Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Pre-colonial Sto:lo-Coast Salish community organization : an archaeological study Schaepe, David M. 2009

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
[if-you-see-this-DO-NOT-CLICK]
ubc_2009_spring_schaepe_david.pdf.pdf [ 16.34MB ]
Metadata
JSON: 1.0058411.json
JSON-LD: 1.0058411+ld.json
RDF/XML (Pretty): 1.0058411.xml
RDF/JSON: 1.0058411+rdf.json
Turtle: 1.0058411+rdf-turtle.txt
N-Triples: 1.0058411+rdf-ntriples.txt
Original Record: 1.0058411 +original-record.json
Full Text
1.0058411.txt
Citation
1.0058411.ris

Full Text

       PRE-COLONIAL STÓ:LŌ-COAST SALISH COMMUNITY ORGANIZATION: AN ARCHAEOLOGICAL STUDY                     by  DAVID M. SCHAEPE  B.A., New York University, 1989 M.A., Simon Fraser University, 1998    A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF  THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY   in   The Faculty of Graduate Studies  (ANTHROPOLOGY)     THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  February 2009   © David M. Schaepe, 2009  iiABSTRACT  This study integrates settlement and community archaeology in investigating pre-colonial Stó:lō-Coast Salish community organization between 2,550-100 years before present (cal B.P.).  Archaeological housepits provide a basic unit of analysis and proxy for households through which community organization manifests in relationships of form and arrangement among housepit settlements in the lower Fraser River Watershed of southwestern British Columbia.  This study focuses on spatial and temporal data from 11 housepit settlements (114 housepits) in the upriver portion of the broader study area (mainland Gulf of Georgia Region).  These settlements were mapped and tested as part of the Fraser Valley Archaeology Project (2003-2006).   The findings of this study suggest a trajectory of continuity and change in community organization among the Stó:lō-Coast Salish over the 2,500 years preceding European colonization.  Shifts between heterarchical and hierarchical forms of social organization, and corporate to network modes of relations represent societal transformations that become expressed by about 550 cal B.P.   Transformations of social structure and community organization are manifest as increasing variation in housepit sizes and settlement patterns, and the development of central arrangements in both intra- and inter-settlement patterns.  In the Late Period (ca. 550-100 cal. B.P.), the largest and most complex settlements in the region, including the largest housepits, develop on islands and at central places or hubs in the region’s communication system along the Fraser River.  These complex sets of household relations within and between settlements represent an expansive form of community organization.  Tracing this progression provides insight into the process of change among Stó:lō pithouse communities.     iiiSocietal change develops as a shift expressed first at a broad-based collective level between settlements, and then at a more discreet individual level between households.  This process speaks to the development of communities formed within a complex political-economic system widely practiced throughout the region.  This pattern survived the smallpox epidemic of the late 18th century and was maintained by the Stó:lō up to the Colonial Era.  Administration of British assimilation policies (e.g., Indian Legislation) instituted after 1858 effectively disrupted but failed to completely replace deeply rooted expressions of Stó:lō community that developed during preceding millennia.     ivTABLE OF CONTENTS   ABSTRACT...................................................................................................................................ii  TABLE OF CONTENTS............................................................................................................. iv  LIST OF TABLES .....................................................................................................................viii  LIST OF FIGURES ..................................................................................................................... ix  GLOSSARY OF HALQ’EMÉYLEM TERMS......................................................................xvii  ACKNOWLEDGEMENTS.....................................................................................................xviii  DEDICATION............................................................................................................................. xx  CHAPTER I     INTRODUCTION.............................................................................................. 1 1.1 Community.......................................................................................................................... 2 1.2 Community and Settlement Archaeology ........................................................................... 5 1.3 Housepits: A Proxy to Political-Economic Relations and Community Formation............. 6 1.4 Relations between Ethnography, Archaeology, and Stó:lō-Coast Salish Community Organization........................................................................................................................ 9 1.4.1 Archaeology and the use of Ethnographic Interpretations and Models .................. 9 1.5 Defining Terms and Developing Questions about the Effects of the Colonial Period ..... 12 1.6 Ethnographic, Archaeological, and Ethnohistorical Discussions of Households, Villages, and Socio-Political Organization among Northwest Coast and Central Coast Salish Peoples .............................................................................................................................. 14 1.7 Defining the Study Area - the Mainland Gulf of Georgia Region.................................... 28 1.8 Defining the Terms Stó:lō and Coast Salish ..................................................................... 33 1.9 Chapter Outline and Descriptions ..................................................................................... 34  CHAPTER II - LINKING COMMUNITY AND SETTLEMENT ARCHAEOLOGY ....... 39 2.1 The Socio-Political Taphonomy of Community: Defining Agents of Formation and Change............................................................................................................................... 39 2.1.1 Knowledge and Economy ..................................................................................... 40 2.1.2 Time and Strategy ................................................................................................. 42 2.1.3 Power and Habitus ................................................................................................ 44 2.1.4 Tradition and Competition .................................................................................... 51 2.1.5 Interrelationships and Transformation .................................................................. 54 2.2 Integrating Community and Settlement Archaeology....................................................... 58 2.3 Examining Current Relations between Community and Settlement Archaeology........... 61 2.4 Stratification and Corporate-Network Relations: A Framework for Investigating the Evolution of Housepit Communities................................................................................. 66       vCHAPTER III - DATA COLLECTION: MAPPING AND TESTING HOUSEPIT SETTLEMENTS ............................................................................................................. 71 3.1 Housepit Form and Indigenous Identity in the Fraser Valley ........................................... 76 3.2 Archaeological Data on Housepits in the Mainland Gulf of Georgia Region .................. 77 3.3 Settlement Data ................................................................................................................. 80 3.4 Feature Data ...................................................................................................................... 82 3.5   Housepit Form - Size and Shape....................................................................................... 83 3.5.1  Step 1 - preparing feature-level maps ................................................................... 84 3.5.2 Step 2a - defining feature outline / perimeter........................................................ 84 3.5.3 Step 2b - defining feature dimensions................................................................... 96 3.5.4 Step 3 - gathering accurate measurements ............................................................ 97 3.6 Measuring Age ................................................................................................................ 100 3.6.1 Test Excavations: Soil Probes, Auger Tests, and Shovel Tests .......................... 101  CHAPTER IV - HOUSEPIT SETTLEMENT DISTRIBUTION......................................... 106 4.1 Mapping Settlement Location and Transportation Systems ........................................... 108 4.2 Mapping Transportation Networks and Housepit Settlement Locations ........................ 111 4.3 Analyzing Primary Water Travel Route Distances......................................................... 114 4.4 Analyzing Secondary Water Route Travel Distances..................................................... 124 4.5 Analysis of Terrestrial Route Travel Distances .............................................................. 127 4.6 Defining ‘Hubs’ of Interaction along the Fraser River and its Tributaries ..................... 130 4.7 Discussion of Communication Networks and Settlement Patterning ............................. 134  CHAPTER V - FRAMING TIME: A CHRONOLOGY OF UPRIVER HOUSEPIT FEATURES ................................................................................................................... 138 5.1 Establishing a Chronology of Housepit Features............................................................ 138 5.2 Defining the Chronology of Settlement Occupations ..................................................... 150 5.3 Chronologies of Rebuilding and Disturbances Affecting Housepit Form...................... 154  CHAPTER VI - ANALYSIS LEVEL I: INVESTIGATING HOUSEPIT AND HOUSEHOLD VARIATION....................................................................................... 158 6.1 Comparing Housepit Areas (m2) and Classifying Housepits Sizes ................................ 159 6.2 Sorting Housepits by Size through Time ........................................................................ 166 6.3 Investigating Social Organization at the Housepit / Household Level ........................... 170 6.4 Investigating Housepit Shapes through Time ................................................................. 180 6.5 Summary of Analyses of Relationships between Housepit Size, Age, and Shape ......... 190  CHAPTER VII - INTRA-SETTLEMENT HOUSEPIT RELATIONS............................... 194 7.1 Defining Settlements and Occupations ........................................................................... 195 7.2 Measuring Intra-Settlement Variation............................................................................. 197 7.3 Exploring Settlement Layouts and Arrangements .......................................................... 205 7.3.1 Settlement Layout ............................................................................................... 220 7.3.2 Settlement Arrangement...................................................................................... 222 7.4 Summary of the Findings in the Analysis of Intra-Settlement Relationships................. 224       viCHAPTER VIII - INTER-SETTLEMENT RELATIONS ................................................... 227 8.1 Housepit Settlement Sizes and Size Classes ................................................................... 228 8.2 The Regional Distribution of Housepit Settlements ....................................................... 236 8.3 Summary of Findings in the Analysis of Inter-Settlement Relations ............................. 244  CHAPTER IX - CONTINUITY AND CHANGE IN THE ORGANIZATION OF STÓ:LŌ-COAST SALISH HOUSEPIT COMMUNITIES ...................................... 246 9.1 Schematic I - Modeling Social Relations through Time................................................. 247 9.2 Schematic II - Modeling Social Organization through Time.......................................... 254 9.2.1 A Comparison of Ethnographic and Archaeological Models of Social Organization.................................................................................................................... 257 9.3 Schematic III - Modeling Political-Economic Relations and Community Organization through Time: Developing an Hypothesis ...................................................................... 263  CHAPTER X - CONCLUSION............................................................................................... 284 10.1 Household Archaeology, Settlement Studies -- Methodological Implications............... 285 10.2 Implications for Political-Economic and Material Economic Relations......................... 286 10.3 Implications about Community....................................................................................... 291 10.4 Implications for an Integrated Community-Settlement Archaeology............................. 294 10.5 Implications for Landscape Archaeology ....................................................................... 300 10.6 Implications for Stó:lō-Coast Salish Community Organization ..................................... 302 10.7 Implications for a Developed Northwest Coast Pattern.................................................. 306 10.8  Implications for Relations between Archaeology and Ethnography............................... 307 10.9 Future Directions of Research......................................................................................... 310 10.10 Conclusion....................................................................................................................... 312  BIBLIOGRAPHY ..................................................................................................................... 316  APPENDIX I - HOUSEPIT DATA FROM SAMPLED UPRIVER SETTLEMENTS - ARRANGED BY SETTLEMENT............................................................................... 341  APPENDIX II - MAPS DEFINING HOUSEPIT POLYGONS AND LINE-WORK USED FOR TAKING AREA AND DIMENSION MEASUREMENTS.................. 346  APPENDIX III - RADIOCARBON DATA SET OF DATED ‘UPRIVER GROUP’ HOUSE FEATURES (n=46). ....................................................................................... 359  APPENDIX IV - RADIOCARBON DATA SUB-SET (extracted from Schaepe, Blake, Formosa, and Lepofsky 2006). ..................................................................................... 364  APPENDIX V - EXAMPLES OF TEST UNIT STRATIGRAPHIC PROFILES .............. 368 Qithyil Island (DhRl-15) - Test Unit Profiles ............................................................................. 368 DhRl-15-F4-SP2 ............................................................................................................. 368 DhRl-15-F6 - Plankhouse feature riverbank exposure - profile...................................... 369 Th’ewá:lí (DgRl-17) - Test Unit Profiles.................................................................................... 370 DgRl-17-F8-AT2............................................................................................................. 370 Eyxel (DiRi-48)  - Test Unit Profiles.......................................................................................... 371 DiRi-48-F1-AT2 ............................................................................................................. 371   viiShxw’ow’hamel (DiRj-30)  - Test Unit Profiles......................................................................... 372 DiRj-30-F13-ST1 ............................................................................................................ 372 DiRj-30-F18-ST4 ............................................................................................................ 373 Xelhálh (DjRi-14) - Test Unit Profiles........................................................................................ 374 DjRi-14-F28-ST2 ............................................................................................................ 374 DjRi-14-F13-ST6 ............................................................................................................ 375  APPENDIX VI - HOUSEPIT SETTLEMENT TRAVEL DISTANCE DATA .................. 376  APPENDIX VII - SPECIFIC UPRIVER GROUP HOUSEPIT SETTLEMENT DATA.. 380   viiiLIST OF TABLES  Table 5.1.  Radiocarbon dates from all 34 radiocarbon-dated house features  in the Central, Upper, and Canyon sections of the Fraser Valley,  sorted by age (p = .05; calibrated at 2σ) -- as of 2006 ...............................139 Table 5.2. Chronological Classification of Housepit Features and Settlement  Occupations by Period ...............................................................................156 Table 6.1.  Frequency, proportion, range (minimum-maximum), and   inter-quartile range (IQR) values for area, maximum length, and   maximum width measurements associated with Housepit Size   Classes I-VI, based on 114 Upriver Group housepits ................................165 Table 6.2.  Cross-tabulation of housepit sizes and ages (Periods I-III) .............................178 Table 6.3. Cross-tabulation of housepit shapes and ages, based on   112 housepit features..................................................................................188 Table 6.4. Cross-tabulation of housepit size and shape, based on                          113 housepit features..................................................................................189 Table 7.1. Quantitative definition of housepit settlement attributes ..................................199 Table 7.2. Summary descriptions of housepit settlement layouts and arrangements.........218 Table 8.1.  Housepit frequencies, roofed area values, and settlement size classes --  for all 17 Upriver Group housepit settlement / occupation options,             arranged chronologically............................................................................229 Table 8.2. Settlement Size Class descriptive statistics.......................................................231 Table 8.3. Cross-tabulation of settlement sizes and ages (Periods I-IIIb) .........................233     ixLIST OF FIGURES Figure 1.1.   Wayne Suttles’ (1987:12) Coast Salish ‘Village Organization Model’ ..........21 Figure 1.2    The Gulf of Georgia Region within the Central Coast Salish Culture Area....31 Figure 1.3.   Geographic sub-sections of the Mainland Gulf of Georgia Region /  lower Fraser River Watershed; FC = lower Fraser Canyon;  UV = Upper Fraser Valley; CV = Central Fraser Valley;  LV = Lower Fraser Valley; and D = Fraser Delta  (Lepofsky, Schaepe, and Blake 2006)..........................................................32 Figure 1.4    Map of the Central Coast Salish area showing the locations of  principal villages in the early 19th century (Suttles 1990:454) .......................35  Figure 3.1.   The location of housepit settlements included in this study ............................81 Figure 3.2.   ‘John Mack Slough’ (DhRl-T1) surface map with housepit features..............85 Figure 3.3.   Qithyil Island (DhRl-15) surface map with housepit features (F1-5)                           and apparent plankhouse platform (F6) .......................................................86 Figure 3.4.   Sqwa:la (DhRl-6) surface map with housepit features ...................................87 Figure 3.5.   Th’ewá:lí (DgRl-17) surface map with features..............................................88 Figure 3.6.   Eyxel (DiRi-48) surface map with housepit features ......................................89 Figure 3.7.   Sxwóxwiymelh ‘South’ surface map with housepit features ..........................90 Figure 3.8.   Composite surface image of Sxwóxwiymelh (DiRj-1) ‘South’ and ‘North’          with radiocarbon results - per Lenert and Lepofsky (2005, 2006)...............91 Figure 3.9.   Shxw’ow’hamel (DiRj-30) surface map with housepit features .....................92 Figure 3.10.  Xelhálh (DjRi-14) surface map with features.................................................93 Figure 3.11.  Welqámex (DiRi-15) surface map with features (Graesch 2006) ..................94 Figure 3.12.  Hiqelem (DhRl-T2) surface map with features (Lepofsky 2006) ..................95 Figure 3.13.  Schematic illustration of the housepit measurement system applied to            Xelhálh (east settlement area) shaded contour/topographic map  (with 5 cm topographic intervals) ................................................................96 Figure 4.1     The spatial distribution of 112 housepit settlements throughout the  mainland Gulf of Georgia Region, showing a pattern of downriver  and upriver settlement groupings ...............................................................109 Figure 4.2.   Schematic illustration of waterways (primary and secondary) and  pathways throughout the region.................................................................110 Figure 4.2.   The structure of primary water route travel distances (km) shown in  a stem-and-leaf plot....................................................................................115 Figure 4.3.   The structure of primary water route travel distances (km) shown in  a box plot....................................................................................................115 Figure 4.4.   Spatial distribution of downriver and upriver housepit  settlement groups (n=112)..........................................................................116 Figure 4.5.   Spatial distribution of ‘upriver group’ housepit settlements (n=104) ...........117 Figure 4.6.   Box plot of primary water route travel distances for   upriver group settlements (n=104).............................................................118 Figure 4.7.   Stem-and-leaf plot of primary water route travel distances for  all upriver group settlements (n=104)........................................................118 Figure 4.8.   Stem-and-leaf plot for primary water route travel distances associated  with upriver settlements -- upriver Group 1...............................................119   xFigure 4.9.   Stem-and-leaf plot for primary water route travel distances associated  with upriver settlements -- upriver Group 2...............................................119 Figure 4.10.  Box plots of primary water route travel distances for  upriver Groups 1 and 2...............................................................................119 Figure 4.11.  Stem-and-leaf plot of upriver settlement Group 2 -- based on  primary water route travel distances ..........................................................120 Figure 4.12.  Stem-and-leaf plot of upriver settlement Group 3 -- based on  primary water route travel distances ..........................................................120 Figure 4.13.  Box-plots of upriver Groups 1-3 (Central Fraser Valley, Upper Fraser          Valley, and Lower Fraser Canyon housepit settlement groups)  -- based on analysis of primary water route travel distances                             to each settlement from the mouth for the Fraser River.............................121 Figure 4.14.  Spatial distribution of upriver Groups 1-3 (Central Fraser Valley,  Upper Fraser Valley, and Fraser Canyon housepit settlement groups)                    -- based on analysis of primary water route travel distances to each               settlement from the mouth for the Fraser River .........................................122 Figure 4.15.  Spatial distribution of upriver housepit settlements defined as  five groups (West-Central Fraser Valley, Central Fraser Valley,                 East-Central Fraser Valley, Upper Fraser Valley, Lower Fraser Canyon)             -- based on analysis of primary water route travel distances to each          settlement from the mouth of the Fraser River ..........................................123 Figures 4.16a-c.  Stem-and-leaf plots of all upriver housepits settlements (n=51)             distanced from the Fraser River by travel along a secondary water route,             and two sub-groups of settlements -- located either near to (< 8 km)                    or far from (> 8 km) the Fraser River -- based on the distance of                travel along secondary waterways .............................................................124 Figure 4.17.  Box-plots of three groups of housepit settlements defined by secondary          water route travel distances (km) from the Fraser River -- Fraser    River-side settlements and those near to or far from the Fraser River         accessed by secondary waterways .............................................................125 Figure 4.18.  Spatial distribution of three groups of housepit settlements defined by        secondary water route travel distances from the Fraser River -- Fraser              River-side settlements and those near to or far from the Fraser River          accessed by secondary waterways .............................................................126 Figure 4.19.  Stem-and-leaf plot of terrestrial route travel distances (km)  for settlements necessarily accessed by overland pathways ......................127 Figure 4.20.  Box-plot of terrestrial route travel distances (km) for  settlements necessarily accessed by overland pathways............................127 Figure 4.21.  Spatial distribution of pithouse settlements categorized by final terrestrial      travel-based groups -- no terrestrial travel, short terrestrial travel                          (< 0.3 km), and moderate-to-long terrestrial travel (> 0.3 km)..................129 Figure 4.22.  The general locations of upriver ‘hubs’ -- central locations in the              distributions of each of the three sub-set upriver housepit groups   and geographic connective points among Central Fraser Valley,   Upper Fraser Valley, and Lower Fraser Canyon housepit   settlement groups .......................................................................................131 Figure 4.23.  Halkomelem language area and Stó:lō tribes of the lower Fraser River    watershed (after Smith 2001) .....................................................................135   xiFigure 5.1.   Chronologically ordered sequence of 34 radiocarbon ages (cal B.P.; p = .05; calibrated at 2σ) showing maximum, minimum, and mid-point values  for all radiocarbon- dated housepit features in the Upriver Group area   (as of 2006), connected by mid-point ........................................................142 Figure 5.2.   Chronologically ordered sequence of radiocarbon ages (cal B.P.; p = .05;    calibrated at 2σ) showing maximum, minimum, and mid-point values               for all 29 radiocarbon dated housepit features from the 12 settlements    included in the Upriver Group sample (as of 2006),                                             connected by mid-point..............................................................................143 Figure 5.3.   Histogram showing the distribution (150 year intervals) of  radiocarbon ages from 29 housepit features among the 12 settlements                   in the Upriver Group sample......................................................................145 Figure 5.4.   Histogram showing of the distribution (250 year intervals) of  radiocarbon ages from 29 housepit features among the 12 settlements                   in the Upriver Group sample......................................................................145 Figure 5.5.   Graph showing the distribution of radiocarbon age ranges and mid-point                 values (cal B.P.; p = .05; calibrated at 2σ) for sampled Upriver Group settlements/occupations, based on of representative housepits..................146 Figure 5.6.   Chronology developed for the purpose of classifying Upriver Group               housepits and settlements...........................................................................147 Figure 5.7.   Bar graph showing the frequency of sampled Upriver Group housepits          (n=114) by the three main age categories -- Periods I - III -- as the                     most robust set of groupings ......................................................................149 Figure 5.8.   Bar graph showing the frequency of sampled Upriver Group housepits            (n=114) by age category sub-groups -- Periods Ia-b, Period II, and                Periods IIIa-b..............................................................................................150 Figure 5.9.   Bar graph showing the frequency of housepit settlement occupations  per Period I (2550-2000 cal B.P.); Period II (1400-950 cal B.P.), and              Period III (550-100 cal B.P.)......................................................................151 Figure 6.1.    Box-plot of illustrating the mid-range, inter-quartile range, and median            area measurements (m2) for of the total Upriver Group’ sample of 114   housepits, identifying three outliers as the largest (‘Very Large’)   category of house sizes...............................................................................160 Figure 6.2.    Stem-and-leaf plot illustrating the shape and spread of area measurements              (m2) for all 114 Upriver Group housepits, identifying the two largest                   size categories --‘Very Large’ and ‘Large’ -- as groups of values               respectively representing outliers and a distinct mode ..............................161 Figure 6.3.    Stem-and-Leaf plot showing the distribution of area (m2) area           measurements and definition of size classes among the truncated               sample of 104 housepits.............................................................................162 Figure 6.4.    Histogram illustrating the distribution of area (m2) measurements for                     all 114 Upriver Group housepits -- defined as Size Classes I-VI ..............163 Figure 6.5.    Box-plots illustrating the range, inter-quartile range, and median area                  (m2) of Size Classes I-VI, including all 114 Upriver Group housepits......163 Figure 6.6.    Frequencies and proportions of Upriver Group housepits  by Size Classes I-VI...................................................................................164   xiiFigure 6.7.   The distribution of housepit area values (m2) by radiocarbon age midpoints            for all dated housepit features (n=27) showing multi-hundred year gaps     between clusters of housepits representing Periods I-III, and a wide                 range of housepit sizes particularly between 100-550 cal B.P...................166 Figure 6.8.   Comparison of inter-quartile range, mid-range, and median housepit                   area values (m2) through Periods Ia-IIIb for all Upriver Group    housepits (n=114), showing a progressive, step-wise increase in    housepit sizes through time........................................................................167 Figure 6.9.    Comparison of inter-quartile range, mid-range, and median housepit area           values (m2) of all Upriver Group housepits (n=114) arranged by age,         showing an increasing trend in housepits sizes through Periods I, II,                  and III, with a particularly large increase between Periods II and III........168 Figure 6.10.  Box-plot comparing of housepit area values (m2) for all 114 Upriver Group housepits arranged by age, illustrating significant difference between     housepits sizes in pre- and post-550 cal B.P. periods ................................169 Figure 6.11.  Plot showing the distribution of housepit area measurements (m2) by                   Size Class through time, across Periods Ia-IIIb .........................................172 Figure 6.12.  Histogram comparing the relative frequencies of housepit Size Classes                   I-VI through time, across Periods Ia, Ib, II, IIIa, and IIIb.  Bar percent             values indicate the percentage of housepits of each size class within               each period .................................................................................................173 Figure 6.13.  Histogram comparing the relative frequencies of housepit Size Classes                    I-VI through time, across Periods I, II, and III.  