. National Museum of Canada Bulletin No. 193. Anthropological Series No. 61. Ottawa: Queen's Printer. pp. 162-183. 1968 Reconnaissance Archeologique Dans La Region de Pond Inlet, Territoires du Nord- Ouest. Centre d' Etudes Nordiques Travaux Divers NO. 21. Québec: Universite Laval. 1973 Preliminary Paper and Discussion. School of American Research Advanced Seminar on Pre- Dorset Problems. Santa Fe. Scheidegger, A. E. and W. B. Langbein 1966 Probability Concepts in GeomorphOlogy. U. S. Geological Survey Professional Paper 500-C. Washington, D. C.: U. S. Govt. Printing. Schiffer, Michael B. 1972 Archaeological Context and Systemic Context. American Antiquity 37(2):156-165. Schurrman, C. J. E. 1965 Influence of Solar Flare Particles on the General Circulation of the Atmosphere. Nature 205:167-168. Schweger, Charles 1974 Forest History and Dynamics of the Fisherman Lake Area, N.W.T. and Implications to Human Ecology. In International Conference on the Prehistory and Paleoecology of Western North American Arctic and Subarctic. Scott Raymond and Peter Schledermann, eds. pp. 173-186. Calgary, Alberta: University of Calgary Archaeological Association. Scott, W. B. and E. J. Crossman 1973 Freshwater Fishes of Canada. Fisheries Re- search Board of Canada Bulletin No. 184. Ottawa: Queen's Printer. Service, Elman R. 1969 Models for the Methodology of MOuthtalk. Southwestern Journal of AnthrOpology 25(1): 68-80. Siegel, Bernard J. 1952 Suggested Factors of Culture Change at Taos Pueblo. In. Acculturation in the Americas. 367 Sol Tax, ed. Chicago: University of Chicago Press. pp. 133-140. Silberman, Charles E. 1970 Crisis in the Classroom; The Remaking of American Education. New York: Random House. Singer, Milton 1968 The Concept of Culture. The International Encyclopedia of the Social Sciences Vol. III. Nzw York: The Macmillan Company. pp. 527- 5 3. Slobodkin, Lawrence B. and Howard L. Sanders 1969 On the Contribution of Environmental Pre- dictability to Species Diversity. In Diversity and Stability in Ecological Sys- tems. pp. 82-95. Brookhaven National Laboratory Report NO. 22. Upton, N. Y. Smith, Thomas G. 1973 The Population Dynamics of the Ringed Seal in the Canadian Eastern Arctic. The Fisheries Research Board of Canada Bulletin No. 181. Ottawa: Queen's Printer. Solberg, Ole Martin 1907 Beitrage zur Vorgeschichte der Ost-Eskimo. Steinerna Schneidergerate und Waffenscharfen aus Gronland 2(1). Solecki, Ralph S. 1951 Notes on Two Archaeological Discoveries in Northern Alaska, 1950. American Antiquity l7(l):55-57. Solecki, Ralph S. and R. J. Hackman 1951 Additional Data On the Denbigh Flint Complex in Northern Alaska. Journal of the Washington Academy of Sciences 41(3):85-88. Solecki, Ralph S., Bert Salwen and Jerome Jacobson 1973 Archaeological Reconnaisances North of the Brooks Range in Northeastern Alaska. Univer- sity of Calgary Department of Archaeology Occasional Paper No. l. Soper, J. Dewey 1928 A Faunal Investigation of Southern Baffin Island. National Museum of Canada Bulletin No. 53. Ottawa: King's Printer. 368 Sopher, David E. 1973 Place and Location: Notes on the Spatial Patterning of Culture. In The Idea of Cul- ture in the Social Sciences. Louis Schneider and Charles Bonjean, eds. New York: Cambridge University Press. pp. 101-117. Sorenson, Curtis J., James C. Knox, James A. Larsen and Reid A. Bryson 1971 Sorenson, C. J. and J. G. Knox 1974 Paleosols and the Forest Border in Keewatin, N.W.T. Quaternary Research l(4):468-473. Paleosols and Paleoclimates Related to Late Holocene Forest/Tundra Border Migrations: Mackenzie and Keewatin, N.W.T., Canada. In International Conference On the Prehistory and Paleoecology of Western North American Arctic and Subarctic. Scott Raymond and Peter Schledermann, eds. pp. 187-204. Calgary, Alberta: University of Calgary Archaeological Association. Spaulding, Albert C. 1958 1973 Spencer, Robert F. 1959 The Significance of Differences Between Radiocarbon Dates. American Antiquity 23(3): 1 309-311. 3 Archeology in the Active Voice: The New AnthrOpology. In. Research and Theory in Current Archeology. Charles L. Redman, ed. New York: John Wiley & Sons, Inc. pp. 337- 354. The North Alaskan Eskimo: A Study in Ecology and Society. Smithsonian Institution Bureau of American Ethnology Bulletin 171. Washington, D. C. Steensby, H. P. 1917 Sterud, Gene 1973 An AnthrOpological Study of the Origin of Eskimo Culture. Meddelelser om Gronland 53:39-288. A Paradigmatic View of Prehistory. In The Explanation of Culture Change: Models in Prehistory. Colin Renfrew, ed. London: Duckworth. pp. 3-17. 369 Stickel, E. Gary and Joseph L. Chartkoff 1973 The Nature Of Scientific Laws and Their Re- lation to Law-building in Archaeology. In The Explanation of Culture Change: Models in Prehistory. Colin Renfrew, ed. London: Duckworth. pp. 663-671. Struever, Stuart 1968 Problems, Methods and Organization: A Disparity in the Growth of Archaeology. In Anthropological Archaeology in the Americas. AnthrOpological Society of Washington Lecture Series, 1966-67. Washington, D. C.: AnthrOpological Society of Washington. pp. 120-131. Stuiver, Minze and Hans E. Suess 1966 On the Relationship between Radiocarbon Dates and True Sample Ages. Radiocarbon 8:534-540. Suess, H. E. 1970 Bristlecone-pine Calibration of the Radio- carbon Variations and Absolute Chronology. Ingrid U. Olsson, ed. New York: John Wiley & Sons, Inc. pp. 303-311. Plates I and II. Swadesh, Morris 1951 Diffusional Cumulation and Archaic Residue as Historic Explanations. Southwestern Journal of Anthropology 7:1-21. Tate, Merle W. and Richard C. Clelland 1957 Nonparametric and Shortcut Statistics in the Social, Biological, and Medical Sciences. Danville, Illinois: Interstate Printers and Publishers, Inc. Tauber, Henrik 1960 Copenhagen Natural Radiocarbon Measurements III, Corrections to Radiocarbon Dates Made with the Solid Carbon Technique. Radiocarbon 2:5-11. 1962 Copenhagen Radiocarbon Dates V. Radiocarbon 4:27-34. 1964 COpenhagen Radiocarbon Dates VI. Radiocarbon 6:215-225. 1966 Copenhagen Radiocarbon Dates VII. Radio- carbon 8:213-234. 1968 370 COpenhagen Radiocarbon Dates IX. Radio- carbon 10(2):295-327. Taylor, Walter W. 1968 1972 A Study of Archaeology. Fourth Printing. Carbondale: Southern Illinois University Press. Old Wine and New Skins: A Contemporary Parable. In Contempoary Archaeology: A Guide to THeory and Contributions. Mark P. Leone, ed. Carbondale: Southern Illinois University Press. pp. 28-33. Taylor, William E., Jr. 1962 1964 1965 1966 1967 1968a l968b Pre-Dorset Occupations at Ivugivik in North- western Ungava. In. Prehistoric Cultural Relations Between the Arctic and Temperate Zones of North America. John M. Campbell, ed. pp. 80-91. Arctic Institute of North America Technical Paper NO. 11. Montreal. Interim Account of an Archaeological