Bar percent values               indicate the percentage of housepits of each size class within                            each period .................................................................................................174 Figure 6.14.  The distribution of housepit area values (m2), defined by Size Class and chronologically grouped into Periods I, II, and III -- showing a                   dramatic increase in the range of housepit sizes associated with                    Period III (100-550 cal B.P.)......................................................................175 Figure 6.15.  Distribution of area measurements (m2), by Size Class, for all Upriver             Group housepit features (n=114) showing a marked difference in                        the distribution of Class V and VI features between pre- and                            post-550 cal B.P periods ............................................................................176 Figure 6.16.  Comparison showing differences in the relative frequencies of pre- and               post-550 cal B.P housepit size classes, forming respectively normal                 and upwardly skewed shapes.  Bar percent values indicate the                percentage of housepits of each size class within each period...................177 Figure 6.17.  Histogram comparing the relative proportions of housepit Size Classes                   in Pre- and Post-550 cal B.P. time periods.  Bar frequency values               indicate the number of housepits of each size class within each period ....177 Figure 6.18.  Comparison of the relative proportions of housepit size classes as                      proxies showing ‘oblong’ and ‘pear-shaped’ social structures and     indicating differences in social stratification and household relations     along a vertical plane of relations in pre- and post-550 cal B.P.     time periods................................................................................................180   xiiiFigure 6.19.  Scatterplot defining square, circular, and rectilinear/oblong shape      characteristics of 114 Upriver Group housepits from all time periods,               based on length/width (W/L) and width/diagonal (W/D)   measurement proportions ...........................................................................182 Figure 6.20. The relative percentages of square, circular, and rectangular/ovoid-                 shaped housepits for 113 Upriver Group housepits throughout all   time periods, excluding DiRj-30-F17 as an uncertain feature type.                       Bar percent values indicate the percentage of housepits of each shape.....183 Figure 6.21.  Plot of 27 radiocarbon-dated Upriver Group housepits (mid-points)   by shape, showing the age distribution of circular house shapes as          clustering within the Late Period / Period III (550-100 cal B.P.),                                   whereas square and rectangular/oblong housepits are present                               in all three time periods..............................................................................184 Figure 6.22.  The distribution of housepit length/width and width/diagonal    measurement ratios, and related shapes for all 114 Upriver Group              housepits, corresponding with Periods I-III; two features (DiRj-1a-F16              and DiRj-30-F17) are uncertain as to their age and feature type ...............185 Figure 6.23a-c. Comparison of housepit shapes through time plotting the distribution                  of housepit shapes within Periods I-III and showing a cluster of all                     but two circular features in Period III (noting the location of the                      two uncertain Period I features otherwise classified as circular)...............186 Figure 6.24.  Relative proportions of square, circular, and rectangular/ovoid housepits associated with Periods I-III (excluding DiRj-30-F17 and DiRj-1a-F16                 as uncertain features in this context). Bar frequency values indicate                  the number of housepits of each shape within each period........................187 Figure 6.25.  Percentages of rectangular/ovoid, square, and circular-shaped    housepits associated with Periods I-III, showing that all circular    features are associated with Period III (excluding DiRj-30-F17 and    DiRj-1a-F16 as uncertain features in this context).  Bar percent values    indicate the percentage of housepits of each shape within each period.....187 Figure 6.26.  Bar graph showing the distribution of housepit shapes by size as                   percentages of rectangular/ovoid, square, and circular-shaped housepit            types associated with Size Classes I-VI (excluding DiRj-30-F17 as                    an uncertain feature in this context).  Bar frequency values indicate                the number of housepits of each shape within each size class...................190 Figure 7.1     Regional map showing the spatial distribution and clustering of Upriver               Group housepit features -- comprising 10 distinct archaeological sites ....196 Figure 7.2.    Plot showing a downward trend in Coefficient of Variation (CV) values               of all 17 settlement options, grouped by time period.................................202 Figure 7.3.    Comparison of the Coefficient of Variation values (CV) by settlement,             arranged chronologically from oldest to youngest (Period I-III),                 showing a ‘V-shaped’ pattern and trends through time .............................202 Figure 7.4.    Comparison of variation in housepit areas with box-plots showing     intra-settlement variation and relationships between settlements     between Periods I, II and III settlements compared to the earlier     periods -- Period III settlements having the widest range of     intra-settlement variation in house sizes ....................................................204   xivFigure 7.5.    Schematic plan showing the linear, double-rowed layout of housepits at Sxwóxwiymelh I and II, noting the direction of access from the                    Fraser River (arrows) .................................................................................207 Figure 7.6.    Schematic plan showing the linear, single-row layout of housepits at Shxw’ow’hamel I and II, noting the direction of access from an                adjacent slough channel (arrows)...............................................................208 Figure 7.7.    Schematic plan showing the segmented, double-rowed, linear layout of                housepits at Hiqelem (total), noting primary routes of access                                             from the Harrison River (arrows)...............................................................209 Figure 7.8.    Schematic plan showing the segmented, double-rowed, and linear layout                 of housepits at Th’ewá:lí, noting probable plankhouse feature locations              and primary routes of access from Sweltzer Creek (arrows) .....................210 Figure 7.9.    Centroid analysis showing a ‘front-back’ arrangement in the plan of                  housepits at Th’ewá:lí, based on size, noting primary routes of                            access (arrows) ...........................................................................................211 Figure 7.10.  Schematic plan showing the single, row linear layout and conceptual                 ‘front-back’ arrangement of housepits at Qithyil Island, noting primary          points of access from the Harrison River (arrows); also depicted are          housepits with square, circular, and rectangular/oblong shapes ................212 Figure 7.11.  Schematic plan showing the double-row ‘C-shaped’ or ‘flanking pattern’          layout of housepits at John Mack Slough surrounding a probable              plankhouse feature, noting primary routes of access from the                      Harrison River (arrows) .............................................................................213 Figure 7.12.  Centroid analysis showing the ‘front-back relations’ in the plan                                 of housepits at John Mack Slough, based on size, showing access                  routes (arrows) ...........................................................................................214 Figure 7.13.  Schematic plan showing the (slightly curvi-) linear layout of housepits                   at Sqwa:la paralleling Hope Slough; also depicting the generally                circular shapes of these features.................................................................215 Figure 7.14.  Schematic plan showing the (slightly curvi-) linear layout of housepits                  at Eyxel, overlooking the Fraser River, opposite Welqámex.....................215 Figure 7.15.  Schematic plan showing the linear-row ‘C-shaped’ or ‘flanking pattern’             layout of housepits at Xelhálh surrounding a probable plankhouse              feature(s), noting the primary route of access to this hill-top                        settlement from Xelhálh Bay (arrow) ........................................................216 Figure 7.16.  Centroid analysis showing ‘front-back relations’ in the plan of housepits                  at Xelhálh, based on size, showing direction of access (arrow).................216 Figure 7.17.  Schematic plan showing the linear-row ‘C-shaped’ or ‘flanking pattern’              layout of housepits at Welqámex I and II surrounding a probable             plankhouse feature, noting the primary routes of access to this from                      the Fraser River (arrows); also noting square, circular, and     rectangular/oblong-shaped features ...........................................................217 Figure 7.18.  Centroid analysis showing ‘front-back / center-side relations’ in the                     plan of housepits at Welqámex I and II, based on size, showing                   primary directions of access (arrows) ........................................................217   xv Figure 8.1.    Box-plot of the distribution of roofed area measurements for all                              17 Upriver Group settlements/occupations -- defining a normal shape                 and showing the breakdown of settlement size classifications                  coinciding with each quartile .....................................................................230 Figure 8.2.   Comparison of roofed area values by settlement indicating the  trajectories of change and increasing diversification of settlement                      sizes through time, over Periods I, II, III ...................................................232 Figure 8.3.   The distribution of roofed areas by settlement size class sorted by                      Periods I, II, IIIa, and IIIb, showing increasing variation through time.....232 Figure 8.4.   The distribution of roofed areas by settlement size class showing                increasing variation between pre- and post-1400 cal B.P. periods,                     from a commonly large group of early and late period Class II and    III settlements (noting DiRj-1i [max. option] as uncertain).......................233 Figure 8.5.   Comparison of roofed area values and housepit frequencies per settlement/occupation with a regression line showing a positive               relationship between settlement size and housepit frequency and              defining both common and divergent (pre- and post-1400 cal B.P.)             forms of housepit settlement composition .................................................234 Figure 8.6.    Spatial relationships between settlements/occupations differentiated                          by size and set within the landscape of the Upriver Group, showing a             strong association between very large settlements and central                     locations (hubs) of communication and transportation..............................238 Figure 8.7.   Spatial relationships between Upriver Group settlements differentiated                       by size class and other upriver housepit settlements classified by                numbers of features, suggesting a set of ‘hub-satellite’ relationships              between small-to-medium sized settlements surrounding      Very Large settlements occupying central places in the    ‘communication landscape’ .......................................................................243 Figures 9.1a-d.  Results of key analyses illustrating changes in housepit / settlement          relations through time ................................................................................248 Figure 9.2.    A schematic framework showing individual trajectories of housepit    and settlement differentiation through time ...............................................250 Figure 9.3.    A framework of stratification and changes in the vertical plane of    relations through time expressed in heterarchical and    hierarchical terms .......................................................................................255 Figure 9.4.    A comparison of ethnographic and archaeological models of Stó:lō-    Coast Salish social stratigraphy and organization (ethnographic     model figure based on Suttles 1985:12, per Carlson 1997b:90;     archaeological model figures based on Figure 6.18) .................................259 Figure 9.5.    Suttles’s ethnographic model of social organization adjusted to fit the archaeological Late Period framework of social structure                            (based on Figure 6.18)................................................................................260 Figure 9.6.    A framework of community development and changes in the horizontal          plane of relations through time ..................................................................270 Figure 9.7.    A social-spatial model of community relations ............................................272   xvi Figure 9.8.    A political-economic model of inter-and intra-group relations and  community organization.............................................................................278 Figure 9.9.   A political-economic model of inter-community spatial relations                   conjoined as a network by the Fraser River communication and          transportation system .................................................................................279   xviiGLOSSARY OF HALQ’EMÉYLEM TERMS  Halq’eméylem is the upriver dialect of the Halkomelem language spoken by the Stó:lō (Galloway 1993).  Translations of the following terms used in this dissertation are, except as otherwise noted, derived from linguist Brent Galloway’s (2004) Dictionary of Upriver Halkomelem.  S’ólh Téméxw - ‘Our World’ or ‘Our Land’ Stó:lō -  ‘river’ (with particular reference to the Fraser River); also “the River of Rivers” (as told to Charles Hill-Tout by a Stó:lō informant; Carlson 2003:55); this term is used in self-identification as ‘People of the River’ (Duff 1952; Carlson et al 2001; Carlson 2003). s’iltexwáwtxw - plankhouse. sqémél - pithouse. si:yám - respected leader, chief, upper class person, boss, master (plural form sí:yá:m). smelá:lh - wealthy, respected person, high class person; wealthy person who knows their history (Suttles 1987:12, 14 and Carlson 1997:90). s’téxem - worthless person, someone who has forgotten their history (Suttles 1987:12, 14 and Carlson 1997:90). skw’iyéth - slave. Ts’elxwéyeqw - as far as you can go with a canoe (Elder Albert Louie in Oliver Wells 1987:160); the indigenous name of the Chilliwack River and people of the Chilliwack River watershed. Yewal Siyá:m - highest ranking siyá:m; primary leader.   xviiiACKNOWLEDGEMENTS I first and foremost thank my Family, especially Lara, Kiah, and Sasha, for their support, encouragement, and patience.  I also recognize my parents -- Peg and Jim, and my brothers, Chris and Steve. I don’t have enough thanks to offer Mike Blake, Dana Lepofsky, Bruce Miller, and David Pokotylo -- my committee members -- for the incredible amount of help, commentary, encouragement, commitment, and general camaraderie they provided throughout the course of my Ph.D. program.   I wish to acknowledge and thank a number of my colleagues, particularly Keith Carlson, Sonny McHalsie, Anthony Graesch, Bill Angelbeck, Jeanne Arnold, Mike Lenert, Sue Formosa, Kisha Supernant, Tracey Joe, and Tia Halstad for their friendship, input, and thought provoking discussion over the years. R.G. Matson provided very helpful commentary on a draft of this dissertation, for which I am grateful.  I thank Sue Formosa, who authored and co-authored a number of the maps included in this dissertation. Herb and Helen Joe and their family, and Gwen Point provided what we might call general support for my well-being, for which I am indebted to them. I received a huge amount of support and assistance from a large number of people associated in many different ways with the Fraser Valley Archaeology Project, through which nearly all of my fieldwork for this dissertation was carried out. Support from the Stó:lō Tribal Council was provided by Grand Chief Clarence Pennier, President; likewise from the Stó:lō Nation by Chief Joe Hall, President.  I am grateful for their interest in, enthusiasm for, and dedication to archaeology as a means of coming to better understand Stó:lō culture and history.  I am also constantly inspired by their integrity, fortitude, and commitment to defending Stó:lō aboriginal rights and title, on behalf of their communities. I thank the many Stó:lō Chiefs, Council members, Community members, and administrative staff who provided input, assistance, support, and approvals for various aspects of the Fraser Valley Archaeology Project, including: Chawathil First Nation - Chief Ron John and Councilors and Band Office staff-members Peter John, Tim Peters, Ruth Peters, and Monica Florence; also Barb and Larry Pete for hosting work at Sxwóxwiymelh which is situated on their CP lands; Shxw’ow’hamel First Nation - Darren Jones, Roger Andrews, Melanie Andrews, Dean Jones, Felix Jones, Darren’s cousin Sean, and members of the Sí:yá:m Council between 2004-2005 for providing access and encouragement to investigate the settlements at both Xelhálh and Shxw’ow’hamel;  Soowahlie First Nation - prior Chief Larry Commodore and Grand Chief Doug Kelly and their respective Councils, including - in particular - Nelson Kahama; also Dan Bisaillon who helped us with the site clearing; Scowlitz First Nation - prior Chief John Pennier; also Allen Williams, and particularly Betty Charlie and Cliff Hall for their parental support and assistance in the field and boating us to Qithyil Island; Skwah First Nation = the late Chief Roy Mussel for allowing us to access and map Sqwa:la; Chehalis First Nation - Chief Willy Charlie, prior Chief Alex Paul, James Leon, and Gordon Mohs, in particular, for inviting to us to connect with his project and permitted our work at John Mack Slough and Hiqelem; Seabird Island First Nation - Elder and Grand Chief Archie Charles, June Harris, and Dwayne McNeil helped with logistics and access to their lands during the initial stages of the Project.   Field assistance was provided by Albert ‘Sonny’ McHalsie, Riley Lewis, Larry Commodore, Dennis Leon, Leeanna Rhodes, and Ian Franck of the Stó:lō Nation Treaty Deapartment // Research Department // Research and Resource Management Centre; Darren   xixJones (Shxw’ow’hamel First Nation); Denise Douglas (Cheam First Nation); and the students of the SFU 2004 Archaeology Field School, and the UBC and SFU 2005 Field Schools, including: 2005 = (from UBC) Kisha Supernant (a very special thanks!), Nick Waber, Marnie Recker, Chris Marchant, Kate Jessup, Patricia Ormerod, and Adrian Sanders; (from SFU) Sonja Aagesen, Hannah Baker, Cinnebar Bertelsen, Melissa Blain, Debbie Castagner, Sandie Dielissen, Mathew Fladmark, Steve Hamm, Marina La Salle, Shana Morin, Amanda Palmer, Sarah Prien, John Sheppard, Chris Springer, Christine Wright, and Corylee Zanatta, as well as Ian McKechnie, Peter Locher, and Morgan Ritchie.   2004 = Meagan Cameron, Brendon Gray, Jarin Hutchinson, Jennifer Jones, Heather Livingstone, Jessica Ruskin, Craig Rust, and Sarah Swayze (from SFU) -- including all the administrative support staff at UBC and SFU who helped out with these fieldschools. An enormous “thank you” goes to Stó:lō Hihiyeqwals Gerald George, Francis Phillips, and their assistants, and all those who helped out in the kitchen as cooks and otherwise helped in preparations for our Burnings - including Yvette John, Ida John, Nikki LaRock, Norma Gabriel, Tia Halstad, and Tracey Joe. Expert boatsmen who provided logistical support for our fieldwork and field trips include Albert ‘Sonny’ McHalsie, Tim Peters Sr., Vince Malloway, Willy Charlie, and Clifford Hall. Chawathil First Nation hosted us graciously at their Telte-Yet Campground in Hope.  Their fabulous Campground Hosts include Viola John, Sharon Blakeborough, Tanya Alex, Patrick Inyallie, and Robert Guiterrez.   I thank Jim Pike of the Provincial Archaeology Branch for administering to the permit for work at DiRi-48. I also acknowledge the generous support and consultation provided to this project by the faculty and staff of the Department of Geomatics at BCIT.  In particular, Bob Harrower and David Martens warrant huge thanks for making available their time, expertise and expensive, high tech equipment to the Fraser Valley Project.  They provided us with a Leica 705R Total Station during the 2004 field season and three Leica 1200 GPS units and two Leica 705R Total Stations during the 2005 field season.  These supplemented the Leica 705R and Garmin GPS unit provided by the Lab of Archaeology, at UBC and the Garmin 76S provided by Stó:lō Nation -- whom I also whole-heartedly thank.  Jerry Maedel, Faculty of Forestry at UBC, provided guidance to the Fraser Valley Project regarding the use of ArcGis.  I am grateful for the help provided by staff at CN Rail, especially for assisting with the georeferencing of our map data at Xelhálh and Shxw’ow’hamel.  Funding to support my research was provided by SSHRC grants No. 752-2003-2318 (Ph.D. Graduate Fellowship) and No. 410-2003-1525 (Fraser Valley Archaeology Project).  I take full responsibility for the presentation of all the data and figures, and all the interpretations and conclusions presented in this dissertation; any errors there-in are mine as well.  My apologies if I happened to forget to mention anyone by name in providing my thanks; they are surely in my mind.   xxDEDICATION   I dedicate this dissertation to my to immediate family Lara, Kiah, and Sasha; those who I consider my extended family T’xwelátse, Naxaxalhts’i, Ey:lí:seleq, and Siyemches; and a few of my long-standing mentors Howard Winters, Eugene Boesch, Owen Lindauer, Knut Fladmark -- for their trust, patience, guidance, teaching, tolerance, humor, inspiration, and friendship.  1CHAPTER I     INTRODUCTION  This study explores the development of Stó:lō-Coast Salish community organization using archaeological data.  I focus on house and settlement patterning in examining the long-term history of pre-colonial community formation among indigenous Stó:lō-Coast Salish peoples of the Pacific Northwest, British Columbia.  Major themes of this case-study originate from issues currently faced by indigenous peoples worldwide struggling against colonial authorities and imposed political economic structures in efforts to recognize and gain legal recognition of their identities and communities (Cormaroff and Cormaroff 1991; Miller 2003; Tveskov 2007).  These efforts center on defining the nature of inter- and intra-community structure and organizational relations of authority and power, particularly between indigenous communities and the state.  Archaeology provides a unique way of developing a history of change connecting present and past forms of community organization among the Stó:lō-Coast Salish, more commonly discussed within ethnographic and historical perspectives.  Some Canadian Northwest Coast indigenous communities are currently exploring opportunities for returning to self-defined systems of governance and expressions of political economic authority by escaping the constraints of federal Indian Act control (Boxberger 1989; Harmon 1998; Harris 2002; Miller 2001, 2007; McHalsie 2007; Schaepe et al. 2008; Thornton 2001).  Such efforts commonly aim to regain political economic authority as a means of regaining a level of self-determination unfettered by state determination, as they now define the meaning of ‘Indian’ in Canada.  The Stó:lō are currently engaged in various types of negotiations with the federal and provincials governments over issues of governance, lands, and resources -- that is, the control of relations between people, places, and things.  These ‘control’ relationships are central to this study as it explores relations of authority in pre-colonial Stó:lō-Coast Salish society over a 3,000 year period.  This study links past and present in common   2anthropological themes and current aboriginal issues regarding indigenous community formation.  Past and present merge in a mix of disciplinary and cultural perspectives, sometimes rather awkwardly, in the attempts to define the future of relations between aboriginal communities and persistent colonial forces.  A confluence of shared research interests by Northwest Coast indigenous peoples and anthropologists, particularly ethnographers and archaeologists, provides dynamic grounds for dialogue on issues of identity, governance, and inter-community relations.  Notions of community pervade this dialogue.  Amidst varied perspectives of legal, anthropological, bureaucratic, and lay-persons debates around these widely-affective issues, there remains no clear or single answer to the question of ‘what constitutes a community?’    1.1 Community A broad literature and range of perspectives has developed from the anthropological and sociological treatment of ‘community’ over the last sixty years (Anderson 1983; Bourdieu 1977; Cohen 1985; Delanty 2003; Habermas 1984, 1987; Murdock 1949; Redfield 1955; Tonnies 1963; Turner 1969; Wolf 1956).  Definitions of community range from the fixed, determined, and bounded to the negotiated, imagined, symbolic, communicative, and porous relationships between people; from functional and organic to ideological and politically motivated entities.  Communities are part of what make us human.  They are uniquely human entities of communication and belonging (Delanty 2003).  I define community within a political-economic framework as forming an integral part of, but not completely determining, the organization of society.  Society is a larger entity and set of relations affected by the   3organization of political-economically based communities, but also informed by a wider range of interactions such as those established around language, warfare, and trade and exchange. A primary assertion of this study is that political economic relations in small scale, pre-state societies affect the formation of communities, linking community with authority.  The negotiation of power within a realm of political economy is a fundamental element of pre-state community formation and organization, developing around central figures of authority.   I incorporate social theory from Pierre Bourdieu’s Outline of a Theory of Practice Theory (1977) and William Roseberry’s Political Economy (1989) in defining community within a political-economic context based on relations of power, linked to material remains of the past.   I treat ‘community’ as actively constructed and constituting a potentially wide-ranging set of political-economic relations stemming from households and extending to relations within and between settlements across a broad political-economic landscape.  Community in this view is neither passively determined by environmental constraints nor intrinsically limited to an intra-settlement, face-to-face set of social interactions.  Community organization is embedded in the physicality of the built cultural landscape.  This study requires a theoretical framework that correlates material remains with socio-political relations.  Developing social and political theory addressing community formation, and integrating community- and settlement-based archaeology are two central objectives of this thesis.    I recognize communication as a fundamental element of community organization (Delanty 2003; Habermas 1984, 1987); that interactions affecting community formation are not limited to the defined and bounded spaces of houses or settlements.  This study examines networks of communication and transportation routes associated with precontact modes of travel throughout the region as a physical infrastructure affecting community formation (Ames 2002; Gorenflo and Bell 1991; Miller 1989a; Schaepe 1999, 2001a; Tromold 1991).  I focus on   4these communicative aspects of the landscape in exploring community organization as expressed through intra- and inter-settlement patterning.    The term ‘community’ is a recent and still contested addition to the vocabulary of regular archaeological use (Canuto and Yeager 2000; Kolb and Snead 1997; MacEachern et al. 1989; Rogers and Smith 1995; Wills and Leonard 1994).  How can archaeologists bring these disparate notions of community into their study of past forms of social organization based on material remains, and beyond that, is it necessary?.. why would they want to?  Not only can ‘community’ be incorporated, but it must to be incorporated into archaeology.    Archaeologists must deal with community in addressing power and political economy as core processes affecting the development and change of social order.  Discussions concerning the evolution of community organization and social structure extend far beyond the anthropology of the Coast Salish and Northwest Coast culture areas.  Questions concerning the nature and evolution of political structures within small- and intermediate-scale, pre-state societies have a long history of treatment by researchers in contexts throughout the world, including: Arnold (1996a, 1996b, 2004) among the Chumash of southern California; Blake and Clark (1999) among the formative cultures of Coast Chiapas, Mexico; Hayden (1992, 2000) among the Stl’atl’imx of the British Columbia Plateau; Steponaitis (1978), Muller (1997), Mehrer (2000), and Pauketat (2000) among the Mississippian societies of the American Mid-West; Feinman, Lightfoot, and Upham (2000) and Redmond (1998) among the Puebloan communities of the American Southwest; Earle (1997) of the South Pacific Hawaiian islanders; and Carniero (1970) in the Andean region.  This study aims to contribute to this dialogue.       51.2 Community and Settlement Archaeology Bringing ‘community’ into archaeology adds political and social theory to a foundation of settlement archaeology defined nearly 40 years ago by Bruce Trigger (1967).  Trigger (1967:151) defined ‘settlement archaeology’ as “the study of societal relationships using archaeological data.”  Settlement archaeology has goals beyond culture history, requiring social theory.  Social theory addresses synchronic (structural) and diachronic (developmental) aspects of cultural relationships (Schiffer 2000).  It stands against understanding aspects of social relationships as traits within complexes of archaeological cultures.  Cultural relations are, rather, understood as “functioning systems of economic, political, and affective relationships” (Trigger 1967:151).  Trigger used the terms system, network, ethnicity, social boundaries, identity, and community in relation to settlement archaeology as a wide-ranging field of study.  The definition of these terms remain relevant and contested within anthropology, archaeology, and in the reckoning of relations of power and authority, today, between Nation states and indigenous peoples around the world.  An aspect of settlement archaeology adopted in this thesis is the aim of establishing a quantitative foundation supporting the application of social theory.    Alternate models of authority and political-economy generating ‘corporate,’ ‘network,’ and ‘heterarchical’ models of political organization, however, remain largely unapplied within this debate -- and within broader application to the Northwest Coast (Crumley 2001; Feinman 2000, 2001; Feinman and Nicholas 2004; Haas 2001).  These recent developments in social theory challenge social-evolutionary ‘chiefdom’ concepts that have persisted since emerging in the 1950s (Yoffee 1993).  Emerging social and political theory helps integrate community and settlement archaeology.   6The house, itself, is increasingly recognized as a basic medium and means by which households relate to one another, defining social relations, and affect community organization (e.g., Sobel 2006).  Houses, for example, act as powerful influences on and of people in the course of their interactions, embodying symbolic capital as ‘the material structure of political action’ (Marshall 1989).  I associate the construction of Stó:lō houses with the negotiation of relations of power and authority.  Tracing political-economic relations archaeologically through the analysis of house features and settlement patterns provides insight into the formation of indigenous, pre-colonial communities among the Stó:lō.    1.3 Housepits: A Proxy to Political-Economic Relations and Community Formation  As a proxy for socio-political units associated with intra- and inter-settlement scales of relations, pithouses (i.e., in-ground houses) provide an ideal unit of analysis in investigating community formation.  Hundreds of individual housepits (sqémél)1 associated with at least 112 recorded settlements (as of 2005) mark the landscape of the mainland Gulf of Georgia Region with a record of in-ground house construction over 5,000 years old (Lepofsky et al. in press; Mason 1994; Schaepe 1998, 2001b, 2001c, 2003).  I assume, as have many archaeologists engaged in household archaeology in the Northwest Coast and elsewhere, that these archaeological remains of houses correlate with household social units (Arnold 2004; Blanton 1994; Coupland 1988, 1996; Hayden and Cannon 1992; Lepofsky et al. 2000; Matson 1996, 2003a, 2003b; Price and Feinman 1995; Wilk and Rathje 1982; Schaepe 1998).  Feinman (2000) and others (Feinman et al. 2000:453) identify house form and size as significant archaeological indicators of social organization among intermediate societies ranging from what they classify as ‘corporate’ to ‘network’ forms of social organization.                                                    1 Sqémél is the term for pithouse in the halq’eméylem language.  See the Glossary of Halq’eméylem Terms for translations of all halq’eméylem terms used in this dissertation.  I italicize all halq’eméylem terms except those used frequently, such as settlement names.   7Drawing on the analysis of housepits in villages in the U.S. Southwest, Feinman et al. (2000:453) correlate variability in house size and form with status differences between household units and with a network form of social arrangement.  Consistently large or homogenous house sizes within and between settlements suggests corporate organization.  On the Northwest Coast, Ames (1995, 2006) identifies the importance of households in the study of social organization, equating households (Ames 1995:156) with the polity and noting that the investigation of the household is crucial to the investigation of status differentiation.   Thus, housepits, as the remains of in-ground houses, represent an archaeological proxy for the investigation of household units and by extension, socio-political-economic organization.    Arnold (2004:173) suggests that “by tracking emergent hierarchical organization alongside this corporate network dimension, we may being to understand the evolution of pithouse communities.”  The archaeological remains of Stó:lō sqémél are an ideal archaeological subject matter for tracking community formation through time.  Households represent a basic unit of interaction central to the process of community formation, particularly in light of their importance among the Stó:lō-Coast Salish.  As Miller (2007:20) points out, “unlike the northern matrilineal tribe of the northern Northwest Coast culture region, the Coast Salish have no clan system, with the result that distribution of members of family groups is not clearly channeled.  This fact points to the importance of examining individual, household, nuclear family, “family”… processes of affiliation.”   While households live together as basic residential family group, the political ties of those individuals may extend to members of other households in other settlements.    Most Northwest Coast household archaeology focuses on the remains of plankhouses (s’iltexwáwtxw) (Sobel et al. 2006).  The Fraser Valley offers the rare opportunity on the Northwest Coast to benefit from exploring the remains of pithouses.  In some cases, co-existent   8pithouses and plankhouses appear to have been built at the same settlement (Barnett 1955; Lepofsky et al. in press).  I correlate residential families or households with housepits -- as opposed to other kinds of houses (e.g., surface-built plankhouses) that leave less tangible signatures and recognizable imprints in the ground -- and are thus easy to identify archaeologically.  Sqémél represent the discrete living quarters of an individual household group.  S’iltexwáwtxw, alternately, represent the collective living quarters of a large if not complete set of households occupying a settlement.  The archaeological footprints of s’iltexwáwtxw do not reveal discrete household quarters from their surficial form, as do sqémél.    I regard sqémél as an external representation of household relations otherwise internalized within s’iltexwáwtxw.   Housepits represent a more spatially discrete rendition of the social order found, also, within co-existent plankhouses.  The same set of inhabitants occupied both forms of houses, perhaps seasonally or perhaps more regularly throughout the year.2  Wealth is also a factor associated with the construction of sqémél in this region.  Barnett (1955:55) states, “Not every family owned or had access to skameL (sic.) since its construction was costly.  It was mainly used by the weak and the infirm who, once in, stayed there most of the time; others resorted to it only during the coldest periods.  There was no general abandonment of the plank house.”  In settlements with co-existent sqémél and s’iltexwáwtxw, the occupation of s’iltexwáwtxw by a collective of extended families (Suttles 1992:214) is less spatially differentiated than the occupation of sqémél by more discrete household units (Mitchell and Donald 1988).  Household organization is thus more readily visible, archaeologically, as expressions of relationships between sqémél.    Another benefit of housepits, representing households, is that they are visible on the landscape and easily plotted in space as a means of exploring settlement patterning.  Housepits                                                  2 Multi-season occupation of sqémél is implied by Graesch (2006) as a result of his detailed investigation of housepits at Welqámex.   9sizes are measurable and comparable in exploring household patterning.  The material remains of their use and occupation often times provide the material required for determining their age and positioning them in time.  Sqémél, thus, constitute the ideal proxy for looking at these potentially complex sets of relations across space and time.     1.4 Relations between Ethnography, Archaeology, and Stó:lō-Coast Salish Community Organization    “What is important is that settlement archaeology forces us to think through problems from a new angle -- that of social relations” (Trigger 1967:158).  Trigger viewed settlement archaeology as an important bridge between archaeology and ethnography, with historic and ethnographic data serving as strong supports to settlement archaeology.   This remains true in the application of a developing community archaeology (Marcus 2000).  The broad temporal scope of this study -- spanning precontact and post-contact times -- bridges archaeological and ethnographic data sets and mediates the ‘contact barrier’ that commonly truncates the application of these disciplinary approaches to the investigation of Coast Salish society.   This direct cross-over between realms of ethnography and archaeology brings up issues regarding in the relationship between disciplines and the use of ethnographic interpretations and models in this study.    1.4.1 Archaeology and the use of Ethnographic Interpretations and Models    Ethnographic research of the last two decades (Miller 2007) supports Wayne Suttles’s declaration of “authority” (Suttles 1989:251) as one of most important topics in contemporary Coast Salish anthropology.  Anthropologists, archaeologists, and ethnohistorians continue to engage in vigorous debate over leadership structure as manifest in the investigation and discussion of inter-community, socio-political relations among the Coast Salish and other   10aboriginal Northwest Coast societies (Ames 1995, 2001; Burley 1980; Carlson 2003; Martindale 2003; Miller 1997; Miller and Boxberger 1994; Mitchell 1983a, 1983b; Mitchell and Donald 1988; Schaepe 2006; Tollefson 1987).    The use of ethnographic analogy within archaeology is a long-standing tradition on the Northwest Coast, not without issue (Grier 2007).   While there are exceptions (Coupland, Martindale, and Marsden 2001; Miller 2007; Moss and Wasson 1998), the relationship between ethnographers and archaeologists working in the Northwest Coast is seldom one of direct collaboration.  Over the last 40 to 50 years, Northwest Coast anthropology has moved further away from a multi-field anthropological approach and achieved a significant degree of separation between anthropological sub-disciplines.  Recent commentary rightfully draws attention to problems associated with ethnographic analogy in Coast Salish archaeology and suggests alternate ways of relating ethnographic and archaeological inquiry as parallel endeavors (Grier 2003, 2007).      I advocate for connecting and developing inter-disciplinary relations, integrating archaeology and ethnography (as well as ethno-history and oral history).  Integration, when possible, works toward achieving a more complete understanding of issues and questions shared between our sub-disciplines.  Such integration provides a potential means of resolving issues that arise from the use of ethnographic interpretations as a basis of analogy in archaeology.    Ethnographic models can and do provide a useful basis for interpreting archaeological data.  For example, I use Miller’s (1989) definition of a corporate “family” group as a dynamic political unit that escapes the more conventional understanding of households as spatially bounded in the scope of their interaction and political affiliation.  Grier (2003) incorporates the concept of a corporate family into his archaeological discussion of extra-household ties affecting Coast Salish interaction throughout the coastal Gulf of Georgia Region.  Community   11relations radiate outward, particularly in light of the negotiation of family-corporate group relations from a foundation of household relations without inherent boundaries.  This said, the application of ethnographic data and conclusions drawn from those data in archaeological interpretation require scrutiny.  Small-pox epidemics and European colonial powers have in many ways affected the nature of Northwest Coast and Coast Salish societies over the last 150 years.  The application of conclusions about Coast Salish authority, for example, drawn from ethnographic data applied back in time to describe socio-political organization 300, 500, or 1,500 years ago is highly suspect in light of post-contact effects on socio-political organization.  The relationship between pre- and post-contact political organization among the indigenous peoples remains unclear.     At issue is the use of ethnographic analogy as accurately depicting a critical period in Stó:lō-Coast Salish pre-contact history that is poorly represented both ethnographically and archaeologically -- this being the post-Marpole period between 1,500 and 150 years ago.  This 1,500 year period currently stands bridged, across the Gulf of Georgia Region, by the use of ethnographic analogy linking that period of the past with the ethnographic present as a static ‘Developed Northwest Coast Culture’ (Matson and Coupland 1995).  I compare the results of ethnographic and archaeological data in reference to Wayne Suttles’ (1987:125) ‘village organization model’ -- discussed below -- as a means of exploring the relationship between ethnographic and archaeological interpretation.  The use of this model aids in understanding the relationship between ethnographic and archaeologically-based studies of authority, and as well as the process of community formation and change leading up to, and as affected by, European contact.         12 1.5 Defining Terms and Developing Questions about the Effects of the Colonial Period    A number of questions addressed in this study are embroiled in the events and impacts of the colonial period on indigenous Northwest Coast societies.  Significant impacts on Stó:lō-Coast Salish peoples, specifically, resulted from colonial expansion into the Pacific Northwest in the late 1700s.  The massive depopulating effects of the small-pox epidemic of 1782 (Carlson 1997a; Harris 1997, 2002; Lamb 1960), interaction with Hudson’s Bay Company traders in the early 1800s (McLaughlin 1998), and the subsequent influence of British colonial authority following the gold rush of 1858 -- associated with the implementation of assimilation policy and reserve creation leading to the establishment of the Indian Act in 1876 (Harris 2002; Tenant 1998) -- all affected indigenous political-economic structures.  Responses to the impacts of these events were surely accounted for by divergent motivations and organizational forces -- ranging from the internal and indigenous to the external and colonial.  Colonial forces of the later 1800s consciously re-shaped the present day organization and relations of authority among the Stó:lō with the intent of minimizing their economic, socio-cultural, and political-economic structures, status, and access to capital (Boxberger 1989; Harmon 1998; Harris 2002).  The British government employed an assimilation policy strategically targeting “aboriginal institutions and life patterns” (Thornton 2001:12).  Stó:lō-Coast Salish reaction to colonial powers also affected their expressions of identity (Carlson 2003; Harmon 1998).  I examine how present-day expressions of identity -- bound within relations of authority and community formation -- relate to those of the precontact past from which they emerged.  These historic events provide a framework for defining precontact, pre-colonial, and colonial periods of time.    I use the term ‘precontact’ to describe that period of time preceding direct contact with Europeans, circa 1800.  Thus, in precontact times before 1800, the Stó:lō had no direct contact with or influence by Europeans, although they were badly affected by a small-pox epidemic just   13prior to that point, circa 1782 (Harris 1997).  Between 1800 and 1850, the Stó:lō developed direct relations with Europeans while yet acting, I maintain, within an indigenous framework of interaction and authority.  I use the term ‘pre-colonial’ in reference to this period of time, preceding the subjugation of aboriginal peoples of British Columbia by British colonial administration and governance.  I establish a cut-off point for the termination of the pre-colonial period at circa 1850, as supported by historians and cultural geographers familiar with the region (Carlson 1996; 2003; Duff 1965; Harris 1997; 2002).  Shortly after 1850, Stó:lō governance and independence was overtaken by Colonial rule.  The commencement of the colonial period in British Columbia is closely associated with its incorporation as a British colony in 1858.  These three events define the basic temporal framework of my study.  Questions about the extent of the effects of contact and colonization underlie the anthropological investigation of Stó:lō-Coast Salish social organization.  Anthropological interpretations of this issue, in turn, affect the definition of Stó:lō rights being negotiated in land claims currently taking place within the political and jural realms of Canada and British Columbia.  Colonial governments consistently institute a ‘local level authority’ model within treaty and litigious environments as a means of limiting and narrowly defining the scope of indigenous peoples’ rights and title. (e.g., the Vanderpeet decision of 1995).  Communities are defined in narrow and bounded terms.  These interpretations are extracted from anthropological interpretation such as Wilson Duff’s (1952:85) assertion that Stó:lō settlements were of a small scale and rarely housed more than 50 people.    A central issue in this contest regards governance in the broader framework of indigenous-State relations.  Tension in developing contemporary indigenous governance frameworks hinges on the relationship between the generally small-scale, spatially bounded, independent corporate units of the Indian Act-structured First Nations of today and the   14indigenous expressions of their political economic systems as they existed prior to contact.  Central to addressing questions emerging from this issue is the relationship between ethnographic and archaeological perspectives on the topic of authority.   1.6 Ethnographic, Archaeological, and Ethnohistorical Discussions of Households, Villages, and Socio-Political Organization among Northwest Coast and Central Coast Salish Peoples   It is necessary to point out that political organization among aboriginal Northwest Coast societies has been only “skimpily and unsystematically treated” (Drucker 1983:86) in the anthropological literature and ethnographic documentation of the area; so too for the Coast Salish.  I draw attention to what has, over the past century, become a prevalent although not unanimous anthropological voice and set of perspectives.  A predominant voice generated from a review of numerous ethnographic works between the 1940s and 1990s tends to characterize the economic and political organization of traditional Northwest Coast-Coast Salish societies as limited to intra/inter-household and/or intra-village level developments.  I reference these below.  Consideration of a broader range of theoretical perspectives beyond that of the long-dominant and somewhat persistent culture-ecology (e.g., Earle 1997) is badly needed.    Jorgensen (1980) characterizes Northwest Coast tribes as engaging in hierarchical distributions, noting that “…the salient feature of the organization of distribution among Northwest Coast kinship groups were the distributions themselves.  These were intra-village and inter-household affairs attending birth, naming, puberty, marriage, death...  these distributions were among the principal mechanisms for local exchanges of goods… Distributions which, more than likely, were expanded to an inter-village scale during the period of contact with Europeans became known by the Chinook jargon word ‘potlatch’..” (Jorgenson 1980:145; emphasis added).  Jorgenson, like some Coast Salish ethnographers and   15ethnohistorians (Carlson 2001a; Suttles 1975; Tollefson 1987), takes interest in and applies ethnographic analogy to reconstructions of past population densities.  Ethnographically based statements like Jorgenson’s that “there is good reason to believe that Northwest Coast populations based solely on extraction could have been larger than those found at first white Contact” (Jorgenson 1980:162) are consistently voiced (e.g., Carlson 2001c).  Archaeologists typically agree with this point (e.g., Ames and Maschner 1998; Huelsbeck 1988).   It is also generally understood across disciplines that “.. resources were abundant throughout these areas and watercraft and waterways were available to transport goods and people.  People from many communities convened to eat vast quantities of food, and on the Northwest Coast enormous amounts of movable property were circulated” (Jorgenson 1980:162; see Ames 2004; Ames and Maschner 1998; Mitchell 1971; Schalk 1977; Suttles 1968a, 1968b).  “Yet” states Jorgenson “political and economic organizations capable of maintaining large populations never developed” (ibid; emphasis added).  The certainty and ‘global’ application of this statement as accurately describing highly diversified Northwest Coast peoples spanning thousands of years of cultural occupation and change is, to say the least, suspect.  This statement re-phrased as a question - did political and economic organizations capable of maintaining large populations develop [among the Stó:lō-Coast Salish]? - resides at the heart of this study and relates directly to Suttles’s reckoning of ‘authority’ as an issue central to Coast Salish anthropology.  Archaeologically, ‘cultural complexity’ stands as a proxy for treating issues of ‘authority.’ Archaeologists have engaged in the investigation of the authority on the Northwest Coast for nearly four decades -- embedded within archaeological investigations of the development of social complexity.  While numerous definitions of ‘complexity’ abound, Arnold’s (1993:77) definition of ‘complex’ as ‘hierarchical political organization on a multi-  16community scale’ is commonly accepted and applied in this field of study (see Matson 2003a:6).  Arnold suggests a standardized definition of ‘complex’ (addressing political and social organization issues) as distinguishing “those societies possessing social and labor relationships in which leaders have sustained or on-demand control over non-kin labor [on a sustained basis] and social differentiation is hereditary” (Arnold 2000:78; 2004:180; see Ames 2001).  ‘Complexity’ (1) “first and foremost means institutionalized control by some people over non-kin labor” and (2) “hereditary inequality and leadership” (Arnold 2000:93) meaning ascribed versus achieved status.  Others view complexity as associated with property ownership, high population density, and political relations that extended beyond the residential kin-group (e.g., Jorgensen 1980).  Archaeologists generally agree that social stratification and ascribed or inherited social status are commonly recognized as achievements of indigenous peoples within the last 2,000-1,500 years, typified by the Marpole Period within the Gulf of George region (Ames and Maschner 1998; Matson and Coupland 1995).  Some argue for its much earlier manifestation (Carlson 1996).  Ames (2001) provides a strong argument for the emergence of slavery by 500 AD, though possibly manifesting earlier, associated with a Coast-wide shift in social relations.  He positions slave labour as critical to the labour supply needed to support elite, corporate groups with ascribed status and ranked households, lineages, and villages that depended on the conversion of labour to wealth to prestige.  This argument links to issues of indigenous authority as control over non-kin labour and hereditary social differentiation factor, also, in Arnold’s (2000) definition of complexity.  These changes in community formation appear archaeologically as changes in the expression of prestige and status display possibly linked to the development of a plank house-based potlatch-type interaction sphere which carried on into post-contact times.     17 I situate Jorgenson’s work as a representative ethnographic voice in this discussion -- a compliment to the thoroughness and significance of his broad body of work and wide-ranging scope of analyses.  Two other elements of Jorgenson’s conclusions are important to present.  He states,   …the impressive force of private property…  in shaping the size of local communities in otherwise heavily populated areas cannot be denied.  In the absence of more complex political centralization, property-owning communities were small and independent” (ibid; emphasis added).  Abundances of resources, either those that occurred in nature or those that were produced from farming, correlated positively with population density but not with political and economic organization.  That is to say, predictable, storable, and abundant food supplies generated the highest population densities in western North America, but these factors were not sufficient to generate complex and centralized political and economic organizations (Jorgenson 1980:164; emphasis added).    Also, the general patterning of political organization indicates that,  …the simplest and most numerically dominant political organization was the residential kinship-group that had no formal ties with any other group, even if several such kinship groups lived side by side in the same community (Jorgenson 1980:211; emphasis added).  These basic assertions are commonly though not always consistently represented in the work of many Northwest Coast-Coast Salish ethnographers of the last two centuries from Charles Hill-Tout (1895) to Wayne Suttles (1987) and beyond.  Underhill (1945:174) summarizes the popular position among 19th and 20th century anthropologists that natives of the Puget Sound “had no idea of belonging to a large group beyond his own village of plank houses” and that political leadership was lodged within locally autonomous villages (if not at the household level) and that any higher level organization was a product of European contact.  Tollefson provides a thorough list of references to ‘classic’ nineteenth and twentieth century Coast Salish ethnographic accounts broadly exemplifying these perspectives (Tollefson 1996:147) -- including Kroeber 1917:369; Ballad 1929:35; Spier 1936; Barnett 1938:119; Ray 1939:8-9;   18Smith 1940, 1941:199-203, 1949:13, 1967:68; Collins 1950:334; Elmendorf 1960:308-313; Riley 1974:79; Drucker 1983:87; Onat 1984:89.   From the Northwest Coast anthropology of the 1980s emerged two significant works by Donald Mitchell (1983b) and Phillip Drucker (1983) focusing on the interplay between resources and settlement patterns, particularly among winter village aggregate groups.  These studies mix ethnographic and archaeological data and perspectives.  Drucker characterizes the Coast Salish as having no large, multi-local winter village groupings in their “less favorably situated upriver divisions” (Drucker 1983:93) farther from the Coast.  Drucker concludes that “there were various large aggregations of local groups among some Northwest Coast divisions.  They existed.  But they were not political organizations.  No authority base resided in such a grouping” (Drucker 1983:95) and local identity was retained by maintaining house-based autonomy (c.f. Carlson 2003).   Donald Mitchell, recognizing the seasonal movements and aggregations of villages of Central Northwest Coast-Coast Salish peoples, questions the nature of political structure and “how villages interacted when aggregated” (Mitchell 1983b:103).  His examination of polities within the context of settlement structure is a significant theoretical application of settlement archaeology to the Northwest Coast.  This study crosses-over between archaeological and ethnographic disciplines.  Using minimal ethnographic sources (e.g., Barnett 1955), Mitchell defined three categories of settlement: camps, villages, and village aggregations.  None of these settlement types are presumed occupied or maintained on a year-round basis, but rather seasonally and with a great deal of variation.  Twenty village aggregations, located within an area including a number of regions are defined as occurring throughout all four seasons.  Winter aggregations were most common among the Coast Salish, although only including the coastal Georgia Strait portion of the region, neglecting the upriver areas.  Of Coast Salish village   19groupings, Mitchell states “Although harmony was sought… there was no formal supra-village political order” (Mitchell 1983b:103).  The only apparent rationale for this statement comes from the line that “as if to underscore the absence of a polity, Barnett (1955:18-34) labeled those that joined these winter assemblies “ethnic divisions..” (ibid).  Mitchell continues to expose his logic, stating:  One would not expect a high order political organization to accompany any of the village aggregations.  After all, what was needed was simply peaceful coexistence at some unusually productive resource locus or particularly desirable wintering location.  No great feats of organization were necessary for employment of technology, for gaining access to the resources, for dividing up the resultant harvest, or for simply waiting out the winter.  The constituent village units of an aggregation merely did, side by side, what other village units on the coast were doing in isolation.  The only difference was the proximity of the village (Mitchell 1983b:104).    While recognizing ‘confederations’ among the Nuu-chah-nulth in which some autonomy was given up within village aggregates providing perhaps “an advantage in making war and keeping peace, and.. may even have encouraged a pooling of other important resources” (Mitchell 1983b:106), he concludes that “although most winter groups on the central Northwest Coast spent a least part of the year assembled with others, they did so with no significant reduction in their independence” (ibid.).  Mitchell, thankfully and to his credit as a widely respected researcher, explicitly states his expectations about the nature of Coast Salish political relations (i.e., not expecting high order political organization).  Positions such a Mitchell’s exemplify a predominant anthropological perception that persists today of aboriginal political organization on the Central Northwest Coast-Coast Salish area.  My study calls into question the application of these interpretations to precontact political economic organization and relations among the Central Coast Salish.     20I focus on the work of Wayne Suttles and Robert Elmendorf as providing the most well-developed models of Coast Salish social organization, testable using archaeological data.  Wayne Suttles and Robert Elmendorf’s perspectives on Coast Salish socio-political relations -- based on their culture-ecological view of networks and interaction -- are worth reviewing in some detail as their work remains influential in the use of ethnographic analogy in Coast Salish archaeology, itself heavily influenced by the culture-ecological movement of the 1970s.  I focus on Suttles’ (1958, 1987) ‘village organization model’ as a key influence on anthropological views of Coast Salish socio-economic organization (Figure 1.1).   Suttles’ model typifies the organizational shape of Coast Salish village-based society (i.e., the ‘village organization model’) as an “inverted pear” (Suttles 1987:14) -- similar to West’s (1945) ‘diamond-shaped’ social structure - describing a ‘normally-shaped’ social demographic in which no part of the population exhibits any significant socio-economical or socio-political distinction.  Suggesting that the majority of individuals and families retain some form of ‘high class’ connection, Suttles situates social stratification in Coast Salish society as a phenomenon that was more imagined than real; no real socio-economic boundaries separated ‘Leaders’ from ‘Good People.’  Suttles asserts that local groups were of equal status (Suttles 1968a; 1987:41).    Elmendorf and Suttles both minimize the existence of classes in Coast Salish society beyond a basic ‘free-slave’ dichotomy.  Elmendorf dismisses references of indigenous Coast Salish terms for ‘classes’ describing a continuum of ‘high’ (i.e., ‘Leaders’ and ‘Good People’) to ‘low’ (i.e., ‘Worthless People’) freeman (Figure 1.1; Suttles 1958, 1987:12; also Carlson 1997b:89-90) as “an apparent discrepancy between ideal and actual social distinctions” (Elmendorf 1971:368).  Elmendorf states that, rather than representing real class differences   21“..ranking of individuals and ranking of groups.. were probably general features of Northwest Coast social structure” (Elmendorf 1971:362), with high rank positively correlated to participation in inter-village activities.  “If so, then the relations of ranked statuses within each local community can only be understood through analysis of the total set of intercommunity relations…” from which he presents the hypothesis that, “Coast Salish social rank within village communities depended upon a total set of intercommunity relations within a network specific to each community” (Elmendorf 1971:363).  I incorporate Suttles’ model as a fundamental element of this study, comparing his ethnographic results with those derived from this archaeological study of community formation.  Figure 1.1.  Wayne Suttles’ Coast Salish ‘Village Organization Model’ of stratification in a Coast Salish community.     22 In the 1974 post-script to his 1958 article describing Coast Salish social classes Suttles loosens the ties of Coast Salish social organization, citing extreme inter-village movement and flexibility, to the extent that he suggests avoiding the term ‘community.’  Suttles states that “since it [i.e., ‘community’] often implies a social unit with a high degree of social cohesiveness, in-group feeling, internal social control, and so forth, the Central Coast Salish village may not have been such a unit.  “Village” or simply “settlement” would be more appropriate, since these terms are more neutral in what they imply about social cohesion..” (Suttles 1987:14).  