Survey in the Central Arctic, 1963. Anthropological Papers of the University of Alaska 12(1): 46-55. The Arnapik and Tyara Sites: An Archaeologi- cal Study of Dorset Culture Origins. Ph.D. Dissertation submitted to the Department of AnthrOpology. Ann Arbor: University of Michigan. An Archaeological Perspective on Eskimo Economy. Antiquity 40:114-120. Summary of Archaeological Field Work on Banks Victoria Islands, Arctic Canada, 1965. Arctic AnthrOpologist 4(1):221-243. The Arnapik and Tyara Sites: An Archaeologi- cal Study of Dorset Culture Origins. Memoirs of the Society for American Archaeology No. 22. American Antiquity 33(4:2) :1-129. Prehistory of Hudson Bay, Part 1, Eskimos Of the North and East Shores. In Science, History and Hudson Bay. C. ST Beals, ed. Ottawa: Queen's Printer. pp. 1-26. 1972 Tener, J. S. 1965 Terasmae, J. 1973 Thomas, D. H. 1972 371 An Archaeological Survey Between Cape Parry and Cambridge Bay, N.W.T., Canada in 1963. Mercury Series, National Museums of Canada Archaeological Survey of Canada Paper No. 1 Ottawa: Queen's Printer. Muskoxen in Canada. Canadian Wildlife Service Monograph No. 2. Ottawa: Queen's Printer. Notes on Late Wisconsin and Early Holocene History of Vegetation in Canada. Arctic and Alpine Research 5(3):201-222. A Territorial MOdel for Archaeology: A Behavioural and Geographical Approach. In Models in Archaeology. David L. Clarke, ed. London: Methuen & Co. Ltd. pp. 671-704. Thompson, Raymond H. 1972 Interpretive Trends and Linear MOdels in American Archaeology. In. Contemporary Archaeology: A Guide to Theory and Contribu- tions. Mark P. Leone, ed. Carbondale: Southern Illinois University Press. pp. 34- 38. Thomsen, Thomas 1928 Eskimo Archaeology. [In Greenland: The Past and Present Population of Greenland (Vol. II). M. Vahl, chief ed. Copenhagen: C. A. Reitsel. pp. 271-330. Thostrup, C. B. 1911 Tuck, James 1973 Tucker, G. B. 1964 Ethnographic Description of the Eskimo Settlement and Stone Remains in North-East Greenland. Meddelelser om Gr¢nland 44:183- 355. Preliminary Paper and Discussion. School of American Research Advanced Seminar on Pre- Dorset--Dorset Problems. Santa Fe. Solar Influence on the Weather. Weather 19: 302-311. 372 Usher, Peter J. 1971 Vasari, Yrjo, 1972 The Bankslaners: Economy and Ecology of a Frontier Trapping Community. Vol. No. 2-- Economy & Ecology. Department of Indian Affairs and Northern Development. Informa- tion Canada Publication. Ottawa: Queen's Printer. Hannu Hyvarinen and Sheila Hicks, Eds. Climatic Changes in Arctic Areas During the Last Ten Thousand Years. University of Oulu Symposium Series A 3(1):511. Oulu, Finland. Vibe, Christian 1967 1970 Vilks, G. 1970 Villiers, D. 1970 Arctic Animals in Relation to Climatic Fluctuations. Meddelelser om.Gr¢n1and 170(5). The Arctic Ecosystem Influenced by Fluctua- tions in Sun-Spots and Drift-Ice Movements. In Productivity and Conservation in Northern Circumpolar Lands: A Symposium. W. A. Fuller and P. G. Kevan, eds. pp. 115-120. International Union for Conservation of Nature and Natural Resources Publications, New Series NO. 16. Morges, Switzerland. Circulation of Surface Waters in Parts of the Canadian Arctic Archipelago Based on Fora- miniferal Evidence. Arctic 23(2):100-1ll. The Central Arctic: An Area Economic Survey. Department of Indian Affairs and Northern Development. Information Canada Publication. Ottawa: Queen's Printer. Wallace, Anthony F. C. 1961 1966 Culture and Personality. New York: Random House. Review Q: The Revolution in AnthrOpology, by I. C. Jarvie. American AnthrOpologist 68: 1254-1255. Wallis, W. Allen and Harry V. Roberts 1956 Statistics, A New Approach. New York: Macmillan Company. 