I suggest that Suttles’s view, like Elmendorf’s, of Coast Salish society as a “social network with no clear boundaries” (Suttles 1960:17) was far ahead of its time and is better accommodated within more recent anthropological and archaeological discussion and definitions of ‘community,’ interaction, and political economy (e.g., Anderson 1983; Bourdieu 1977; Canuto and Yeager 2000; Roseberry 1988).   Suttles’ comments on community could be linked into Anderson’s (1983) concept of an ‘imagined’ community.  While the extensive networks of bi-laterally reckoned Coast Salish genealogies may have connected many people to some upper class relative, not all of those people enjoyed the rights and privileges of those that were, versus those that imagined themselves to be, of name-carrying upper class status.  Typically each village had only one ancestral name linked to its origin signifying ‘the man in charge of the resources’ (Suttles 1987:21), thereby creating a significant imbalance between the number of possible and the actual name carriers.  Demand was greater than the supply.  Significant competition surrounded the acquisition and inheritance of high status names, and the rights carried with them, including the ownership of resource sites such as fishing rocks and reckoning of access to resources such as salmon.     23Suttles’s earlier work describes Coast Salish communities or villages as sufficiently bounded to accommodate social stratification within the arrangement of houses.  “The lower class often occupied separate houses in its own section of the community or in a location sufficiently separate so that it might be regarded as a lower class community subservient to an upper class” (Suttles 1960; 1987:17).  He also notes differences in rights and privileges between upper and the lower class families (Suttles 1960).  Suttles states that high status is directly linked to the production of resources.  This assertion links to the inter-household and inter-village production-based ‘task groups’ forming Bruce Miller’s (1989) ‘family corporate groups’ discussed below and implying, perhaps, a level of complexity beyond what Suttles recognized at the time.   Anthropological developments in post-1970s Coast Salish ethnography diverge from Suttles and Elmendorf in their understanding of ‘interaction’ as limited in scope to intervillage ties of affines within a culture-ecological (i.e., economic focused) model of spatially and temporally clumped resources (e.g., Suttles 1960, 1961, 1968).  The emergence of new theoretical perspectives in the 1980s and afterwards served to augment, affect, and build on the anthropological understandings of Northwest Coast-Coast Salish community organization.  Consideration of interaction between corporate groups and households extending beyond affinal ties to include ‘family corporate group’ structures (Mooney 1974; Miller 1989b) affects notions of ‘community’ among the Coast Salish.  Miller’s (1989b) ‘family corporate group structure’ connects directly with archaeological concepts of complexity expressed by Arnold (1996a) as existing beyond intra-household and immediate kin-based relationships.  Founded on working cooperatives comprised of non-kin cohorts, Miller’s social unit moves toward recognizing the centralized control of non-kin labour for the life-span of each corporate family working group -- the maintenance of which was both competitive and political.    24  In a recent work, I employ Miller’s concept as a key factor explaining the linkages between settlements.  Such linkages may be expressed archaeologically, for example, as within an inter-village network of defensive fortifications in the lower Fraser Canyon (Schaepe 2006).  Additional ethnographic support for inter-settlement class differences is provided by Haeberlin and Gunther (1930:58).  Barnett (1955:19, 21, 23, 32-33) describes the importance of house size and placement in the arrangement of houses within a settlement, adding substance to the social-spatial relationships hinted at above.  These descriptions are analogous to the expression of status in a pattern of centrality expressed in the sizes and spatial arrangement of houses on an intra-village scale.  They also allude to a network of inter-settlement relations expressed similarly as social-spatial patterns based on centrality.  Both of these sets of ethnographic descriptions hint at a social-spatial model of inter-household and inter-community arrangements with centrality and spatial distance factoring significantly into socio-political relations.  This field of social-spatial relations was later developed by Kathleen Mooney (1976, 1979) and by Bruce Miller (1989) using quantitative methods to measure centrality in socio-economic and political relations among the Coast Salish.     A theoretical shift among late 1980s and early 1990s post-modernists worked toward achieving an emic understanding of culture involving human agency, breaking away from the culture-ecological paradigm of their predecessors.  Bierwert (1986), Miller (1989), Boxberger (1989), and Miller (1999) were leaders in embarking on this new path.  Focusing on the Stó:lō, Bierwert (1986) covers a wide range of topics including spatial maps, settlement layout, longhouse layout, and directional orientation.  In comparison with Suttles’s environmental approach, she integrates culture and environment within a dynamic process of learning and thinking, as a reproductive (i.e., economic) process.  This shift moves away from the paradigm of understanding based on a ‘separate’ ecological base that has so profoundly the affected the   25anthropological and archaeological views of social development among Northwest Coast societies including views of authority and community organization (Schalk 1977).  “Oral history” she states, “is open to redefinition of the salient links with the past.  The process of interpretation does not quit” (Bierwert 1986:429), bringing dynamism to the traditionally static notion of ‘traditional’ community or society previously sought after within ‘memory anthropology’ -- also providing a counterpoint to those who subsume the formation of aboriginal identity entirely within colonial experience (Oliver 2006).   The physical location of houses and settlements within and between communities is likely to maintain multiple meanings, simultaneously, and be affected by the on-going negotiation of power and agency.  The landscape of settlement patterns, itself, takes on meaning as a topologic landscape of meaning.  The topologic aspect of relations between houses and settlements -- involving the construction and conveyance of knowledge attached to the meaning of things -- delivers us squarely into a discussion of community founded in Bourdieu’s (1977) Practice Theory and Roseberry’s Political Economy (1989). Jay Miller’s (1999) ‘anchored radiance’ model, when viewed apart from the broader, religion-dominated ‘Shamanic Odyssey’ in which it is presented, provides an integration of people, place, and power -- spiritually, economically, and politically -- among the household and settlement patterning of the Skagit-Coast Salish.  Miller builds on the work of Sally Snyder (1964) and Bierwert (1986) in remodeling a topological version of Smith’s (1940) topographic ‘watershed-based systems’ of social relations.  “Within each watershed, group alliances became more expansive in terms of (a) hearth mates eating together, (b) households of all residents, (c) birthright locals - those born there in contrast to in-laws, visitors, and foreigners, (d) settlements and resorts, (e) community networks, (f) tributary drainages, and (g) the entire drainage” (Miller 1999:19).  Culturally, these units, increasing in size, included notions of person (combining   26body, mind, and soul with spirit allies); of house (including what he refers to as ‘hearthers’ -- those who share hearths in their residence --  locals, and distant kin); of canoe (transport across time and space, distinguished as forest, prairie, river, or sea); and of world (the drainage linked both to residential immortals and to more remote peoples and places through marriage, ritual, and trade).   Major nodes in this overall system were cedar plank houses located along the shore near spots rich in local resources.  Beyond the ‘house node’ were at least three concentric rings occupied by allies, by competitors for regional status, and by strangers -- working within a social-spatial distance model of networks and interaction.  By prudently using locally ‘anchored’ resources, a household could add to its regional ‘radiance.’  The crux of the entire system and the basic reason for gathering people together was the display of bonds with particular immortal powers (Miller 1999:19-21).   Miller supposes a settlement hierarchy with the largest settlements (i.e., towns) associated with these interactive hubs, radiating spatially outward up attached drainage systems and diminishing in size (i.e., villages and hamlets).  Prestige and status were attained by engaging in ‘social networks’ beyond one’s own community -- as described by Suttles and Elmendorf (see Kennedy 1995).  Social ranking is considered by Miller as a regional matter (Miller 1999:90), with class sometimes assigned at a community level.  Miller, like Bierwert, recognizes ideation and social ranking within ‘social-spatial’ settlement patterns.  Miller’s model, incorporating political economy, is useful and can be translated into a framework that can be examined archaeologically through the quantification and analysis of house and settlement sizes through time and space.  An application of Jay Miller’s settlement hierarchy model to the Stó:lō is presented in A Stó:lō-Coast Salish Historical Atlas (Carlson 2001a:24-26).     27In recent work, ethno-historian Keith Carlson (2003) identifies the lower Fraser River system of the Stó:lō -- between the lower Fraser River Canyon and the Fraser Delta -- as unparalleled in the social-geography of the Coast Salish world in its capacity to support inter-tribal and inter-watershed transportation and communication.  Carlson identifies the lower Fraser Canyon as a particularly important centre of exchange between the local, salmon rich inhabitants, and those from elsewhere - ‘strangers’ from which economic profit could be drawn and exchanged into social -- and I would add symbolic and political -- capital.  Importance is placed on travel and relations between settlements and watersheds and the control of both resources and communication by individuals of high status.  These relations were primarily acted out upon the avenues of transportation and communication specific to the region (Schaepe 2001a).  Relations between high and low class are characterized by extra-local versus local levels of access and privilege; controlling access and knowledge versus being subject to control; the pursuit of prestige versus that of survival.   Carlson’s view of Stó:lō community as a highly complex series of nested networks leads, again, to a social-spatial model of households and communities with differences of status and wealth expressed in spatial terms.  Recognizing these relationships requires a broad view of the settlement landscape.  These later views of Coast Salish ethnographers and ethnohistorians have not been taken into consideration in the common application of ethnographic analogy used by Northwest Coast archaeologists -- generally still using a Suttles-founded, culture-ecology based set of interpretations guiding our understanding of Coast Salish community organization. The application of the ‘local village’ or ‘local political unit’ model long held in Northwest Coast ethnographic literature cannot be summarily projected into the precontact past.  Within this objection I accept the use of ethnographic models in interpreting archaeological data providing insight into precontact social organization.  Network and communication theory are   28indeed relevant and useful to Coast Salish-Northwest Coast archaeological pursuits (Miller 2007).  The interpretation of archaeological data in this work relies on such ethnographic models, independent of and in comparison to ethnographic data.  The crux of this objection resides in questions about the ‘fit’ of those ethnographic data entered into these models and the products of interpretation as representing precontact, and particularly pre-colonial, forms of socio-political arrangements.  It is necessary to critically evaluate and differentiate between the content of ethnographic conclusions and the particular models developed by ethnographers.  1.7 Defining the Study Area - the Mainland Gulf of Georgia Region  Donald Mitchell (1971) recognized the unique interrelationships between the natural and cultural attributes of the Gulf of Georgia Region that support the use of the term ‘landscape’ in current anthropological meaning (Ashmore and Knapp 1999).  The mainland Gulf of Georgia Region is a ‘landscape’ coincident with the lower Fraser River Watershed of southwestern British Columbia and northern Washington State.  This landscape is unique in at least three critical ways compared with other major watersheds and regions of the Northwest Coast.  First, the area is associated with multiple, large, closely connected watersheds joined by an extensive intra-regional network of transportation and communication leading, via the stó:lō (river; Fraser River), to the Georgia Strait-Coast Salish Sea.  Secondly, the region is home to a range of abundant oceanic, deltaic, riverine, and montane-based resources.  Lastly, unique ecological qualities support highly effective ‘wind dry’ salmon processing and storage techniques factoring into abundant economic opportunities for trade and exchange.  This region is ideally suited to the objectives of this study.    The abundant resources of this environment play a significant role that cannot be underplayed in examining the economic, social, and political development of the Stó:lō, as   29distinct from their Coast Salish and other Coastal neighbors.  The region’s surrounding oceanic, deltaic, riverine, and montane environments provide year-round access to diverse and plentiful localized resources available for harvesting within a predictable and reliable scheduling of yearly activities (Lepofsky et al. 2005).  As a fishery, the Fraser River stands out as the largest salmon-bearing river in the world, traditionally rivaled only by the Columbia River (Northcote and Larkin 1989).  The lower Fraser Canyon provides access to the largest and most important traditional aboriginal salmon fishery north of the Columbia.  The meteorological characteristics of the lower Fraser River Canyon permit the rapid and reliable wind-dry processing of salmon suitable for long-term storage.  The mainland Gulf of Georgia region is distinguished by the Fraser River as a central waterway and artery of transportation and communication within a watershed system unlike any other on the Northwest Coast.  Within the lower 150 km stretch of the lower Fraser River Watershed are approximately 17 major tributary rivers and watersheds.  Most of the other Coastal regions represent the individual watershed systems with fewer and smaller tributary systems (e.g., Skagit River; Skeena River).  Stó:lō-Coast Salish peoples have a long-term history of extensive interaction geographically connected to the Gulf of Georgia Region, extending from the lower Fraser River Canyon, downriver to through the Delta, and across the Georgia Strait to include southeastern Vancouver Island (Burley 1980; Grier 2003; Pratt 1992).    The socio-cultural and environmental attributes of the lower Fraser River Watershed define this area as a central place of longstanding regional interaction among the Stó:lō-Coast Salish.  The network of tributary rivers and watersheds connected by the lower Fraser River -- supporting extensive networks of communication and transportation -- forms a geo-social/political structure well suited to the investigation of community formation through time.    30The archaeological and social-spatial ingredients required of this study are provided by the geo-cultural attributes of this region.    The long-term in situ development of cultural traditions and interactive relations among and between the Stó:lō and their Halkomelem / Coast Salish neighbors also provides an ideal context for investigating community formation.  The long-standing maintenance of this network of relations is supported linguistically by the distribution of Island, Downriver, and Upriver dialects of the Halkomelem language family (Galloway 1993; Smith 2001; Suttles 1990).    The long-term interaction among Stó:lō-Coast Salish peoples of the Gulf of Georgia Region can be demonstrated archaeologically.  Common ancestral material relations are recognizable as the ‘Charles Culture’ -- more accurately reiterated in archaeological terms as the ‘Charles Horizon’ -- signifying the widespread distribution of similar archaeological features and artifacts throughout the region between 5,500-3,500 years ago (Borden 1975; Carlson 1994; Pratt 1992).  The archaeological record provides evidence of significant trade and exchange among these Coast Salish peoples throughout pre-contact history (Blake 2004; Brown et al. 2008; Carlson 1994; Carlson 1996; Clark 2000; Grier 2003; Lepofsky et al. 2000; Mitchell 1963).  The ethnographic record and contemporary history account for a continuity of interaction among and between Stó:lō-Coast Salish peoples (Carlson et al. 2001; Duff 1952; Elmendorf 1971; Pennier 2002; Suttles 1987, 1990).  Archaeologists generally agree on a process of long-term in situ cultural development in the region since, if not preceding, the Charles ‘Horizon’ of approximately 5,000 years ago (Burley 1980; Clark 2000; Mitchell 1971; Matson 1976).  My study area begins at the Fraser River Delta and extends approximately 240 km eastward and northward along the Fraser River to a point including the lower Fraser River Canyon, about eight kilometers north of the town of Yale, B.C. -- the eastern edge of the region   31marking the transition, culturally and environmentally, from the coast to the interior.  Mitchell (1971) recognized the lower Fraser River Watershed as an element of the Gulf of Georgia Region -- a unique natural, ethnographic, and archaeological region (Figure 1.2).  This area can be sub-divided into five interconnected though geographically and environmentally distinct locales.    Figure 1.2 The Gulf of Georgia Region within the Central Coast Salish Culture Area.   Moving downriver from east to west, the lower Fraser River Watershed includes the lower Fraser Canyon (the Canyon), Upper Fraser Valley, Central Fraser Valley, Lower Fraser Valley, and Fraser Delta (Figure 1.3).  Each sub-section of this region has slightly different characteristics creating a spectrum of differences from east to west such that the environments of the two geographic extremes -- the Canyon and the Delta -- are very different from one another.  The eastern extent of the region, at the upriver limit of the Canyon, defines the connective edge between Coastal and Interior Salish peoples.     32 Figure 1.3.  Geographic sub-sections of the Mainland Gulf of Georgia Region / lower Fraser River Watershed; FC = lower Fraser Canyon; UV = Upper Fraser Valley; CV = Central Fraser Valley; LV = Lower Fraser Valley; and D = Fraser Delta (Lepofsky, Schaepe, and Blake 2006).   I use those geographic definitions developed by Lepofsky, Schaepe, and Blake (2006) as most accurate and effective at incorporating indigenous Stó:lō perspectives describing the physical characteristics of the region’s land base and waterways (Figure 1.3).  The lower Fraser River Watershed is shaped like funnel, with its constricted neck representing the Canyon, to the east, and its flared opening beginning in the Central Fraser Valley and opening to the sea at the western margin of the Fraser Delta.  The Fraser River flows southward from the Interior through the steep-sided rocky gorge that is the Canyon.  The river is navigable from the ocean to the abrupt entrance (i.e., downstream end) of the Canyon, above which seasonal rapids restrict access.  The traditional mode of transportation changes from consistently river-based to a more mixed form of river- and land-based travel upriver of the entrance to Canyon.  From that point, the river flows through the Upper Valley which remains a narrow though heavily forested   33valley; after abruptly turning west (near the ‘UV’ label in Figure 1.3) it levels and widens as it grades into the Central Valley.   Through the Central Valley, the floodplain broadens significantly and is braided with many intertwining river and slough channels past the town of Chilliwack to the western toe of Sumas Mountain.  At this point, the Cascade Mountains to the south give way to a broad coastal plain (the Puget lowlands) with Coast Mountains to the north continuing to bound the Lower Fraser Valley, through which the river flows in a more singular channel forming an estuarine environment affected by tidal waters of the ocean yet some 70 km to the west.  The Delta, beginning at the foot of the uplands, is an ancient alluvial fan of Fraser River sediments deposited and expanding westward over the last 5,000 years (Clague, Hebda, and Luternauer 1983; Schaepe 2001c:16-19).  The Delta is flat and broad with two main branches of the Fraser River that traverse the Delta before connecting with the oceanic waters of the Georgia Strait (the Coast Salish Sea).     1.8 Defining the Terms Stó:lō and Coast Salish  Throughout this thesis, I use the term ‘Stó:lō’ in reference to the indigenous people inhabiting the lower Fraser River Watershed, i.e., the mainland Gulf of Georgia ‘Landscape,’  roughly coincident with my study area (Figures 1.2 and 1.3).  The Stó:lō represent a collective of indigenous Halkomelem-speaking peoples.  ‘Stó:lō’ is a word in the Halkomelem language literally meaning ‘river’ with particular reference to the Fraser River, but also used in self-identification by these people as ‘People of the River’ (Duff 1952; Carlson et al. 2001; Carlson 2003; Schaepe 2007).  Historical records provide evidence of this term being used in self-identification going back to the late 1800s (e.g., as used in the title of the St. Mary’s Roman Catholic “Stó:lō hymnals” of the 1880s).  Not everyone in all First Nations communities along   34this stretch of the river currently identify themselves as Stó:lō.  Individuals frequently identify themselves within a range of identities based on family, Band/First Nation, and/or broader collective units including tribal (e.g., Ts’elxwéyeqw) or supra-tribal (i.e., Stó:lō) affiliations -- oftentimes situationally and sometimes simultaneously.    Some degree of confusion also currently exists between cultural entities (i.e., Stó:lō ) and a number of contemporary political / service delivery agencies that use the name Stó:lō  (e.g., Stó:lō  Nation; Stó:lō  Tribal Council).  The Stó:lō, as I refer to them here, are collectively associated with a cultural group of indigenous people known in anthropological terms as the ‘Coast Salish’ (Figure 1.4).  The Stó:lō maintain close family, linguistic, and cultural ties to their neighbors and relatives (e.g., Nooksack, Squamish, Skagit) living throughout the broader Gulf of Georgia and Puget Sound regions of the south-central Northwest Coast (Suttles 1990).  The Stó:lō call their homeland S’ólh Téméxw (‘Our World’ or ‘Our Land’).    1.9 Chapter Outline and Descriptions  This thesis includes 10 chapters moving beyond this introduction into discussions of theory and methods, followed by specific analyses, and interpretive and concluding chapters.  In Chapter II, I define my theoretical approach to this study, informing the framework which I apply to the interpretation of my findings.  I discuss the context of the study as drawing from and incorporating elements of settlement archeology, interaction theory, practice theory, and community archaeology.  I discuss anthropological perspectives on definitions of ‘community’ leading to an essay on the application of practice theory and political economic theory to the interpretation of findings on housepit features and settlements.  Lastly, I develop a set of expectations with which to compare and interpret my data and findings.   35  Figure 1.4. Map of the Central Coast Salish area showing the locations of principal villages in the early 19th century (Suttles 1990:454). (Note: This map does not show all Central Coast Salish tribal groups, languages, and First Nations within the region, and a number of local groups are not identified. The main objective of figure is to illustrate the geographic relationship between speakers of Halkomelem and their immediate Central Coast Salish neighbors).    36 Chapter III deals with the data collection methods used in addressing the data gaps and methodological issues affecting the regions’ housepit and settlement data at the outset of this study.  I describe methods used in mapping and testing of housepits and settlements in my study area as the primary means of gaining control over those dimensions of space and time critical to this study.  These methods describe the collection of new spatial and temporal data from 114 housepit features within 11 settlements acquired as part of the Fraser Valley Project between 2004-2006 (Graesch 2006; Lepofsky et al. 2003; Lepofsky et al. 2005; Lenert 2008; Lenert and Lepofsky 2005, 2006; Sanders and Ritchie 2006; Schaepe, Blake, Lepofsky, and Formosa 2006).  I also describe the basic, exploratory quantitative analytic methods used in this study.  Chapter IV involves the broadest scope of analysis in this stud, using housepits settlements as a unit of analysis as distributed throughout mainland Gulf of Georgia Region.  I deal with the issues of space and address questions about the locations of the 112 housepits settlements recorded in study area as of 2005.  I define the region as a network of transportation and communication routes, both riverine and terrestrial.  Within this ‘transportation and communication framework’ I quantify travel distances between each recorded settlement and the Georgia Strait.  I use an Exploratory Data Analysis (EDA) -- defined in Chapter III -- as a quantitative approach to exploring, organizing, and analyzing these data, providing a means of defining groups of settlements by their location along the region’s transportation corridors.  This analysis also serves to assess long-term trends in settlement patterns defining where the Stó:lō established their settlements within this geographic and interactive framework.  These results provide reference and context for the more detailed analyses of both time and space, including a sample of housepits and housepit settlements, conducted in Chapters V-VIII.  Chapter V deals with issues of time and serves to answer questions about the age of housepit features and settlements in the study area.  I use both new and existing radiocarbon   37data to define a temporal framework spanning 3,000 years between A.D. 1850 and 800 B.C., divided into four major periods.  This framework serves to situate in time, either directly or indirectly, each the 114 housepit features included in this study.  In Chapter VI, I focus on housepits features as independent units of analysis.  Central questions addressed in this chapter focus on variations of size, shape, and age as attributes of housepit features.  Using EDA, I quantify the relationships through time and space of my sample features classified and sorted by variables of size, shape, and age as indicators of socio-political differences between households.  I extrapolate and compare household populations per feature across housepit features.  These analyses inform discussion of household size and organization as it developed through time.  Chapter VII expands the scope of analysis to the level of individual settlements.  Retaining housepits as the basic unit, analyses in this chapter address questions relating to variation through time of the spatial relationship between housepits and households within settlements as a basic layer of community organization.  Incorporating findings from the preceding chapters focusing on regarding age and size, I define and compare intra-settlement housepit arrangements.  These analyses inform discussion of community plans and social organization within settlements as they developed and changed though time.  Chapter VIII returns to the broadest level of analysis comparing settlements and communities through time across the entire region.  This chapter culminates as a synthesis of findings in Chapters V-VII, with controls over time and space re-applied to those regional settlement patterns identified in Chapter IV.  I address questions regarding variability in inter-settlement relationships through time based on size and community populations.  Positioning settlements as my unit of analysis I explore and classify settlement sizes and arrange them in time as discrete occupations.  Similar to the analyses in Chapter VI, I extrapolate and compare   38community populations per settlement/occupation.  These analyses inform discussion of the regional system of relations between settlements as they developed and changed among the Stó:lō over the 3,000 years preceding Colonization.  Chapter IX includes a discussion and interpretation of my findings within the intellectual and theoretical frameworks set out at the beginning of this thesis.  This chapter provides archaeological insights into the organization of housepits, households, settlements and communities within mainland Gulf of Georgia Region.  My goal is to address questions about the nature of pre-colonial Stó:lō community structures and political organization, examined in terms of corporate and network forms of socio-political organization.    I conclude in Chapter X by revisiting my original research questions, summarizing my findings, and evaluating the strengths and weaknesses of my interpretations.  I suggest areas for future research highlighted by topics of my study.  I link my findings to the broader field of anthropological questions and theories, as well as to the current interests of the Stó:lō communities, related to the utility of archaeology and the investigation of issues concerning authority and community organization.     39CHAPTER II - LINKING COMMUNITY AND SETTLEMENT ARCHAEOLOGY  2.1 The Socio-Political Taphonomy of Community: Defining Agents of Formation and Change   Understanding community within an integrated community-settlement archaeology framework requires defining the ways in which community is expressed in material form.   Pierre Bourdieu’s (1977) Practice Theory and William Roseberry’s (1988) Political Economy provide two sources of theory central to developing an integrated community-settlement archaeology aiding in deriving meaning from the material world of the past.  Bourdieu’s ‘practice theory’ and ‘habitus’ (Bourdieu 1977, 1980, 1994) have been recently adopted by a number of archaeologists addressing identity (Dietler and Herbich 1998; Robb 1998), regional or settlement studies (Mackie 2003), households (Coupland 2006), and cultural practices (Pauketat 2001).  Citation of Roseberry remains essentially absent in the archaeological literature, despite the applicability of his work and approach to mediating materialist and ideationist perspectives.    The benefit that both Bourdieu and Roseberry offer archaeologists is their recognition that people make investments of knowledge and symbolic capital in material culture.  This theoretical step -- growing out of Sahlins’ (1976) Culture and Practical Reason -- helps mitigate conflicts between materialist and symbolic theoretical camps recognizing ‘ideals’ and ‘values’ as having economic ‘praxis’ and ‘practical reason.’  The material culture of houses, as forms of symbolic capital and investments of households, relate to the production of household prestige derived from relations of power and authority.  Relation of power and authority act as central elements of community formation that can be measured along vertical and horizontal planes of relations (Blake and Clark 1999; Crumley 1987; Feinman 2000).    Linking houses with symbolic capital requires examining the way in which ideals and values are created within the production of knowledge; a system which integrates material and   40political economies (Cohen 1985; Delanty 2003).  I present, below, eight essential ‘taphonomic’ elements of community -- framed as a political economic system -- derived largely from Roseberry (1988, 1989) and Bourdieu (1977, 1994).  Central to the definition of community applied in this study are: knowledge and economy, time and strategy, power and habitus, and tradition and competition, negotiated within a political economy.  The dynamic process of mediating tensions inherent in these interrelated elements of community is a generative process of community formation, integrating material and reason, expressed as a factor of transformation.   2.1.1 Knowledge and Economy    Knowledge is an essential ingredient in the production of meaning and therefore integrally connected to the economy of the meaning of things, including houses, through which expressions of community are manifest.  Bourdieu’s (1977) practical mode of knowledge, as described in Practice Theory, is a fundamental element of community formulation.  Knowledge, as a central tool of structuring agents of household, settlements, and communities, is achieved through action and produced within an economic framework; economic in the sense of being produced, distributed or exchanged, consumed, and reproduced.  These four sequentially inter-linked actions represent the basic economic cycle.  The production of knowledge, akin to Sahlins’s practical reason (1976), is recognized by Bourdieu as a basic element in the generation of culture.  The generation of knowledge occurs in interactions and relations between individuals and broader social collectives.  Herein lies an essential assumption of community formation -- knowledge, and thus culture, is relational, it is interactive, and in this sense is objectified.  While individuals can produce and consume knowledge, they cannot fulfill the distributive or reproductive aspects of the economic cycle.  Thus, knowledge must be shared   41and reproduced between groups of individuals in order to be realized in cultural practice.  Sharing and reproducing knowledge requires communication (Habermas 1981, 1987). The generation of knowledge (i.e., the economy of practical knowledge), and thus the formulation of culture, is generated within an economic process negotiated through the interactions and relations between people; between individuals; between individuals and collectives -- as within the formation of communities.  As a relational system, the economic process requires interplay between at least two people, i.e., a group.  While individuals can produce and consume, they cannot fulfill the distributive or reproductive aspects of the cycle.  Individuals, in failing to fulfill the economic cycle, fail to account for culture.  Thus, knowledge is discursive; it must be shared, in practice, between individuals in order to be realized (Foucault 1972:182-1853).   From this basic set of relations stems a basic Bourdieuian-influenced concept of community as a group (or collective) of people engaged in practicing an economy of shared knowledge (i.e., shared culture).    Households are basic groups of people active as agents of practical knowledge influencing community formation.  In application, the knowledge of in-ground house construction and the meaning of those houses is shared and practiced in the construction of housepits, and ‘housepit communities,’ within and between settlements.  It is upon the recognition and maintenance or reproduction of a community, like that manifest in the social-spatial relations between households and their houses, that the economies of knowledge and culture are dependent.    Knowledge is not purely unconscious, structured in a fixed and deterministic way, mechanistic and without self-awareness as it is often portrayed by Structuralists, Structural Marxists, Functionalists, and Evolutionists (Foucault 1972; Sahlins 1976).  Practical knowledge is a matter of self-reflection among those who are conscious of it and who pose “the questions   42of the conditions which make... knowledge possible” (Bourdieu 1977:4).  Knowledge, thus, is not a representation of human experience; it is a combined product of human self-awareness, agency, and action.  Among those conscious of how knowledge is produced, it becomes a tool in the negotiation of social relations and standing.   Relationships between individuals, action, and time all form essential ingredients of strategy in the mechanism of community formation.  Individuals are agents of practical knowledge and action (Pauketat 2001).  Action moves from ‘agency’ (conceptualized; conceptual knowledge; conceptual choice; ideational; imagined) to ‘agents’ (realized; actual knowledge; actual choice; practiced choice; manifest) by way of practice; particularly the control of practice.  This relationship between control and practice is played out between individuals and a broader collective community as captured in the concept of habitus (Bourdieu 1977), discussed below.  The ‘reproductive’ element of the economic cycle fosters another key element (also linked to habitus) – that of time.  If knowledge accounts for structure, form, and behavior as manufactured and practiced through an economic process, time provides the medium within which knowledge is practiced and structure is replicated and restructured.     2.1.2 Time and Strategy  Time itself, as linked to knowledge, is manufactured in relation to the economic cycle.  Archaeologists, for example, manipulate, manufacture, and reproduce concepts of time (Smith 1992).  Through time, changes in such things as house form and location relative to other houses within and between settlements represent changes in the relations between individual households, their actions, and the production of knowledge.  Such changes indicate an alteration in the strategy of relations between households by which knowledge is produced and manifest as symbolic capital within house structures and settlement layout.  The appreciation of a   43Bourdieuian sense of time affects how we understand the place of strategy in interaction and exchange.  He states:  ..everything takes place as if the agents’ practice, and in particular, their manipulation of time, were organized exclusively with a view to concealing from themselves and from others the truth of their practice, which the anthropologist and his model bring to light simply by substituting the timeless model for a scheme which works itself out only in and through time… To abolish the interval is to abolish strategy…. Time derives its efficacy from the state of the structure of relations with which it comes into play; which does not imply that the model of that structure can leave it out of account… To restore to practice its practical truth, we must therefore reintroduce time into the theoretical representation of practice which, being temporally structured, is intrinsically defined by its tempo. …To substitute strategy for the rule is to reintroduce time, with its rhythm, its orientation, its irreversibility (Bourdieu 1977:6-9).     Thus, the concept of tempo is linked to practice as an emic and meaningful measure of time and economic practice.  This definition contrasts with ‘science’ and anthropological models wherein ‘time’ is linked to ‘rules’ and is de-temporalized and divorced from practical reason; divorced from knowledge; divorced from a self-aware strategy.  It can become lost in a processual view of the longue durée (see Ames 1991).  The practical mode of knowledge is characterized by strategic implementation and realization, utilizing time as a tool.  The economy of practical knowledge is marked by an intrinsic tempo of production, exchange, consumption, and reproduction.  The strategy of the agent in this economy, e.g., the household, is linked to the tempo of exchange in practice.  Tempo is a production linked to a strategy; a strategy linked to purpose.  Controlling the tempo of change is comparable to controlling an essential ‘rule of the game’ (Bailey 1969).  For what purpose, then, what ‘game’ supports an economy of knowledge and culture?  The answer to this question links largely to competition for prestige achieved through contests of power.       442.1.3 Power and Habitus  Power is a factor of influence on the economy of knowledge and culture and is necessary in converting labour to wealth to prestige (Arnold 1993, 1996).  Power influences the negotiation of authority and provides purpose and motivation to practice.  Power is conceptualized within agency (e.g., the idea of a sqémél) and maintained by an agent (e.g., the household) and manifest through practice (e.g., materializing / building a sqémél).  Power situates individuals among a collective as, for example, manifest in the construction and placement of a larger house by a more prestigious household relative to a lower status household within the same community (Matson 1996).  Power is manifest as an influence on the economy of knowledge and culture actualized, through practice, as an accepted form of material culture among many alternate possibilities -- for example, as in the persistence and reproduction of square sqémél versus other possible shapes (e.g., circular).  Power is manifest as influencing community formation -- in both intra- and inter-community relations -- in the spatial arrangement of houses within a settlement and the establishment of one settlement versus another near a ‘powerful’ place (e.g., resource patch; spiritually significant feature; junction of transportation and communication).  The attributes of power must be commonly recognized.  Power must be enacted in order to be recognized and made effective in shaping community organization.  Houses provide a very effective medium for enacting power as physical manifestations and markers of social-spatial relations speaking to household status and power (Bourdieu 1973, 1994; Coupland 2006).  To have power, then, is to control the means of production in the economy of culture and knowledge.  Culture, incorporating knowledge, becomes the means of production.  Competition among the collective becomes the mode of production.  From an investigative standpoint, economic interaction is an indicator -- with strong material correlates -- of elite /   45non-elite relations recognized as a continuum (Blake and Clark 1999; Clark and Blake 1994).  Elite / non-elite relations are tied to the definition of ‘community’ and measures hierarchical differentiation between individual actors (e.g., households) and collectives (e.g., settlements; local residential communities) defined through social-spatial distance (Miller 1989a; Miller 1999).  Power binds knowledge, culture, and economy with individuals and collectives as the motivation for interaction.  Relations of power between households are expressed in material form as social-spatial relations between houses, both within and between settlements (Pauketat 2000).   Thus ‘political economy’ is founded on a platform of economic process influencing knowledge, culture, and time, through strategy, exerted by powerful actors aiming to achieve prestige and honour.  But, how is practice theory played out?  What are the mechanism(s) of such interaction?  What are the formative processes by which communities develop?  A final and significant element of this theory is ‘habitus.’  ‘Habitus’ is, in short, the product of the mode of production and a fundamental element of intra-community relations.  It is worthwhile quoting at some length some of Bourdieu’s commentary on habitus.  He states: … The homogeneity of the mode of production of habitus (i.e., of the material conditions of life, and of pedagogic action) produces a homogenization of dispositions and interests which, far from excluding competition, may in some cases engender it by inclining those who are the product of the same conditions of production to recognize and pursue the same goods, whose rarity may arise entirely from this competition.  The domestic unit, a monopolistic group defined… by the exclusive appropriation of a determinate type of goods (land, names, etc.) is the locus of a competition for this capital, or rather, for control over this capital… (Bourdieu 1977:63).  He follows up this definition of habitus, and precedes others to follow, by stating: [it is necessary] to pass beyond the necessity of the realism of structure [methodological objectivism] by passing from opus operatum to modus operandi – from statistical regularity to the principle of the production of the observed order… to “construct a theory of practice, or, more precisely, the theory of the   46mode of generation of practices, which is the precondition for establishing an experimental science of the dialectic of the internalization of externality and the externalization of the internality, or, more simply, of incorporation of the objectification (Bourdieu 1977:72).  Here we see the interplay and negotiation of the dialectic relations defining the ‘practical mode of knowledge.’  He continues:  The structures constitutive of a particular type of environment (e.g., the material conditions of existence characteristic of a class condition) produce habitus, systems of durable, transposable dispositions, structured structures predisposed to function as structuring structures, that is, as principles of the generation and structuring of practices and representations which can be objectively ‘regulated’ and ‘regular’ without in any way being the product of obedience to rules, objectively adapted to their goals without presupposing a conscious aiming at ends or an express mastery of the operations necessary to attain them and, being all this, collectively orchestrated without being the product of the orchestrating action of a conductor…. always tending to reproduce the objective structures of which they are a product, they are determined by the past conditions of which have produced the principle of their production, that is, by the actual outcome of identical and interchangeable past practices, which coincide with their own outcome to the extent that the objective structures of which they are the product are prolonged in the structures within which they function (ibid., emphasis added).  To paraphrase Bourdieu and others (Maton 2008), habitus refers to a way of acting and valuing things affected by a dynamic of (1) internal social structures, including the acquisition of knowledge and capital, and (2) external forces acting upon them.  Habitus is perpetuated through time as tradition -- somewhere between fixed and fluid behavior.  I emphasized a section of this quote that strongly relates to the idea of cultural tradition -- the ‘folk’ element of society.  This “common sense world” (Bourdieu 1977:80) is fabricated but is practiced and reproduced in what ranges from conscious to unconscious practice between people of different social standing and place within the relations of power and authority.   The relationship between the individual, their capital and place in society, and external forces acting upon them is know as the ‘field’ (Bourdieu 1984; also see Elmendorf 1971).  The field is a convergence; a ‘central’ realm where both influential and influenced forces meet and are negotiated as the practice of political-economy.  The outcome of negotiations situates the   47actor(s) within community.   Housepits, for example, represent an inseparable material and ideological capital that, as constructions of habitus and full of meaning (Bourdieu 1973), are factors of negotiation within the field.     That individuals strive to distinguish themselves from the collective through hierarchical position recognized as differences in power is another essential assumption of community formation.  Powerful individuals or households -- elites -- establish this economy as both informing and as limited by habitus.  Those with power influence this economy while mediating a tension between both practicing culture and generating habitus.  Again, the attributes of power must be commonly accepted among the collective.  Individual reckoning of those attributes does not constitute a collective culture. Those lacking power are relegated to greater degrees of practice and lesser degrees of influence serving to reinforce habitus.  The benefits of power are linked to prestige and honor.    Gramsci’s transcendent “man-in-the-mass” (Gramsci 1971; Roseberry 1989:46) -- the one who ‘knows thyself’ (i.e., linked to knowledge) and transcends the collective ‘folk’ -- moves into the position of negotiating ‘community’ and affecting history.  History relates, in this sense, to Wolf’s (1982) world of interconnectedness.  The transcendent individual -- as powerful, wealthy, and of elite status -- is linked to the community through kinship relations that form the means and relations of production in the ‘natural’ or ‘pre-hegemonic’ political economy (Blake and Clark 1999; Clark and Blake 1994; Meillassoux 1981).  Individual-and-collective relations, as linked by Roseberry to community, are also linked to historical process(es) that interrelate locally, regionally, and perhaps more broadly -- as within and between households; within and between settlements; and within and between watersheds throughout the cultural landscape of S’ólh Téméxw (the Stó:lō world of the lower Fraser River Watershed).       48The level of consciousness within which culture is practiced depends upon political position and provides a means of accessing and manipulating power relations through networks of inter- and intra-settlement relations on numerous scales.  The level of political interaction escalates with the intensification of production, in Sahlins’s terms (1972), corresponding with the shift between household to corporate and network modes of relations (Crumley 1987, 2001; Feinman 2000).  Variation in the attributes of housepits and housepit settlements are attributed to differences of social standing between households, within and between local residential communities.  From a broad view, these variations provide a basis for describing social arrangements as corporate or network modes of relations, and issues of political interaction. It is natural that politics should be the privileged arena for the dialectic of the official and the useful: in their efforts to draw the group’s delegation upon themselves and withdraw it from their rivals, the agents in competition for political power are limited to ritual strategies and strategic rituals, products of the collectivizing of private interests and the symbolic appropriation of official interests (Bourdieu 1977:40-41).  Power, political status, and wealth are linked as aspects of habitus differentiating households (or settlements) collectively recognized as elites from that range of households (or settlements) otherwise lacking such recognition. “Wealthy people work to reproduce relations [interested in acquiring honor and symbolic capital -- tied to material economic exchange]; poor people settle for the ordinary practical [marriage] relations available to them” (Bourdieu 1977:40).  Miller (2007) recognizes these differences in Coast Salish society distinguishing sí:yá:m (wealthy, respected leaders with unblemished ancestry, extra-human support, and extra-village relations) from the ‘rank and file.’3  “People affiliate with the idiom of kinship strategically based on calculations of social distance to the siyá:m and the resources they control and distribute” (Miller, personal communication, 2008).  The ‘tool-kit’ and options of relations                                                  3 The halq’eméylem term sí:yá:m (i.e., leaders) is the plural form of siyá:m (i.e., leader).  Both forms of this term are used throughout this dissertation.   49between sí:yá:m and those lacking such status is certain to vary, affecting strategy for maintaining, attaining, or closing the social-spatial distance to ‘siyá:m’ as a central and significant idiom defining Coast Salish community.   From an archaeological view, the association of housepits with measurements of household status is expected to manifest in differences in size, shape, and location within and between settlements.  Social-spatial relationships between houses act as expressions and indicators of various levels of political economic standing, including ‘eliteness.’  Within Coast Salish society, for instance, the construction of larger houses is not simply an outcome or expression of material economic success based on controlling access to labor and resources (Ames 1995).  Rather large houses are important products of an active and consciously controlled knowledge production system and used as part of a strategy for acquiring prestige, accessing material resources, and gaining control over labor (Coupland 2006; Hayden 1994; Matson 1996).  Recognizing, influencing, and navigating the political-economic landscape are critical to success.  These measures differentiate households and settlements of higher or lower prestige, greater or lesser wealth, and more or less power and influence within the community.  Meaning is thus produced and attached to houses whose traces manifest today in the form of depressions on the landscape. Those of elite status work to manufacture and reproduce habitus, while those on the lower scale of ‘rank and file’ carry on the practice of habitus as tradition; tradition being the common bond between them maintaining their connection within ‘community.’  Elites work to establish forms of symbolic capital in things such as house architecture and placement that convey meaning which sets them apart from the rest of the group, as documented ethnographically among the Coast Salish (Barnett 1938, 1955; Haeberlin and Gunther 1938).  This system of production is dynamic and affected by relations with others of the same elite   50class as well as those of lower classes.  Elites and those of lower standing alike have agency to affect and change things either directly as sí:yá:m (recognized leaders) of powerful and dominant constituencies (kin- and extra-kin network) or otherwise as leaders of resistance groups vying for power.  Symbolic capital must be recognizable among elites as representing elite status, serving to distinguish them from others of lesser standing.  It must simultaneously be acceptable to, supported by, and reproduced by the other members of their communities from whom elites attempt to distinguish themselves.  The symbolic capital of elites must therefore appeal in some way to the ‘rank and file’ or otherwise risk exposure, resistance, and rejection (Miller 2007).  Elites hold political power and to some degree modus operandi (i.e., generative principles); non-elites, as those practicing opus operatum (i.e., objective structure), don’t.  Thus, “each agent, wittingly or unwittingly… is a producer and reproducer of objective meaning” (Bourdieu 1977:79).  In short, the habitus, the product of history, produces individual and collective practices, and hence history, in accordance with the schemes engendered in history.  The system of dispositions – a past which survives in the present and tends to perpetuate itself into the future by making itself present in practices structured according to its principles, an internal law relaying the continuous exercise of the law of external necessities – is the principle of the continuity and regularity which objectivism discerns in the social world without being able to give them a rational basis (Bourdieu 1977:82).   Lastly,   The habitus is the product of the work of inculcation and appropriation necessary in order for those products of collective history, the objective structures (e.g. language, economy, etc.) to succeed in reproducing themselves more or less completely, in the form of durable dispositions, in the organisms (which one can, if one wishes, call individuals) lastingly subjected to the same conditions, and hence placed in the same material conditions of existence (Bourdieu 1977:85).   Habitus, thus defined, is an important concept with powerful application to archaeology, bringing forward explicit definitions of relations between tradition and continuity; innovation   51and change; collective unconscious and individual free-will; symbolic meaning and material economy; cognition and social structure; action and identity; culture and hegemony; and interaction and community organization.  All of these relationships are expressed and objectified as material culture, integrated into identity formation, and played out in the relations of material phenomena through time and across space.     2.1.4 Tradition and Competition  Tradition -- including values and social norms of meaning and behavior -- is set by the dominant power mediating and affecting the relationship between elements of the community, produced and reproduced through time.  Roseberry (1989:45) states that,   differential power is critical in the determination of control over the means of cultural production, the means for the selection and presentation of tradition.  But what makes hegemony cultural and not simply ideology is that it appears to connect with the experience and understanding of those people who do not produce it, people who lack access, or have sharply diminished access, to wealth and power….tradition can emerge despite the fact of differentiation.    Roseberry, as with Bourdieu, integrates the dialectic of Weber’s social action, social actors, and webs of relations with Durkheim’s organic social values and norms (i.e., collective unconscious), and establishes Marx’s economy as the mechanism by which this dialectical tension is played out.  This integration communicates across divides resulting from attachment to one or the other of these theoretical perspectives and serves to mediate these oppositional tensions in the negotiation of a political economy.  It is this historical aspect of political economy, and practice theory, that is significant in --  its attempt to understand the emergence of particular peoples at the conjunction of local and global histories, to place local populations in the larger currents of world history… Historical political economy does not simply assert that particular societies are part of world history.  It also asserts that the attempt to draw rigid boundaries around, say, the South, or Navajo.. is to reify culture.  Because populations are not formed in isolation, their connections with other   52populations and, perhaps, with the larger currents of world history, require attention.  To ignore these connections is to treat societies and cultures like ‘billiard balls,” in Eric Wolf’s telling words (Wolf 1982:6)… (Roseberry1989:50-51).     Returning to the relations between the individual and society requires combining a number of the concepts discussed so far, addressing the ‘natural economy’ and its links to community.  The term ‘natural economy’ describes ‘pre-capitalist economy’ (Roseberry 1989:201), something akin to Sahlins’s (1972) household mode of production, as based not on supply and demand (i.e., market forces) but rather on a form of ‘traditional culture.’  Forces of traditional culture emphasize values and traditions in the derivation of social standing and relations, within and between households for example, affecting community formation (Roseberry 1989:199-200; as derived from the works of E.P. Thompson).  Culture, values, and tradition develop within the “experience of community.. [of which it is noted that].. in Thompson’s early work, a cultural feeling of community was seen to rest in the actual experience of community-based social relations..[associated with].. traditional rights and customs” (ibid; notes added).    I reject the notion of ‘natural economy’ as a factor of tradition and community formation, supplanting it, rather, with political economy involving class consciousness, competition and habitus.  The ‘natural economy’ concept is problematic in its organic nature and lack of explicit discussion of social action in the formulation of community values and traditions that shape ‘class’ consciousness.  The ‘natural community’ remains organic, unproblematic, cohesive, homogenous, and bounded (i.e., non-individualistic).  Just what constitutes the internal dynamics of community remains unclear.  Understanding community requires addressing this ‘problem of tradition.’  I approach this by contextualizing the idea of a ‘natural’ community with one that manifests its negotiation of power in practice as a ‘practiced   53community’ -- as a community linked, at least in part, to conscious competition for prestige, authority and power.  Contrary to their organic idleness within the ‘natural economy,’ the ‘folk’ community is differentiated from the powerful community in the extent of their consciousness of and inclusion in relations of power, perhaps even forming diasporic fringe communities of their own.    Roseberry addresses this dynamic in re-evaluating Thompson’s work by introducing Benedict Anderson’s “imagined community” (Anderson 1983).  ‘Community’ in Anderson’s scheme is “imagined because the members of even the smallest nation will never know most of their fellow-members..” (Roseberry 1989:226).  ‘Community’ is presented as an unbounded collective of individual relations surpassing ‘face-to-face contact,’ as Anderson notes “all communities larger than primordial villages of face-to-face contact (and perhaps even these) are imagined.  Communities are to be distinguished, not by their falsity / genuineness, but by the style in which they are imagined” (Anderson 1983:15).  Communities are then imagined by whom? Constructed by whom?  As we look as specific histories, we see that political communities are not formed around images of “the state” itself but around particular social and cultural oppositions that create a group or community feeling among heterogeneous folk (Roseberry 1989:226).  The creation of community is situated in the realm of social action and negotiation among ‘heterogeneous folk,’ among agency-bearing individuals subject as ‘folk’ to ‘that which came before.’  Community is the manifestation of interaction between differentially equipped and positioned individuals acting within the context of the ‘traditional’ past and strategizing for the future.  As Roseberry states,  Central to our analysis of counter-hegemonic communities [natural economies], then, should be an examination of the cultural forms and symbols around which alternative images of community can be built, and an exploration of the organizational or institutional forms through which such images can be   54given political expression… ..one should not isolate cultural form or content from political process… …[to] see the politics in culture (Roseberry 1989:228-232; notes added).  The production of knowledge that becomes accepted and repetitively practiced over the long term as tradition signifies a successful strategy among elite relations.  This otherwise signifies a persistent cap on the manifestation of new forms of social relations engendered by the broader collective.   Differences in relations, either elite or egalitarian, are known to manifest in the variation and positioning of households and houses representing corporate or network-based communities (Canuto and Yeager 2000).   Households, as basic actors, strategize to distinguish themselves by generating and utilizing symbolic capital such as houses (Coupland 2006; Hodder 1982; Marshall 1989, 2006).  Community identity, subsuming those individuals who identify with the community in ways that are ‘stylistic’ and ‘imagined,’ is tangibly defined through attachment to and control of symbolic capital.  Human history speaks of the inter-actions between core and periphery (Wallerstein 1974, 1979; Wolf 1982) -- core and peripheral power; core and peripheral identity; core and peripheral relations set against a natural backdrop appreciated in the form of culture.  Competition -- for power, prestige, and authority -- is a basic political-economic mechanism at work in the erosion of a domestic mode of production and the development of social-spatial relations centered around elite households (Hayden 1995).  Community formation, then, is significantly influenced through household competition for prestige within a political economy.    2.1.5 Interrelationships and Transformation  The powerful principles of ‘interrelationship’ (Geertz 1973) and ‘transformation’ provide keys concepts in the formation of community.  These concepts serve to reconcile differences between extremist positions on material economy and symbolic meaning,   55dichotomies of science versus history, Marxism/materialism and practical reason versus culture, and political economy versus symbolic anthropology (Roseberry 1989:32).  As implicit in the term ‘political economy,’ Roseberry joins with Bourdieu in focusing on the realm and understanding of economy as a central part of this mediation process that is possible “if we reject the analogous positioning of the pairs” (ibid.).  The mediation of pairs, or oppositional dualities, is a universal theme implicit in the reckoning of community formation and relations of power within a landscape of social-spatial relations.   Transformation is a factor of mediating two opposing and competing forces resulting in a single outcome.  