373 Watson, Patty Jo, Steven A. LeBlanc and Charles L. Redman 1971 Explanation in Archaeology: An Explicitly Scientific Approach. New York: Columbia University Press. Wauchope, Robert, Ed. 1956 Seminars in Archaeology: 1955. Memoirs of the Society for American Archaeology No. 11. Wax, Murray L. 1973 Cultural Pluralism, Political Power, and Ethnic Studies. In Learning and Culture. Solon T. Kimball and Jacquetta H. Burnett, eds. Proceedings of the 1972 Annual Meeting of the American Ethnological Society. University of Washington Press. Weidick, Anker 1968 Observations on Some Holocene Glacier Fluctu- ations in West Greenland. Meddelelser om Gr¢n1and 165(5):194. 1972 Notes on Holocene Glacial Events in Climatic Changes in Arctic Areas During the Last Ten Thousand Years. Y. Vasari, H. Hyvarinen and S. Hicks, eds. University of Oulu Symposium Series A 3(1):177-204. Oulu, Finland. Wendland, Wayne M. and David L. Donley 1971 Radiocarbon--Ca1endar Age Relationship. Earth and Planetary Science Letters 11:135- 139. Wexler, H. 1956 Variations in Insolation, General Circulation and Climate. Tellus 8:480-494. Whallon, Robert, Jr. 1968 Investigations of Late Prehistoric Social Organization in New York State. In, New Perspectives in Archeology. S. R. Binford and L. R. Binford, eds. Chicago: Aldine. pp. 223-244. White, Leslie A. 1954 Review nI Culture: A Critical Review of Concepts and Definitions, by A. L. Kroeber and C. Kluckhohn. American AnthrOpologist 56:461-486. 1959 The Evolution of Culture. New York: McGraw-Hill. 374 Wilcox, John J., Phillip H. Scherrer, Leif Svalgaard, Walter O. Roberts and Roger H. Olson 1973 Solar Magnetic Sector Structure: Relation to Circulation of the Earth's Atmosphere. Science 180:185-186. Wilimovsky, Norman J. and John N. Wolfe, Eds. 1966 Environment of the Cape Thompson Region, Alaska. U. S. Atomic Energy Commission Publication PNE-481. Washington, D. C.: U. S. Government Printing. Wilkinson, P. F. 1972 Ecosystem Models and Demographic Hypothesis: Predation and Prehistory in North America. In Models in Archaeology. David L. Clarke, Ed. London: Methuen & CO., Ltd. pp. 543- 576. Willett, H. C. 1951 Extrapolation of Sunspot-Climate Relation- ships. Journal Of Meteorology 8:1-6. 1961 The Pattern of Solar Climatic Relationship. Annals of the New York Academy of Science 95(1):89-106. Willey, Gordon R. 1966 An Introduction to American Archaeology, Vol. I. Englewood Cliffs, N.J.: Prentice Hall. Wilmeth, Roscoe 1971 Canadian Archaeological Radiocarbon Dates. Contributions to Anthropology VII: Archae- ology and Physical AnthrOpology. National Museums of Canada Bulletin No. 232. Ottawa: Queen's Printer. pp. 68-126. Wilson, Godfrey and MOnica Wilson 1945 The Analysis of Social Change. Cambridge: University of Cambridge Press. Yellen, John, and Henry Harpending 1972 Hunter—gatherer populations and archaeologi- cal inference. World Archaeology 4(2):244- 253. Yeomans, K. A. l968a Statistics for the Social Scientist: 1 Introducing Statistics. Baltimore, Md.: Penguin Books, Inc. 375 l968b Statistics for the Social Scientist: 2 Applied Statistics. Baltimore, Md.: Penguin Books, Inc.