Power is a factor of this transformative process.  This transformative and integrative ‘economic’ process is inherent in the formation of communities.  Community is generated within the realm of socially situated productive processes requiring two essential ingredients -- the “presence of social and cultural differentiation. Reference to differentiation is, in part, reference to the connections between culture and relations of power and domination” (Geertz 1989:25); and “a concept of culture as material social process” (ibid.).  Roseberry champions this culture concept, arguing for its extraction from the epiphenomenal realm of the ‘vulgar’ materialists and Marxists.  He states that “the point is an intersection of the concerns of political economy and symbolic anthropology, an intersection that is based on an emphasis on meaningful action and recognizes that action is shaped by the meanings people take to their action even as meanings are shaped by people’s activities” (Roseberry 1989:32).  The effect of this ‘intersection’ is the movement of production of community toward and ultimately beyond Geertz’s recognition of cultural meaning “as socially constituted and socially constituting” (Roseberry 1989:20).  It moves culture into the realm of production itself, social production linked to people and individuals rather than process; rather than material determinism.  These processes generate a material expression and   56enactment in the construction of houses, for example, and other forms of material culture.  These two aspects of culture are integral to the definition and placement of culture within the ‘productive machine’ of community relations and the negotiation of power and authority.  “Cultural creation is itself a form of material production, that the abstract distinction between material base and ideal superstructure dissolves in the face of a material social process through which both “material” and “ideal” are constantly created and recreated” (Roseberry 1989:26).    Roseberry’s reading of Marx (Marx and Engels 1846; Marx 1867) also implies this positioning of culture within the productive apparatus of the economy.  Roseberry finds these original sources, and others (e.g., Taylor and Rebel 1981) supportive of his definition of culture as both product and production; socially constituted and socially constituting; negotiated through political means and tied to the (trans)formation of social class, social domination, and hegemony.  Culture cannot be separated from what people say, do, or have done to them.  Culture is defined through action.    Culture, then, as something practiced and not simply something produced is embodied in history (Pauketat 2001).  History becomes a construction of social action representing the political economy of culture, not simply a process… an outcome of material economy separate from culture.  Again, these thoughts echo those of Bourdieu and incorporate the same dialectic relationship between ‘objective and subjective’ in the co-mingling of relations of culture (e.g., houses) and power (e.g., control of knowledge) defining community.    Nature is ‘activated’ as something constructed and something practiced, conscious to some, unconscious to others dependent upon relative social position.   “Marx himself stressed that the interaction between humans and nature in the production process transformed nature.  The whole understanding of the human/nature relationship to which the concept of productive forces refers is more active, with humans as the subject of a process by which they transform nature and, in the process, transform themselves” (Roseberry 1989:157; see Marx and Engels 1970 [1846]:62-63; Marx 1964 [1844]:177-193).   57 Nature is no longer conceived of as passive, objective, or apart from culture.  Among the Stó:lō, the landscape is transformational (McHalsie et al 2001).  It is achieved as a result of interactions between powerful people and inhabited by powerful beings.  The built landscape of the Stó:lō world is further achieved and transformed by the houses marking their past community relationships and their past negotiation and integration of culture and nature.4  As documented ethnographically in Stó:lō-Coast Salish society, power relations are negotiated and expressed within a realm of social-spatial forms and relationships (Bierwert 1989; Miller 1999).  These are expressed as mediations between oppositional forces including, for example, the arrangement of household groups at gatherings or houses within a settlement layout -- between large and small, front position and back, center position and side, parallel or perpendicular orientations… at local and more widely regional scales of relations (Bierwirt 1989).  A core element of this study is to examine housepit patterning within and between settlements for patterning across horizontal and vertical planes of relations, as an indication of community structure.    I view the formation of community as an interconnected formula of knowledge and communication, time, strategy, power, and habitus constituting elements of a political-economic process.  That the built landscape of Stó:lō-Coast Salish community relations is partly manifest in the construction of housepit settlements is a core assumption, and orientation of exploration, of this archaeological study.  Central to the theoretical underpinning of this study is a flow of logic connecting community, authority, and archaeological housepit features.  Community manifests and creates physical patterns through a dynamic economic process of production (e.g., building houses), distribution (e.g., relations between houses / settlement layout /                                                  4 The remains of both plank houses and pithouses constitute the built landscape of the Stó:lō.  I focus on the physical remains of housepits in this study (see Chapter I for a discussion of the relationship between pithouses and plank houses).   58settlement patterning), ‘consumption’ (i.e., inhabitation), and reproduction (e.g., maintenance of houses and patterning of housing within and between settlements throughout the region).  Community is enacted and physically manifest, in this case, as variation within and between houses and settlements.  House and settlement social-spatial patterns are integrally related with the geo-political landscape in which they are established; that they, in fact, worked to establish.   The relations of settlement production are driven by agents of a community whose politics, values, knowledge, and culture are embedded and materialized by and between the basic socio-political elements of those settlements -- households.   Incorporating this perspective on the production of community is central to the development of an integrated community-settlement archaeology addressing issues of political economy and communities formed around relations of authority.  2.2 Integrating Community and Settlement Archaeology  Developing an integrated ‘community-settlement archaeology’ builds on long-standing definitions of settlement archaeology and addresses the need for integrating social theory and archaeological perspectives on settlement patterns incorporating ‘community’ as discussed above.  Bruce Trigger (1967) defined settlement archaeology as including three basic levels of analysis including: (1) the individual structure regarding the structuring of the family, residential units, class divisions, and occupational specialization within a community; (2) the settlement including information concerning social relationships; the location and nature of buildings indicating something about government, religious, economic, and other social institutions/structures in the community; and (3) settlement distributions including spatial relationships between communities, still treated as embodied in separate sites, revealing ecological and political arrangements.  Trigger’s three basic levels of analysis directly inform   59my use of housepits as a basic unit of analysis (i.e., individual structure) as well as my basic analytic framework.   My attempt at integrating these perspectives recognizing issues of power and contextualizing settlement patterning within a context of built, cultural landscape of interconnected places (Cannon 2002; Mackie 2002).    Combining social theory on the formation of communities with settlement archaeology forms an integrated approach to the study of settlement patterning and community formation in archaeology.  Community-settlement archaeology provides an interpretive framework linking measurable housepit attributes (e.g., dimensions, shapes, age, and locations) forming spatial arrangements within and between settlements with relations of power and authority influencing community organization.    I maintain a fundamental outlook in this interpretive process -- that indigenous reasoning anchored in the construction of houses and expressed in their various forms and locations remains embedded in these features of the built landscape.  The political economy of remnant houses -- surviving as archaeological housepits -- remains active and continues now, as before, to transmit, convey, and play out the reasoning and reckoning of authority and power of an ancestral community.  As these housepits have persisted through time so too have the thought processes of the people who built them.  Those thoughts and decision-making processes are embedded in the physical forms of and relationships between these features.  Housepits, as reasoned features and forms of practical knowledge, embody a set of total cultural relations integrating society, politics, economics, religion, and environment as elements of a built cultural landscape.    The position of housepits and settlements within any political economic system is likely to ebb and flow through the rising and falling in power of households and large collectives.  Power may be negotiated in new ways and re-situated in new locations, away from the places   60where households previously built their homes and attached to land and resources in the past.  Such shifts may incorporate an enduring connection to these places within a broader context of reckoning relations within and to a cultural landscape.  Such shift may be caused by any number of reasons including shifts in the location of important resources in the environment and technological adaptations which provide different access to and uses of resources.  Changing social relations and the development of new networks providing control over labor and/or links to more distant resources are equally significant causes of change.  These shifts, whether environmentally, technologically, or socially motivated, come as part of incorporating change into the production of knowledge and reckoning of relations and meaning between people, places, and things expressed as material culture.  Tracking the changes in the relations of housepits through time requires measuring and describing differences in housepit form and location at multiple levels of analysis coincident with definition of settlement archaeology (Trigger 1967).    Changes in the physical dimension of housepits and their arrangements within and between settlements represent patterns affected by political economic influences.  Political economic factors can be described on two basic levels of horizontal and vertical planes relations.  These planes are useful in describing the degree of social stratification within society representing heterarchical or hierarchical forms of social organization; and the extent of spatial relations between settlements representing corporate or network modes of relations (Feinman 2000), discussed in more detail below.  Changes in these two basic forms of relations are affected by the political economic processes described below and can be traced out within a framework of changing community formation.  Describing these changes over the long-term leads to a framework for investigating changes in the formation of Stó:lō-Coast Salish ‘housepit’ communities.   612.3 Examining Current Relations between Community and Settlement Archaeology The relationship between community and settlement archaeology hinges on Bruce Trigger’s (1967:158) recognition of the need to address social relations.  This position still holds currency as a cornerstone of settlement archaeology and provides a basis from which archaeologists have approached its modernization over the last 40 years.  This aspect of settlement archaeology connects with the more recent outgrowth of what has been called ‘community archaeology’ (e.g., Marcus 2000).  The addition of social theory augments the interpretive framework of settlement archaeology which was largely methodological in its definition.   Community archaeology incorporates elements of interaction theory and interaction studies.  Schortman and Urban (1992) define interaction studies as “research founded on the notion that individual societies, or ‘cultures,’ are not viable but depend on inputs from other societies for survival and reproduction from generation to generation” (Schortman and Urban 1992a:3).  Interaction studies are complementary, rather than adversarial, to culture ecology and processual studies.  The ultimate goal of interaction research has been identified as writing “total histories of ancient societies, histories which place local developments within the rich network of connections any one society maintained” (Schortman and Urban 1992b:248).  Embracing elements of feminist archaeology and post-processual theory attributing agency to individual actors, the basic unit of analysis in interaction studies equates to “a dynamic network of inter-actors whose transactions link their respective polities and affect sociopolitical developments within these units” (ibid:238).  A central question arising from this point of view, then, is what scale of interaction is most critical in defining a community?   Recent practice among archaeologists dealing with ‘community’ addresses the relationship between interaction and scale in different ways.  Kolb and Snead (1997) treat   62community as a “fundamental unit of human society” (Kolb and Snead 1997:609) located socially between the family and their extensive social networks, and archaeologically between the site and region.  Their view of community is quite different from Johnson’s (1982) ‘sequential or simultaneous hierarchies,’ Service’s (1962) ‘tribes,’ or Johnson and Earle’s (1987) ‘local groups.’  Kolb and Snead (1997:611) consider community to be “a minimal, spatially defined locus of human activity that incorporates social reproduction, subsistence production, and self-identification.”  They derive their definition from Hollingshead’s (1948:145) earlier view of ‘community’ “1. as a form of group solidarity, cohesion, and action around common and diverse interests; 2. as a geographic area with spatial limits; or 3. a socio-geographic structure which combines the ideas embodied in 1. and 2.”   Asserting that the blending of regional and local level archaeological investigation creates “fuzzy pictures,” Kolb and Snead (1997:612) develop a methodological program of “microregional analysis” incorporating three analytic strategies examining differential labour investment (family, festive, corvée), spatial relations between community elements (houses, etc.), and boundary maintenance (physical, stylistic markers).  This approach is intended to enhance the “operational ability of archaeologists to examine local and medium-scale social organization, better meshing the dimensions of system and time with the dimensions of space and ecology” (Kolb and Snead 1997:622).  Kolb and Snead’s attempt to balance social-spatial and material-ecological relationships provides a middle-ground between processual and post-processual theory and practice.  Their ‘local’ approach, however, remains limited in its emphasis on a ‘small world’ apart from broader, regional collectives and extensive community networks (e.g., Steponaitis 1978).   I maintain the household as the basic building block of community, located socially, economically, and politically between the family and their extensive networks, and   63archaeologically between the feature and region.  Housepit features, in this case, correlate with households that may include extended or multiple families.  I do not incorporate the rather severe limits Kolb and Snead place on the extent of those networks potentially stemming from micro-regional types of relations.  The self-reflexive socially reproductive aspects of their work provide a lead into Practice Theory as incorporated by Yeager and Canuto (2000). Yeager and Canuto’s (2000) Interactive-Practice based approach to the ‘archaeology of community’ rejects Murdock’s functionalist notion that the site equating to community.  They counter a view inherent in most processual studies of settlement patterns maintaining the site as the basic social unit.  Their definition of community is founded on “the conjunction of people, place, and premise… an ever-emergent social institution that generates and is generated by supra-household interactions that are structured and synchronized by a set of places within a particular timespan” (Yeager and Canuto 2000:5).  A community is “not a spatial cluster of material remains to be observed, but rather a social process to be inferred” (ibid.).  Community, for them, is the outcome of social action of which interaction is a central tenant.   I do not entirely agree with the premise of ‘practice’ as understood and utilized by Yeager and Canuto.  Archaeologically, community is represented by social-spatial and temporally clustered material remains.  Within these arrangements one can observe the socio-political processes by which they were formed.  Community, as the conjunction of people, place, and premise is completely understood with the inclusion of ‘things’ as in resources and material goods.  Communities at all levels have a powerful need to materialize themselves for a number of reasons, from warding off the prospect of an abstract, ambiguous, and ever-changing existence to self-identification and self-actualization in relation to an ‘other.’  This need involves the competition over and negotiation of status and authority as part of their actualization and organizational form.  The perpetuation and survival of a community is, in part,   64connected to the production and reproduction of a recognizable material manifestation over time.  Housepits are ‘things’ of community built by households affected by social, economic, and political processes.  From a Stó:lō-Coast Salish point of view it is also necessary to include religion or spiritual forces.  Meaning is anchored and expressed in things (Cohen 1985; Delanty 2003).  These are essential parts of community that permit archaeological investigation.     Peterson and Drennan (2005:6) focus on “patterns of intensity across space” with interaction coming into focus “at different scales to reveal different structures that exist simultaneously in a given region.”  They accuse Yeager and Canuto of being too ideational, lacking grounding in the ‘reality’ of influences of the environment and the substance of direct face-to-face contact and daily interactions.  They focus more on the level of daily face-to-face interactions as the significant substance of community.  I support Peterson’s and Drennan’s argument only in so far as they identify a lack of balance between the ideational and material composition of community, pointing to a divide between settlement and community archaeology that I address in this study.  They neglect the power inherent within the relationship between an imagined community and a material realm. The notion that people -- versus more organically structured societies -- are the basic agents of interaction situates this form of analysis squarely within realm of social theory as applied to regional analysis.  Primary social factions motivating involvement in interactive relations are elites.  Elites are those powerful individuals dominating political-economic networks of interaction, controlling labour and the relations of production, and most actively involved in defining societal habitus.  The influence of elites is discussed at great length by archaeologists focusing on the development of complex social organization among intermediate-level societies (e.g., Arnold 1996a; Clark and Blake 1994; Hayden 1995).     65Modeling interaction among elites within a community requires considering the possibility that elites engaged in the daily, day-to-day interactions of local village life and (or perhaps even moreso) maintained more extensive relations between settlements with households of equally high status.  Thus, powerful households maintain an elite stratum of community described by a broad social-spatial scope and potentially limited daily interaction with lower-status families living in the same settlements.  Elite households can be both locally and regionally engaged, and influential in the establishment and maintenance of a multiplicity of community strata operating simultaneously.  This view of community fundamentally differs from those described above by recognizing multiple layers of community, stratified by wealth and status, operating simultaneously in different scales of interaction -- differentiated and sorted by access to power.  Blake and Clark (1999) argue that this is one of the ways that elites can overcome the egalitarian-maintaining processes in tribal societies.   Relations between households operating within a domestic mode of production and having minimal social stratification are the most localized expressions of community (Sahlins 1972).  Feinman (2000) describes this expression of community relations as a form of corporate organization exhibiting a limited range of horizontal relations with others and a ‘flattened’ vertical scale of internal social differentiation.  This form of organization may not change but may be added to as upper echelons of those with elite status and having broader networks emerge from this basic form of community organization.  Sahlins (1972) connects this process with the intensification of production moving away from the domestic or household mode of production.  Alternately, more localized levels of community operate within a domestic mode of production as an under-class of households within a broader landscape of social stratification.  Such households exist as an element of a wider-ranging, nested hierarchy of relations   66dominated by networks of elite or upper class households simultaneously operating beyond the limits of household production.   The process of removing the cap from or operating outside a domestic mode of production leads toward the development of what Feinman (2000) calls a network strategy and form of socio-economic/political organization.  Network relations are associated with relations over a broad horizontal plane and a more stratified social structure on an expanded vertical scale.  Thus, multiple elements and levels of communities may arise and simultaneously function in an integrated way as differentiated by access to power, prestige, and authority.    2.4 Stratification and Corporate-Network Relations: A Framework for Investigating the Evolution of Housepit Communities    A foundation of my interpretive framework is the relationship between housepit and/or settlement size and socio-economic and socio-political status, and thus power.  As is commonly acknowledged in Northwest Coast household studies (Ames 1995; Coupland 1996; Matson 1996), larger houses (therefore larger housepits) represent households of higher social status5.  This assertion is supported by the oral history and ethnographic records of Stó:lō-Coast Salish household relations (Barnett 1938, 1955; Bierwert 1986; Haeberlin and Gunther 1938; Smith 1940; Snyder 1964; Miller 1999).  The application of practice theory and political economy equates households, and therefore the house in which they occupied, with a central place in the production of knowledge, relations of authority and power, and development of Stó:lō-Coast Salish community.                                                    5 I recognize that large houses are not a necessity of complex social organization and that other forms of relationships beyond those expressed in house or settlement size may support the development of complex social systems, such as among the Chumash who operated out of small households (Arnold 2004:173).  I also recognize that this association is not universal.  Support for this association among the Stó:lō-Coast Salish, however, comes from the ethnographic and archaeological information drawn from the previous chapter.   67 My interpretive framework is influenced by Arnold’s (2004:173) statement about understanding the evolution of community organization where “by tracking emergent hierarchical organization alongside [a] corporate network dimension, we may begin to understand the evolution of pithouse communities.”  This framework describes community formation relationally along two axes and continuums of measurement defining (1) heterarchical-to-hierarchical structures (Crumley 1987; 2001), and (2) corporate-to-network socio-economic strategies (Feinman 1995, 2000; Feinman et al. 2000).   Addressing power, political action, and practice, Feinman (2000) adds the ‘orthogonal’ dimension of the ‘corporate-network’ continuum to the prior monolithic ‘egalitarian-stratified’ view of social organization (Fried 1967; Service 1962, 1971; Flannery 1972; see Haas 2001).  He challenges the view that “political complexity can be equated generally with (or measured by) marked centralization or rampant self-aggrandizement or individualism” (Feinman 2000:31).  I recognize this view, among the Coast Salish, as the emergence of complex community relations within a collective, rather than individualistic, society.   Corporate and Network strategies tend to be opposed modes “at the polar ends of a more continuous dimension that can be conceptualized as orthogonal to (or cross-cutting) the familiar axis of hierarchical complexity” (Feinman 2000:32).  Corporate-mode relations and access to power are less individualized, more apt to be shared, and often a product of group membership.  Network-mode power can also be group-oriented.  This mode of power assumes elements of the traditional model that “hierarchical complexity, greater wealth stratification, and the increasing centralization of power always co-occur” (Feinman 2000:31).  The concept of ‘network’ opens the possibility, however, that this co-occurrence can be manifest at a broader, more collective level rather than in a more individualistic sense, as personified by accumulators (Hayden 1990), aggrandizers (Clark and Blake 1994), and entrepreneurial elites (Hayden 1995).  Network mode   68power, too, acknowledges that the formation of hierarchical complexity (i.e., expressed as centrality in social arrangements) can be spread out across a wider plane of horizontal relations.  “If centralization is defined as the greater consolidation of wealth and political power in the hands of a single individual or ruler, then larger, more hierarchical polities are not always more centralized” Feinman argues (2000:31).  The two axes of measurement described by Arnold are required to assess this suggestion.  A basic rationale for developing this framework is to force archaeologists to focus on power and the different ways it is achieved, while providing a framework for the comparative examination of political action.    Emergent corporate or network forms of interaction are expressed on a horizontal plane of relations represented by social-spatial relations between households at two levels -- within settlements and between settlements.  Differences between households may not be evenly distributed between the co-existent settlements of the region.  Measures of variability may be consistent or otherwise vary between settlements, indicating heterarchical or hierarchical relationships that extend beyond the settlement-specific level.  Exploring the horizontal plane of relations requires describing variability in the spatial arrangement of housepits of different sizes across at least two levels analysis. Defining corporate or network modes of relations between households requires looking at the entire system of horizontal and vertical relations, manifest in the three levels of analysis described in Trigger’s (1967) definition of settlement archaeology.  Horizontal and vertical planes of relations are not independent of one another but represent an interconnected system.  I visualize this relationship as the two axes joined together with the vertical axis centrally anchored on a horizontal plane.  The relationship between heterarchy and hierarchy on the vertical plane and corporate or network relations on the horizontal plane is associated with two things: the height and angle or tilt of the vertical axis; and the spread of the horizontal plane.     69A corporate organizational model, for example, includes a perfectly upright or straight vertical axis of social relations.  No matter how hierarchically differentiated, corporately organized communities can be thought of as constituting a ‘self-contained’ set of social-spatial relations represented as series of expanding concentric rings (i.e., social strata) contained by an outer ring.  The level of hierarchical authority diminishes and becomes more heterarchically arranged with the flattening of the vertical axis and reduction of variation between households.  This corporate arrangement constitutes a normally-shaped community demographic replicated across the landscape.  This type of formation is only one of many possible manifestations of community organization, one that would indicate a heterarchical foundation of hierarchical authority (Crumley 1987, 2001).   Expanding this system into a more complex system of community relations -- associated with a hierarchically arranged, network mode of relations -- is slightly more difficult to describe.  Such models require conceptually tilting the vertical axis.  Tilting the vertical axis expands the range of relationships across a broader horizontal plane.  The relationship between the height of the vertical axis and its degree of tilt describes stratification as socially and spatially differentiated sets of community relations.  The greater the tilt of the vertical axis and the wider the horizontal plane of relations, the greater the degree of hierarchically arranged network-type relations.  This is particularly so when matched with high order stratification and more hierarchical forms of social order -- expressed as pronounced inter-settlement variation among households within and between settlements.  Taken to an extreme, some settlements -- particularly those between the extreme upper or lower echelons of society -- may become so disassociated or off-set from each other that they, in essence, form a sub-set of ‘communities within communities.’  This network societal form constitutes a non-normal shape in statistical   70terms whereas the shape of the previous (corporate) model, with a straight vertical axes, is far more normal or ‘circular’ (i.e., normal) in its demographic patterning.   This interpretive framework bridges mathematical and socio-political languages.  The measurement and comparison of housepit dimensions and locations captures community relations and translates them into the quantitative terms using mathematical language.  These data are essential in investigating horizontal and vertical planes of relations.  Meaning is derived through the process of translating quantitatively-based patterns describing housepits and their groupings into political economic terms describing patterns of social organization and change.  Changes in housepit and settlement attributes described in mathematical terms reveal changes in Stó:lō households and communities over the nearly 3,000 years reviewed in this study.     71CHAPTER III - DATA COLLECTION: MAPPING AND TESTING HOUSEPIT SETTLEMENTS     Neither a regional settlement pattern (i.e., top-down) approach nor a household (i.e., bottom-up) approach are sufficient, on their own, in undertaking an archaeological investigation of community organization.  A community exists both emically as “a network of interactions among families, residential wards, real and fictive kinsmen… [down to the earth and resources themselves]” and etically or archaeologically as “a cluster of artifacts and ruined structures that exists in space” (Marcus 2000:239).  The term region has at least two meanings as “a mental concept on the part of its inhabitants… [and]… a reality that exists in space” (Marcus 2000:238; also see Kantner 2008; Tabor 2004).  Investigating households and villages is a good start, but “ahead lies the difficult task of developing methods to recover the units defined by indigenous societies.  It is a task that must inevitably unite excavators and settlement pattern archaeologists with ethnohistorians and ethnologists” (Marcus 2000:240), not to mention geographers, geologists, linguists, and indigenous oral historians.  I address this task by examining three basic variables of housepit form (size, shape), location, and age across three levels of analysis spanning ‘the housepit’ and ‘the region.’  I apply this process in connecting archaeological data and ethnographic models regarding Stó:lō-Coast Salish community organization and change through time.     A number of intermediate questions link my broader objectives concerning Stó:lō-Coast Salish community formation with the study of archaeological housepits and settlements.  These questions include:  Where are housepits located throughout the landscape?  Is there patterning evident in housepit settlement locations, particularly when examined within a framework of travel routes and avenues of communication?  What is the range of housepit sizes and has this changed over time?  What is the range of housepit shapes -- circular, square, rectangular, or   72oval -- how has that changed through time?  How are housepits spatially arranged within settlements, considering differences in size and form?  Does settlement composition -- layout and arrangement -- change through time?  What is the range of settlement sizes and how have those dimensions changed through time?  Are there quantitative patterns in housepit and settlement data -- location and form -- when viewed through time?    A lack of fundamental information made previous attempts at answering such basic questions about housepit form and size untenable.  Addressing these questions required collecting new data describing the location, form, and age of housepits, as a basic unit of analysis.  Archaeological information on above-ground houses (i.e., plankhouses) is far more limited than for in-ground houses.  While this study provides new information on plankhouse features and includes them in the discussion of settlement patterns (see Chapters VII and IX), their limited archaeological visibility and the available information about them effectively excluded them from consideration in developing the basic framework of this study.  While incorporating housepit data from a broad range of sources, I rely mainly on data recently collected from 11 housepit settlements.  These data, collected between 2003-2006 as part of the Fraser Valley Archaeology Project, form a sample of approximately 10 % of recorded housepit settlements in the mainland Gulf of Georgia Region (as of 2005).  I focus on the analysis of 114 housepits derived from the mapping of these 11 settlements, approximately 15 % of those recorded in the study area.   I use Exploratory Data Analysis (EDA) techniques described by Drennan (1996:3-73), Fletcher and Lock (1991), Hartwig and Dearing (1979), and Tukey (1977) as a means of quantifying and exploring the structures of these data sets.  I draw on the basic principle set out by Tukey (1977:v) in his introduction to EDA that “It is important to understand what you CAN DO before you learn to measure how WELL you seem to have DONE it.”  EDA differs from   73classical ‘confirmatory’ statistics in its focus on “looking at data to see what it seems to say… It regards whatever appearances we have recognized as partial descriptions, and tries to look beneath them for new insights.  It concern is with appearance, not with confirmation” (ibid.).   EDA allows more subjectivity than classical or Bayesian statistics, suiting studies like this one aimed at the preliminary exploration of data structures and data organization rather than testing of hypotheses.  This preliminary approach to data organization and analysis serves a fundamental purpose of supporting follow-up confirmatory analyses, defining what we can do before we enter that realm of statistical testing.  EDA is, thus, a non-parametric approach to statistical analysis aimed at exploring and understanding data structure rather than hypothesis testing.   EDA provides a range of methods for exploring data structure.  A basic approach used here aims to explore data structure by defining and describing the center, spread, and shape or distribution of values in a batch or group of data.  The center of a group of data -- where values bunch together -- is indicated by its measure of central tendency or mid-point (Drennan 1996:17).  Outside EDA, the mean or average value is commonly used as a measure of a group’s center.  In EDA, the median, or most commonly occurring value, is used to identify center though often in conjunction with the mean.  Spread, or dispersion, describes the extent to which the values in a group of data are all-together spread out or bunched together (Drennan 1996:27).  Spread is measured in various ways including the range of values above and below a group’s center, the inter-quartile range (IQR) or mid-spread (i.e., middle 50 % of values), and variance or standard deviation.   Shape, or distribution, refers to the way in which values in a group of data “are distributed along the number scale, apart from [the center] and spread… There are two principle aspects to the shape of a [group]: number of peaks and symmetry” (Drennan 1996:53).  Also, “the presence of multiple peaks in a [group] is always an indication   74that two or more fundamentally different kinds of things have been thrown together and measured… [groups with multiple peaks cannot be analyzed further.  The only correction for this problem is to subdivide the [group] into separate [groups] for further analysis” (Drennan 1996:14-15).    EDA is visual.  I use graphic techniques including stem-and-leaf graphs, box plots, and histograms to explore data structure, compare groups, identify anomalies (e.g., groups with two centers or modes), and define sub-set groups with single centers and more normal shapes.6   Throughout this study I use stem-and-leaf plots in preference to histograms, though sometimes both together (Chapter VI), as a way of maintaining transparency of data values in presenting the results of analyses.  These techniques are used to identify relationships between center, spread, and shape, as well as outliers or anomalous values that fall far from the group and cause its shape to become asymmetrical or non-normal.    EDA uses methods and values more ‘resistant’ to being affected by widely spread or outlying values within a group, such as median rather than mean to measure a group’s center, as                                                  6      The following definitions of box-plots and stem-and-leaf plots are based on Drennan (1996:39-41) and Tukey (1977:675).  Stem-and-leaf plots are a fundamental data organization tool used to order data along a scale; corresponding to the more commonly used histogram.  The stem section defines the scale and the leaf attaches each value to its place within that scale, exposing patterns in the distribution of data.  Stem-an-leaf plots present all numeric values, whereas histograms do not.  A stem-and-leaf plot graphically represents the spread and shape of a group of data, where data values are portrayed in place of the bars common to histograms.  A stem-and-leaf plot “is a generalized two-digit display, in which the lefthand portion of the values displayed is given by a stem value, while the righthand portion makes up a leaf” (Tukey 1977:675).   As in a histogram, the stem represents a bin- or stem-width defining a range of values (e.g., units of 10; as in values within sets of ranges from 1-10; 11-20; 21-30 and so on).  The stem forms a vertical column.  A plot with a stem-width of 10 (km), is represented by a vertical, lefthand column with a sequence of values of, for example, 6, 7, 8, 9, and 10 equating to measurements of 60-69, 70-79, 80-89, 90-99, 100-109 km (e.g.,  units of distance).  Actual values of data in a group are attached as leaves projecting from each stem, organized by values falling within each stem-width.  Thus, using the above example, a sequence of leaf-values of 2, 4, 4, 5, 7 attached in the righthand column to the stem value of 6, represent values of 62, 64, 64, 65, and 67 km (e.g., actual distance travelled). Stem-widths are identified in each of the stem-and-leaf plots included in the following analyses.         Box-plots (or box-and-whisker plots) graphically represent the center and spread of data within a group.  The median represents the center.  The mid-spread (middle 50 % of values) is represented as a central box; the upper and lower quartile ranges (upper and lower 25 % of values) as whiskers stemming from the box including the largest values that are not outliers; and any outliers (values that fall far outside the central bunch) as dots located beyond the whiskers.  Extreme outliers are values that fall more that two times the midspread, or inter-quartile, range from the center.      75noted above.  Using techniques described above, outliers or extreme values (i.e., values far away from the bunch of values in a group), and coarse or inaccurate data, that cause asymmetrical shapes and skewness of a group’s center, shape, and/or spread, can be identified. Outliers, extremes, and coarse data can be treated in a number of ways including elimination as a means of smoothing data and organizing groups based on normal shapes and distributions.  Re-expression or transformation (Drennan 1996: 56; Tukey 1977:57) is another EDA technique used to treat groups with asymmetrical shapes.  Arithmetic functions (e.g., square root or logarithm transformations) may be used to remove a group’s shape or set it to a standard, single-peaked, and symmetrical form to produce a new group that may be better suited to analysis.  EDA provides a choice of a range of exploratory options.  I do not apply re-expression in this study, though potentially useful, in deference for maintaining the transparency of unaltered data values.  The results of EDA, as applied here, aid in data organization and lead to the more confident use of classical, confirmatory statistical tests (e.g., Chapter VI).    The exploratory principles and objectives EDA are well-suited to the exploratory principles and objectives of this study.  This study, at its outset, produced a new set data describing 114 housepits about which little could be said of patterns underlying their appearance.  The housepits represent 14 distinct occupations in settlements located throughout the Fraser Valley, spanning nearly 3,000 years between 2,550 cal B.P. and 100 cal B.P.  EDA is also used in Chapter IV to explore travel distance values and the distribution of housepit settlements with the region’s transportation system.  The quantitative foundation of this study rests on an objective of looking into this new set of data describing housepit form and seeing what it seems to say; looking beneath its appearances for new insights; finding out what we can do with it.  Results of these analyses provide a basis for describing and classifying housepit   76features and settlements by size, age, and geographic location.  These EDA-based results support the use of statistical tests addressing questions of association and independence of data sets (e.g., housepit size and age).     3.1 Housepit Form and Indigenous Identity in the Fraser Valley   Variation in house form fuels anthropological debate about the relationship of Fraser Valley housepits -- and Stó:lō peoples -- to the Coast and/or the Interior.  Although housepit features of various types are commonly found throughout coastal areas of the Pacific Northwest (Barnett 1938; Mackie and Christensen 2003; Acheson 1995), archaeological perception often and questionably takes an ‘Interior’ rather and ‘Coastal’ view of housepits in the Fraser Valley.  Rectangular features are more commonly and acceptably recognized as ‘Coastal’ house forms.  A long-standing perception among archaeologists is that Fraser Valley housepits are most commonly circular in shape and therefore related to the pithouses of Interior Plateau peoples (e.g., Borden 1970, 1975; Carlson and Wilson 1980; Duff 1952; Hanson 1973).  This premise of Interior-oriented relations historically affected the development of archaeological models describing Lower Fraser River Canyon and Fraser Valley material culture as significantly affected by the diffusion of traits from the Interior to the Coast (e.g., Borden 1968; Burley 1980; Mitchell 1963; Matson 1976).  This in turn influenced an anthropological view of ‘upriver’ Stó:lō identity as composed of ‘mixed’ cultural traits, origins, and affiliations -- neither really Coastal nor of the Interior.  These assumptions remain unfounded and unsupported by any solid set of housepit data.   In my experience among the Stó:lō, this anthropological identity crisis is not commonly shared within an indigenous reckoning of Stó:lō-Coast Salish identity.  The Stó:lō collectively view themselves as a community with extensive social, economic, linguistic, and cultural   77relations centered within the lower Fraser River Watershed and extending across the Georgia Strait, into Puget Sound, and upriver among the Interior Salish (Carlson 2001a:24-28; Smith 2001).  A downriver, coastal orientation is predominant in these sets of relationships (Bierwert 1986).  Exploring housepit form as a factor of community formation in this study incorporates landscape -- beyond individual features -- as an aspect of “identity geography” (Carlson 2001:26) and serves to test archaeological assumptions about housepit form affecting the anthropological understanding of Stó:lō identity.   3.2 Archaeological Data on Housepits in the Mainland Gulf of Georgia Region  The data collection strategy in this study was designed to augment existing spatial and temporal information supporting the study of housepits and settlements in the mainland Gulf of Georgia Region -- particularly in the Fraser Valley or ‘upriver’ portion of the region.  In 2004, the provincial archaeological database accounted for 112 housepit settlements containing at least 560 individual housepit features (Schaepe 2004).  The majority of these records are founded on archaeologically documented sites (74 %) while others are based on oral historical (25 %) or historical records of site locations (1 %).  Previous overviews of housepits in the region are provided by Duff (1949), Little (1996), Mohs (1990), and Schaepe et al. (2001).  Useful information about Stó:lō housepits and settlements comes from both archaeologists and ethnographers.  A number of ethnographers (Barnett 1938, 1955; Duff 1950, 1952; Hill-Tout 1902, 1904; Jenness 1955; Smith 1945, 1947; Wells 1987) working in the region over the past 100 years described the two archetypical Stó:lō house types -- sqémél (‘in-ground’ pithouses) and s’iltexwáwtxw (‘on-ground’ shed-roofed plankhouses).  Other less well known house types such as mat-lodges and ‘hybrid-structures’ combining elements of both pit- and plank-houses are also described (Smith 1945).  The earliest anthropological descriptions of sqémél come from ethnographer Charles Hill-Tout in 1895.  Archaeological investigation of   78housepit features otherwise began in the 1940s (Duff 1949; Smith 1947) and intensified in the 1950s and 1960s (e.g., Borden 1950, 1960, 1961; Kenyon 1953; Kidd 1968).  Investigations carried out between the 1940s and 1960s account for approximately 53 % of all documented housepits settlements in the mainland Gulf of Georgia Region.  On-going work carried out largely within the context of cultural resource management projects of the 1970s (e.g., Hanson 1973), the 1980s (e.g., Arcas 1986), and the 1990s (e.g., Schaepe 1997) account for most of sites recorded as of 2004.  Many housepit settlements recorded in the region have been severely impacted by 19th and 20th century development and land alteration.  The majority of housepit features recorded in the region have suffered disturbances that limit their usefulness for, or in some cases eliminate the possibility of, archaeological investigation.  Based on my review of site records, I estimate that about 7 % of all recorded housepit sites in the mainland Gulf of Georgia Region are destroyed, while 66 % have suffered significant impacts, leaving only about 27 % of documented housepit settlements largely intact (Schaepe 2004).  While archaeologists have investigated a few housepit settlements (Mitchell 1963; Hanson 1973; Schaepe et al. 1999; von Krough 1980), data from these sites and features have never been systematically analyzed to provide even basic descriptions of the range of variation in size, shape, or age.    Initial ‘pilot’ attempts at sorting housepit settlements and features by any level of detail beyond basic site type and location were frustrated by lack of data (e.g., Schaepe 2001a).  Analysis and comparison of housepits and settlements was limited by the lack of a uniform unit of measure for plotting sites and features in space.  Given that the majority of sites were recorded prior the 1980s, most researchers used latitude and longitude while others, later on, used the much more intuitive, Universal Transverse Mercator (UTM) system.  The British Columbia Archaeology Branch maintains a database of all recorded archaeological sites in the   79province, one field of which contains spatial location information.  Generally, spatial data were available in one form or another but only for sites, not specific features.  Very little spatial information existed beyond site location.  Within-site spatial information came only from hand-drawn maps, if available, on the archaeological site forms.  These maps are generally detailed enough to show the location of housepits in a site plan.  The scale and level of detail of these maps varies widely, sometimes for the same site as exemplified by differences in depictions of the well-known ‘Katz’ (DiRj-1) site developed by Hanson (1973), Coupland (1996), and Schaepe (2001b).  Housepit features tend to be represented as circles and are often ambiguous, recognizing that feature shape is difficult to determine without extensive vegetation clearing and detailed mapping techniques.  The housepit dimensions provided in site forms were typically determined as a factor of expedient, low resolution mapping techniques -- useful as general estimates of size based on length and width measurements but not suitable for determining area calculations (as discussed below).   Dimension measurements were sometimes found to be lacking altogether.  Settlement area measurements were based on a single polygon or site boundary including the entire site area and were not accurate enough to be of use in my study which required area measurements at the level of the individual house feature.  Typically, only limited topographic information is provided describing the landscape in which these sites are located -- again, as an understandable factor of expedient mapping techniques.  The methodological objectives of this study address these spatial issues as a means of allowing feature-level, settlement-level analysis, and inter-settlement-level analyses.    The limits of existing temporal data were found to be even greater than those affecting the spatial dimension of this study.  Only seven of the 112 housepit settlements documented as of 2005 had radiocarbon dates directly associated with house features.  Of these, only five fell   80within the temporal limits of this study.7  A number of dates from above-ground house features indirectly supplement this body of information, primarily derived from work at the Scowlitz site (Lepofsky et al. 2000).  Additionally, Mitchell (1963:133) provided a relative age for one housepit feature at the Esilao site (DjRi-5).  Arnold and Schaepe (2003) also provide a relative age for a housepit feature at the Ts’qo:ls site (DiRi-1) which was excluded from this study due to its Colonial-era age (i.e., post-1850 AD).  Supplementing the available set of temporal data was critical to addressing the objectives of to this study.    Thus, existing spatial and temporal data describing the 112 housepit settlements recorded as of 2005 proved to be limited in their usefulness for the purposes of the present study.   A pilot study that I carried out in early stages of the project helped define issues and pinpoint those areas where data collection was most needed.  The data collection strategy employed in this study thus facilitated the study of housepit features by adding to the radiocarbon record, providing for uniform and comparable units of spatial measures, and producing high resolution maps for a sample of entire settlements and the individual houses within them.  Fieldwork carried out between 2004 and 2006 implemented the data collection strategies described below.      3.3 Settlement Data   Spatial locations at the settlement-level were established using UTM coordinates for each of the 11 settlements included in this study (Figure 3.1).  Calculation of UTM coordinates was done with ArcView 3.2 geographic information system software using the standard North American Datum (NAD) 1983.  Having UTM coordinates at a fine scale of resolution provides accurate settlement plots from which to investigate broad, regional-level analysis of settlement                                                   7 Data from the Maurer site - DhRk-8 (LeClair 1976; Schaepe 1998, 2003), the Xa:ytem site - DgRn-23 (Mason 1994), Iy’oythel - DgRl-10-Feature 2 (Schaepe et al. 1999), and Scowlitz site - DhRl-16-Features 4 and 8 (Lepofsky et al. 2000) were excluded from this study as pre-dating 3,000 cal B.P.   81  Figure 3.1.  The location of eleven housepit settlements included in this study.     82locations.  UTM location data were adjusted to a sub-meter level of accuracy (e.g., +/- 10 cm).  A table of housepit settlement UTM-location coordinates is presented in Appendix I.    Using UTMs has advantages including allowing the accurate plotting of features and sites using simple scatterplots -- as found in standard statistics packages like SPSS -- where the X-axis is UTM mE (metric easting) measurements and the Y-axis is UTM mN (metric northing) measurements.  Settlements and features represented as points on a spatial plane can be directly incorporated into the outcomes of analyses and easily, accurately, and expediently plotted in space as various classes or groups (e.g., size, age) of data.  This was found to be particularly useful in analyzing intra-settlement arrangements of housepits.  Also, UTMs are the standard spatial unit of measure used in Golden Software’s Surfer 8.0 mapping and surface imaging software, used extensively as an integral feature of data analysis in this study as described below.  UTMs also permit the easy movement between levels of analysis (e.g., housepit/feature, settlement/site, and region) and facets of research (e.g., surveying and mapping, testing, formal excavation unit layout, carbon sample or artifact proveniencing, analysis) as a common unit of measure with a single datum point and a flexible resolution range that accommodates measurements ranging from centimeters to kilometers.  3.4 Feature Data   Housepits and other features at the 11 housepit settlements were mapped using a high precision GPS unit (Leica 1200 GPS; 1200 GPS rover) to tie each site into the UTM grid, and establish datum points and mapping station locations.  This process resulted in centimeter-level accuracy on a global scale.  A Leica Total Station (703R or 705R) was used to collect three-dimensional topographic data across the surface of each site, with particular focus on all housepit features.  Brush was cleared at each site to provide visibility for mapping.  This   83activity, in itself, proved to be a very useful way of gaining insight into the organization of features at each settlement.    Detailed information on this mapping process is provided in Schaepe, Blake, Formosa, and Lepofsky (2006:7), a sample of which is provided below: Through trial and error, our mapping methodology was refined in order to collect enough data to make accurate surface models for the sites, while at the same time, avoiding over-sampling and potentially collecting more data points than necessary to create the surface model. The full surface of a site was walked either in a grid with regular capture intervals between measurements, or in radiating lines extending out from the total station. Housepit depressions were sampled more intensively, with data collection points spaced every 30-50 cm depending on the degree of vertical rise between points.  Smaller cultural depressions, such as cache pits were sampled at 20 cm intervals. Mounds were sampled around the perimeter then points were collected at approximately 30 cm intervals with an additional point or two at the top of the mound.  Portions of the site with no visible surface features were sampled at approximately 1-2 m intervals.  Any structures – stone walls, cairns – were surveyed as standing structures separate from the topographic survey.  UTM coordinates marking the center-point of each housepit feature were extracted from these precise measurements (Appendix I).  These data were used to make detailed settlement maps which in turn were used to make accurate and reliable measurements of feature form (e.g., dimensions, area, and shape).   3.5   Housepit Form - Size and Shape  Digital topographic data collected using the Total Station and GPS units were processed and projected in various map forms (e.g., surface maps, topographic maps, shaded contour maps, etc.) using Surfer 8.0.  Field point measurements were interpolated using a kriging algorithm with grid size of 0.33 m in order to create a digital elevation model for each site in the study.  Surface maps of each of the 11 settlements included in this study are based on data from Schaepe, Blake, Formosa, and Lepofsky (2006), Graesch (2006), and Lenert (2008), and Lenert   84and Lepofsky (2006).  The following sections describe the three-part process by which measurements of housepit feature size and shape were ‘pulled’ from these maps.  3.5.1  Step 1 - preparing feature-level maps  Surfer 8.0 map-making software was used to develop a set of high resolution maps of housepit features at each settlement (see Figures 3.2-3.12).  These maps were developed from the three-dimensional surface points recorded for each site.  I then produced maps that integrated topographic map (5 cm intervals) and shaded contour maps as illustrated in Figure 3.13 which, in combination with surface maps, provided a range of viewing options for each settlement and set of features.  Visual manipulation of these maps allowed for control over perspective and the definition of vertical scale to highlight and define each feature’s form.  While the features of small settlements (e.g., Sqwá:la / DhRl-6) could all be captured in a single map, multiple maps were produced for sections of  larger settlements (e.g., Xelhálh / DjRi-14) in order to achieve a common level of resolution -- important to the next steps of this process, described below.        3.5.2 Step 2a - defining feature outline / perimeter  Using Surfer 8.0, with a set of combined shaded contour and topographic maps, the outline of each feature was drawn around the apex of the rim or high-point of the ground surface immediately surrounding each depression feature (Figure 3.13).  In this way, a polygon was created for each feature that defined its perimeter and established the basis for calculating the area (m2) and analyzing the shape of each of the 114 housepits included in this study.  The process for defining each housepit’s area as an explicit measurement is described, below, in Step 3a.  The process developed to define each housepit’s shape was less explicit and more    85   Figure 3.2. ‘John Mack Slough’ (DhRl-T1) surface map with housepit features.   86  Figure 3.3. Qithyil Island (DhRl-15) surface map with housepit features (F1-5) and apparent plankhouse platform (F6).     87  Figure 3.4. Sqwa:la (DhRl-6) surface map with housepit features (note: the blue -- or dark grey if depcited in black and white -- in F2 and F3 is a factor of surface elevation coloration and does not represent water).  88  Figure 3.5. Th’ewá:lí (DgRl-17) surface map with features (note: southeast bank of Sweltzer Creek not shown).    89    Figure 3.6. Eyxel (DiRi-48) surface map with housepit features.    90  Figure 3.7. Sxwóxwiymelh ‘South’ (Katz) surface map with housepit features.   91Sxwóxwiymelh (DiRj-1 Northern Portion)  Figure 3.8.  Composite surface image of Sxwóxwiymelh (DiRj-1) ‘South’ and ‘North’ with radiocarbon results                          - per Lenert and Lepofsky (2005, 2006) (Note: dashed, shaded, and solid lines represent pipeline,                                    highway, and railway right-of-ways, respectively).   92  Figure 3.9. Shxw’ow’hamel (DiRj-30) surface map with housepit features.    93  Figure 3.10. Xelhálh (DjRi-14) surface map with features.     94         Figure 3.11. Welqámex (DiRi-15) surface map with features (Graesch 2006).      95   Figure 3.12. Hiqelem (DhRl-T2) surface map with features (Lepofsky et al. 2006).   96exploratory in nature.  In some cases, a feature’s shape (i.e., square or circular) was clear and explicitly apparent in the visual appearance of the perimeter line.  In many cases, though, feature shape was not readily apparent from simple visual inspection.  The qualitative process of determining shape more often than not was found to be highly subjective.  A quantitative analysis was developed for the purpose of making shape determinations based on an objective, reliable, and systematic mathematical process as described in Chapter VI.  Figure 3.13. Schematic illustration of the housepit measurement system applied to Xelhálh (east settlement area) shaded contour/topographic map  (with 5 cm topographic intervals).  3.5.3 Step 2b - defining feature dimensions  Using the same set of feature maps and based on the established housepit polygons (i.e., perimeter lines), three additional lines were drawn between bounding edges within each feature   97representing its maximum length, maximum width, and maximum diagonal dimensions (Figure 3.13).  For each feature, a line representing its maximum length was drawn between two points of the feature’s perimeter line, across the longest distance parallel to the long axis.  A second, perpendicular line representing maximum width was drawn between two points of the feature’s perimeter line across the longest distance of the short axis.  A third line, representing the maximum diagonal distance across each feature was drawn between opposing ‘corners’ set farthest apart from one another.  The dimensions of each feature were gathered from measuring the distances of each of these three lines (see Step 3 below).    These measurements represent linear variables describing feature dimensions that can be analyzed in various ways.  They can be viewed as stand alone variables; in combination and comparison with each other; or in combination and comparison with each other and with measurements of area as descriptors of housepit form including size and shape.  For example, all three dimensions will be the same size in features with a perfect circular shape.  The area of that feature within its perfectly circular perimeter outline will match exactly the calculation of area based on the relevant dimension measurements.  Mathematical representations and expressions of perfect shapes, such as circles, squares, rectangles, and ovals, can be used as ideals in comparison with the actual mathematical representation of each feature, by which feature shape can be systematically described and determined.  The specific calculations developed and used in this process are further described in Chapters VI and VII for addressing housepit features and settlements, respectively.  3.5.4 Step 3 - gathering accurate measurements   Measurements of the line-work created in Step 2 were established using ArcView (version 3.2).  The line-work drawn in Steps 1 and 2 were saved as shape-files and exported   98from Surfer to ArcView.  Functions within ArcView were used to translate the line-work for each feature into metric measurements of maximum length, maximum width, maximum diagonal dimension, perimeter distance, and area.  Housepit dimensions (length and width) were measured from ‘rim-crest to rim-crest’ on perpendicular axes as others have done in B.C. (Archer 2004; Arnold 2004; Graesch 2006; Hayden and Spafford 1993; Mackie and Christensen 2004; Wilson and Carlson 1980).  Calculations of length, width, and diagonal dimensions required manually tracing over each line, from end to end, with ArcView’s distance measuring tool.  Calculations of the area and perimeter line distance for each housepit feature were carried out automatically as a factor of ArcView’s shape-file processing capacity.  Accuracy was maximized by zooming in on each individual feature so that it was generally shown as a single polygon and set of line-work, or nearly so.  Three measurements were generally taken for each line, with the average of the measurements serving as the final calculation.  These repeated measurements were generally found to be the same or to vary only within a range less than 10 cm over a 10 meter distance.  Likewise, each feature’s area value correlates with the ‘aperture’ (i.e., area at aperture) defined by the perimeter of the feature at its rim-crest or otherwise the peak of the immediately surrounding ground surface.   This strategy attempts to minimize the effects of in-filling and erosion on the area measurements used.  These taphonomic forces generally cause greater disturbance to the interior shape of housepits including their base and side-walls, in comparison to the apex which better withstands the ravages of time as a structurally more stable element of these features.  Taking dimension measurements from the perimeter -- at the apex -- also serves to reduce the effects of erosion and infilling in the same way.  Also, judging from clear plan views, obvious areas of disturbance such as slumping can be avoided and adjusted for when taking dimension measurements.  The most common taphonomic effects like slumping and infilling tend to   99diminish the size of housepits rather than make them bigger.  If anything, these effects tend to minimize rather than exaggerate the area of housepits.  Maps of each settlement showing individual feature area polygons are presented in Appendix II.  The dimensions and areas of the 114 housepits included in this study are presented in Appendix I.    Variation within these measurements, while present in my dimension calculations, is minimal -- far less so than field-based measurements affected by lack of perspective and visibility impacting on the discrimination of minute topographic variation apparent in the Surfer-based settlement maps.  The mapping process described above permits the inclusion and consideration of a wider range of data in defining feature perimeters and related measurements.  Variation in dimensional data is inescapable no matter how it is determined.  The process used here minimizes that variation and establishes a systematic process, the reliability of which can be readily evaluated and tested by future researchers.   It is necessary to establish some basic assumptions when using archaeological data describing housepits that may have been rebuilt or reconditioned throughout the period of their occupation.  As a basic assumption, I maintain that the physical dimensions of housepits become larger, if affected at all, as a result of rebuilding.  As such, the maps and dimensions of the features analyzed in this study describe the size of each feature at its time of last occupation.  If occupied and occasionally rebuilt over a long period of time, the size of a housepit may not accurately reflect its original dimensions.  Rebuilding episodes or perhaps subsequent building activity nearby may have changed the attributes of a housepit.  Applying these measurements to the initial construction and occupation of each housepit -- some lived in and re-modeled over hundreds of years -- will, if anything, exaggerate the initial size of the house.  Acquiring the data needed to address the construction and taphonomic histories of each and every housepit in my sample is beyond the scope of this study.  While these assumptions are thought not to affect   100the results or interpretations of the qualitative analyses in this study, they should be kept in mind when viewing the distribution of housepit sizes through time presented in Chapter VI.   Having access to the three-dimensional Surfer maps, while aiding significantly in determining some elements of feature form, does not provide sufficiently accurate information to calculate or address housepit depth as a variable in this study.  Accurate measurement of feature depth(s) can be gained only through supplementary excavation locating the feature floor(s) in relation to the ground surface.  As such, I exclude data from the third dimension (depth) as beyond the scope of this study.   Sub-surface testing revealed that the relationship between the modern ground surface and housepit base elevations is not a good indicator of actual housepit depth.  Taphonomic effects of post-occupation erosion and in-filling can significantly obscure the actual depth of the floor surface(s) below the modern ground level.  Useful information on ‘depth’ can be derived at some level from the Surfer maps, and in a few instances I make qualitative references regarding general relationships between measurements of area and volume in Chapter VI.    3.6 Measuring Age  This study required defining the occupations of each settlement based on the grouping housepits that were lived in at the same time.  A sample of housepits was tested within each settlement with the specific intention of exposing stratigraphic profiles and collecting radiocarbon samples from those strata (i.e., occupation surfaces and/or floors) directly associated with each of the housepit occupations.  Establishing radiocarbon dates for each of the mapped features was beyond the financial scope of this project.  Rather, a sampling strategy was employed that could provide tight definition of co-resident housepits and settlement occupations based on analysis of stratigraphy and careful processing of carbon samples from  those collected at any given site.  Prentiss et al. (2008) demonstrated an effective investigation   101strategy that minimized testing and maximized dating of 100 % of housepits at the Bridge River site in the Middle Fraser Region.  Their work represents an ideal for testing settlements that the researchers of the Fraser Valley Archaeology Project would liked to have achieved, and which we hope to achieve in the future through on-going work at the 11 sites included in this study. Seventeen AMS dates from 15 features in six sites were collected specifically for this study through fieldwork carried out in 2005 and reported by Schaepe, Blake, Formosa, and Lepofsky (2006:104-115).8  These included ‘John Mack Slough’ (DhRl-T1), Qithyil Island (DhRl-15), Th’ewá:lí (DgRl-17), Eyxel (DiRi-48), Shxw’ow’hamel (DiRj-30), and Xelhálh (DjRi-14).  An additional six AMS dates and one conventional date were collected from six features at Welqámex (DiRi-15) by Graesch (2006:66).  Lepofsky et al. (2005:4-5) recovered two AMS dates from two features at Hiqelem (DhRl-T2).  Five AMS dates were provided for five features at Sxwóxwiymelh (DiRj-1) by Lenert and Lepofsky, as reported in Lenert (2008) and Lenert and Lepofsky (2005).  These data are presented in Appendix III.  Radiocarbon data from Schaepe, Blake, Formosa, and Lepofsky (2006) is presented in Appendix IV.   All radiocarbon age estimates represent a 95 % confidence level calibrated at two sigmas using Calib 5.0.2 (Stuiver et al. 2005).  Conventional dates resulting from Hanson’s and von Krogh’s work were also calibrated using Calib 5.0.2.  These dates provide a foundation for developing a temporal framework and context for the analyses of housepits and settlements carried out in this study, presented in Chapter V.   3.6.1 Test Excavations: Soil Probes, Auger Tests, and Shovel Tests  The methods by which radiocarbon samples used in this study were collected vary between researchers.  The methods used in extracting radiocarbon samples from the sites that I                                                  8 Beta Analytic processed all the radiocarbon samples from Fraser Valley Archaeology Project included in this analysis.   102was directly involved in testing -- i.e., co-directing the ‘Stó:lō Pithouse Settlement Mapping and Testing Project’ -- constitute a form of “evaluative testing” (Apland and Kenny 1998:12-13).  In comparison to what may be more commonly recognized as a formal excavation process (e.g., Graesch 2006; Hanson 1973; Lenert 2008), this program of testing aimed to minimize both the areal extent of excavation and the recovery of artifacts.  This strategy maximized the recovery of information on the stratigraphy, composition, and integrity of housepit deposits, as well as associated carbon and soil samples while minimizing site disturbance.9   Thus, small-scale shovel-, auger-, and soil probe testing served, variously, as means of executing small-scale excavations aimed at minimizing artifact collection and maximizing the recovery of stratigraphic data and radiocarbon samples.10  This evaluative testing strategy involved,                                                  9 This strategy was guided by the principle in the Stó:lō Heritage Policy Manual (Schaepe and McHalsie 2003:7) of “taking only what you need.”  10 Schaepe, Blake, Formosa, and Lepofsky (2006:9-10) -- “The Oakfield Probes allowed us to collect 25 cm-deep core increments.  Each 25 cm core increment was removed, described, and photographed before continuing down through the deposits.  This continued until we reach the C Horizon.  The Oakfield probe provides an exposed ‘window’ on one side of the coring tube, allowing effective examination, stratigraphic description, and identification and collection of radiocarbon samples from core sample while still contained in the probe.  Depths of the core  were established for each probe, as it progressed downward, allowing for an accurate description of sediment compaction that tended to occur within the core, itself, as a result of pushing the probe into the ground.  Photographs and stratigraphic drawings included the ‘corrected’ depths for each core, based on direct measurements inside the test unit.    Auger and shovel test excavations were carried out stratigraphically when possible. They were excavated in arbitrary 10 cm levels within strata exceeding 10 cm in thickness.  This strategy was more effective for the shovel tests than for the bucket auger tests, because they afforded a better view of the unit’s side walls and base during excavations.    Auger tests were usually excavated in 15 cm increments, the length of the auger bore.  This precluded examination of layers or stratigraphic contact zones within the 15 cm span of the auger.  All excavation units were precisely plotted and provided ‘real-world’ three-dimensional provenience (UTM / mASL) from established stations on the site using the Leica Total Station, as noted above.  Each test type was differentiated by a code and assigned a unique number which was also recorded on the total station.  Shovel tests had the four surface corners surveyed beginning with the test datum.  The same sequence was repeated at the base.  Auger tests had four surface points taken around the perimeter beginning with a point above the best profile for the test. The same sequence was repeated at the base.  Soil probes had a surface point taken.  Most of the time, the diameter of the hole created by the soil probe was too small for the prism rod so no reliable basal readings could be taken.  The depth of the probe was measured by tape and calculated by subtracting from the surface elevation reading.  Information on the excavations was maintained on excavation and sample forms created for this project.  All units were given ‘numbered’ designations (e.g., ST-1) on a site-by-site basis (e.g., DgRl-17-F8-ST-1; DgRl-17-F3-ST-2; DgRl-17-F20-ST-3), progressively increasing by test type throughout the testing program at each site (as opposed to a feature-specific numbering system).  Excavation unit locations are provided in the data tables presented in Appendix I - Mapping Data.  Elevations were established for the surface and base of   103minimally, probing with a 2 cm diameter Oakfield soil probe, and/or augering with a 13 cm diameter bucket auger or digging square or rectangular shovel tests generally measuring 20-25 cm per side.  Where possible, the least obtrusive testing methods were used to recover data.  Existing soil exposures and profiles with visible stratigraphy were opportunistically investigated (e.g., DhRl-15-F6).  All test excavations terminated at the C horizon which was culturally sterile -- with the exception of a few features, including Feature 13 at Xelhálh, excavation was halted to minimize unnecessary impacts.  High precision three-dimensional proveniences were established for the locations of all collected carbon samples (Schaepe, Blake, Formosa, and Lepofsky 2006).  Some of the detailed profiles from tested housepit features are included for reference in this thesis (Appendix V).11 Tests were generally placed within the central floor area (generally within a 1 m radius circle) of each feature.  Systematically placing tests this way provided some ability to compare patterning in the internal features of housepits, particularly hearth locations.  These testing methods proved effective at exposing carbon-rich hearth deposits associated with the earthen floors in many of the housepits.                                                                                                                                                                   each unit; including each corner of the shovel tests.  All matrix was screened through a 6 mm (1/8”) mesh size in order to recover all archaeological material.  These collections were recorded according to site number, test unit designation, level/stratigraphic layer, and/or depth below ground surface.  Artifacts discovered in situ were recorded three dimensionally using the total station.   Paleobotanical samples were collected when possible, but these collections were limited by the small size of the test excavation units.  Paleobotanical analyses were carried out on the soil sample collected from DiRj-30-F12.  Profiles from at least one wall per unit were drawn and photographed.  A hand-held digital camera was used to take photographs of each excavation unit (overview location) and exposed profiles, including detail profile shots from within each unit (taken by reaching down into and photographing the ‘inside’ of the unit, progressively from the top to bottom, and, when necessary, specific parts of the profile).  Plan views were drawn and photographed where features were noted in the base of any excavation unit.”  Additional information on the testing strategies and methods employed at the other settlements included in this study can be found in Graesch (2006), Lenert (2007), Lenert and Lepofsky (2005), Lepofsky et al (2005), and Ritchie (forthcoming M.A. thesis).   11 Additional information on housepit stratigraphy can be found in Graesch (2006), Lenert (2007), Lenert and Lepofsky (2005), Lepofsky et al (2005), Ritchie (forthcoming M.A. thesis), and Schaepe, Blake, Formosa, and Lepofsky (2006).    104I use the term ‘floor’ as representing a discrete living surface of an earth-floored house.  Floor strata are best thought of as a zone developed over a period of time and comprised of debris from that period of occupation mixed together to create that stratum.  These floors are not like the hardened surfaces associated with clay or wood, but rather represent a somewhat broader timeframe of living that permeates the softer earthen surfaces on which people lived.  These were recognized in many housepit features as dark, often black, bands developed by the accumulation of debris on that surface, commonly creating an occupation zone no more than a centimeter thick.  Floor deposits tended to be carbon-rich and yielded many carbon samples, of which only a small portion were dated, limited by funding.  These carbon-rich floors may be the result of encountering numerous individual or intersecting hearth features in our testing.   Multiple floors overlaying and separated from one another as discrete strata were common (e.g., DiRj-30-F18).  Multiple floors within a single housepit tended to be composed of distinct materials, with each floor having different attributes (e.g., compaction, color, and consistency).  Episodes of floor surface rebuilding apparently involved adding a capping layer of construction fill over the old floor that could serve as the base of the next floor.  Material associated with these floor renovation episodes constituted distinct strata between floors and were often no more than a centimeter thick.  Basal and terminal occupations were generally obvious as distinct strata set against B horizon or C horizon sediments, and generally unencumbered by pre-housepit occupation midden.  Well-stratified deposits were encountered in most of the tested housepits, with the exception of those at Th’ewá:lí (DgRl-17) which was composed of a thick deposit of midden accumulation throughout the site.  This allowed us to collect radiocarbon samples that were reliably associated with floors, and in some cases we were able to recover samples from a superimposed sequence of floors in the same house.     105In summary, the mapping and testing results of the Fraser Valley Archaeology Project, including all 11 documented settlements, represent 10 % of the housepit settlements documented in the region as of 2006.  The ‘Stó:lō Housepit Mapping and Testing Project’ alone documented eight of these settlements and accounted for a 7 % sample of housepit sites, and approximately 12 % all recorded housepits in the region.  The radiocarbon results from this project effectively doubled the number of dated housepit settlements in the region.     106CHAPTER IV - HOUSEPIT SETTLEMENT DISTRIBUTION    Viewing the mainland Gulf of Georgia Region as a network of transportation and communication systems provides a backdrop for exploring the arrangement of housepit settlements within this broad landscape.  For thousands of years, the two basic means and avenues of communication traditionally used by the Stó:lō included canoeing along water-based transportation systems following rivers and slough channels, and walking or running along over-land trails (Schaepe 1999, 2001a).  Traveling by canoe was the basic and fundamental link between settlements dependent on the network of rivers and sloughs within the region.  These routes provided a fundamental link between settlements within and beyond the region, as influenced by the unique geographic structure of the lower Fraser River Watershed.   I view the relationship between geography, transportation, and communication as integrated within the Stó:lō cultural landscape.  Waterways and pathways used as transportation and communication routes are, thus, significant elements of interaction affecting community formation within a social-spatial framework.  I explore broad-scale spatial and temporal relations between housepit settlements based on their position within this network described by travel distances. Watershed landforms and river channel locations are significant elements of environment brought into cultural appreciation, through practice, as avenues of interaction.  These features are integrated into cultural landscapes as elements of the economies of transportation and communication.  Travel routes are recognized and applied in analyses throughout North America as important links to understanding social interaction and relationships between settlements (e.g., Ames 2002; Earle 1991).  The structure of water-based routes of transportation and communication traditionally navigated by canoe throughout the Gulf of Georgia Region forms the basic framework of this spatial analysis.  Measurements incorporated into this analysis reproduce the waterways and pathways of the region rather than using   107arbitrary straight line measurements blanketing an undifferentiated space.  The locations of major waterways and pathways within the study area remain largely the same now as they were over the past 3,000 years -- identifiable and measurable as a foundation for this analysis.   Terrestrial travel routes factor into this analysis, affecting those settlements not directly accessible by water.  A number of settlements throughout the region are not accessible by canoe, including those located in the upper reaches of the lower Fraser River Watershed beyond the entrance to the lower Fraser River Canyon; the majority of the Chilliwack River Valley; and the majority of the Chehalis River Valley.  Water-transport to and from these places was limited, if impossible at times, due to the local hydrology -- thus requiring use of trails as a regular or primary means of travel.   Within this landscape-based framework of transportation, I maintain a basic assumption that interaction regulating the flow of information, access to resources, and communication is a significant factor of community formation.  It is critical to note that different modes and means of transportation drastically effect the way in which this basic assumption becomes manifest in the tempo and scale of interaction.  For example, the development of canoe technology played a significant role in the development of complex social organization and chiefdom-level authority among the Chumash of southern California (Arnold and Graesch 2004).  Water-based travel is recognized as facilitating a tempo and scale of interaction far more intensive and extensive than land-based travel (Ames 2004).  Water-based travel using canoes serves to move more people and things over greater distances far more efficiently and effectively than walking along trails.  This assumption is substantiated by numerous Stó:lō and Coast Salish ethnographies (e.g., Duff 1952; Elmendorf 1993; see Peterson and Drennan 2005:5-6).  My objective is to quantify, explore, and define cumulative, ‘longue duree’ patterns of Stó:lō-Coast Salish housepit   108settlement relative to a system of transportation along the waterways and/or pathways particular to the landscape of this region.  4.1 Mapping Settlement Location and Transportation Systems The Fraser River, linking 24 watersheds within its lower 200 km stretch, is the central feature of intra-regional transportation and communications historically affecting inter-settlement relations among the Stó:lō.  The capacity for this river system to sustain interaction between watershed-based populations of settlements is unsurpassed and largely unparalleled by any other region of the Northwest Coast.  This analysis defines sub-set portions of the region based on spatial relations between 112 housepit settlements set against a framework of regional transportation routes.  The results of this quantitative exploration provide a context for the more detailed analyses of the 11 settlements and corresponding housepits on which the rest of this study focuses.   Plotting the geographic distribution of the 112 housepit settlements (Figure 4.1), a 100 % sample of all recorded housepit settlements in the region as of 2006, reveals several clear patterns.  It is immediately apparent that a large majority (n=106; 95 %) of housepit settlements are located in the upriver portion of the region, from the Central Fraser Valley to the east.  Many settlements in this ‘upriver group’ are located along the Fraser River although some are located, quite far up tributary drainages including the Chilliwack and Harrison rivers.  Only five recorded housepit settlements are located near to the mouth of the Fraser River and its confluence with the Georgia Strait, constituting a small ‘downriver group.’     109 Figure 4.1 The spatial distribution of 112 housepit settlements throughout the mainland Gulf of Georgia Region, showing a pattern of downriver and upriver settlement groupings.  The definition of basic ‘upriver’ and ‘downriver’ groups is not a factor of differences in survey coverage or data documenting housepit settlements in the region.  While significant ground disturbance is a factor of 20th century urbanization along its lower reaches, the Fraser River corridor was subject to similar archaeological survey coverage both up- and downriver from the Central Fraser Valley, throughout the ‘gap’ separating these two groups of housepit settlements.  Downriver people of the Lower Fraser Valley and Delta built ‘on-ground’ plankhouses as their predominate form of housing, generally establishing settlements without pithouses.  In the analyses that follow, I explore how patterning can be further identified and refined beyond this impressionistic picture of broad, upriver and downriver housepit settlement groupings.     110 Figure 4.2.  Schematic illustration of waterways (primary and secondary) and pathways throughout the region.  As a foundation to this analysis, I established a basic framework of two types of transportation routes, waterways and pathways, describing the region (Figure 4.2).   The locations of these routes are based on the physical structure of the lower Fraser River Watershed, reconstructed as of 1800 A.D (Schaepe 2001c:19), prior to the industrial modifications that now affect this landscape.  Waterways are sub-divided into primary and secondary waterways.  Terrestrial pathways are defined as travel routes beginning from the point where water-travel is recognized as no longer possible.  I made these determinations based on my knowledge of the region’s landforms and waterways as informed by geography, oral history, historic accounts, and personal experience of extensive travels throughout this country-side.  This schematic provides the conceptual backdrop against which I plot all 112 pithouse settlements used in this analysis.    1114.2 Mapping Transportation Networks and Housepit Settlement Locations  Quantifying housepit settlement locations along the region’s network of travel routes followed a basic framework of establishing the travel distance (km) following an up-river travel route starting at the mouth of the Fraser River and ending at each of the 112 settlements.  I set the starting point of this route at the mouth of the Fraser River as marking the access point to this intra-regional travel network.  Travel beyond this point and within the Georgia Strait provides access to many neighboring regions throughout the broader ‘Coast Salish Sea’ and Central Northwest Coast.    Total travel distance measurements were established according to a system of travel between the mouth of the Fraser River and each settlement.  This measurement was sub-divided, as determined by each specific journey, into three possible sequential parts defining travel distance(s) along: (1) the ‘primary water route’ (i.e., the Fraser River); (2) a ‘secondary water route’ (i.e., a tributary of the Fraser River); and (3) a ‘terrestrial route’ or pathway.   The ‘primary water route’ value provides a measure of travel distance up the Fraser River as a required initial leg of travel common to all 112 settlements.  This measurement identifies travel distance to either a settlement, directly, or to a point where travelers exit from the Fraser River as a central feature of this system and continue upriver along a secondary waterway.  The ‘secondary water route’ value is required for those settlements not directly located on the banks of the Fraser River but otherwise along a tributary waterway such as the Harrison, Sumas, or Chilliwack rivers, or various slough channels of the Central Fraser Valley.  This value is a measure of distance traveled along a tributary waterway between its confluence with the Fraser River and the settlement itself, or a point of connection to a pathway leading to the settlement.  The ‘terrestrial route’ value is a measure of distance traveled on foot as a final leg of the journey to any settlement requiring such a shift to a terrestrial means of travel -- including those in the   112upper Chilliwack and Chehalis river valleys.  All travel distance measurements are thus modeled on upriver travel by canoe from a starting point (0 km) at the mouth of the Fraser River, and increasing as one follows the respective sequence of travel routes needed to reach each of the 112 settlements.  Measuring travel distance serves as a proxy for measuring social-spatial relations between settlements, akin to the ‘anchored radiance’ model (Miller 1999) of Coast Salish settlement.  Primary water route travel distances define locations of ‘exits’ where travelers the left the Fraser River either to reach a settlement directly or take a subsidiary route.  Analysis of the distribution of primary water route travel distances provides insight into these ‘exit’ locations as prospective clusters of points of connection or ‘hubs’ of travel and communication along the Fraser River.  Such hubs define clusters or groupings of proximate settlements -- either co-existing or other representing longer-term patterns of use and occupancy (Cannon 2002; Mackie 2003).  Hubs, representing more densely or consistently occupied portions of the landscape, indicate potential ‘central places’ of interaction.  Defining settlements associated with secondary or tertiary water route travel distances is also insightful as a measure of social-spatial distance away from the central Fraser River corridor.  As travelers enter tributary watershed systems, they enter more locally-oriented networks of travel and communication -- one step removed from direct connection to the inter-regional waterway.  Settlements requiring terrestrial travel (i.e., with terrestrial travel distances) are the most peripheral to this system and maintain the greatest degree of social-spatial separation from the intra-regional network of relations.    Defining the region by travel distances, while based on a topographic foundation, helps recognize a topologic view of the landscape.  Travel time is a significant factor of this process and is linked directly to the concept of tempo, discussed in Chapter II, as a factor of the   113community formation process.  The tempo of travel (i.e., travel time), moreso than distance, influences the potential for communication and affects the perception proximity between settlements.  Travel time, of course, would vary between up-river and down-river trips depending upon current, tide, wind speed, weight, technology, and so on.  Distance measurements used in this analysis represent one-half of a round-trip journey from the mouth of the Fraser River to each settlement.  In this model, travel conditions similarly affect all travelers on their journey upriver.  Distance serves as a proxy for travel time, assuming a constant speed of travel -- equalizing time as a factor in this analysis.  More complex modeling, beyond the scope of this study, is required to create a truly topologic perspective of the region based on factors of travel distance, travel time, and different means and modes of transportation. Recognizing travel time in conceptual terms does, however, provide insight into a topologic view of the region’s landscape defined as a factor of travel-distance:travel-time ratios.  The impact of time on a topologic landscape would most significantly affect the transition from water to land-based travel.  This transition would effectively warp and ‘extend’ physical space relative to the increased time:distance ratio for settlements accessible only by trails.  Conceptual (topologic) space grows as travel time increases.  Topologically, describing social-spatial distance, settlements accessible only by terrestrial means would be farther removed than water-accessible settlements.  Topologic warping of the region would, considering the distribution of housepit settlements, most significantly affect the Chilliwack River Valley, Chehalis River Valley, and lower Fraser River Canyon, as noted above.        An objective of this analysis is to define groups of settlements based on measureable differences in relationships between three travel distance variables -- primary water route, secondary water route, and terrestrial route.  This approach sets the results of bivariate statistical analyses of these variables against a geographic landscape.  The geographic context provided in   114plotting these statistical descriptions of housepit settlements is critical to the interpretation of these findings.  The inventory of documented housepit settlements used in this study was plotted on a 1:180,000 scale map approximating the region’s geographic landscape circa 1800 A.D.  Travel distances were measured in kilometers using a digital planimeter (Scalemaster Classic v2.0 Digital Plan Measure) set to this scale (Appendix VI).  In Figure 4.1 and subsequent maps in this chapter, X- and Y-axis scatterplot coordinates represent geographic UTM mE and mN measurements, respectively.  This geographic referencing provides a means of coding settlements according to the results of these statistical analyses and plotting them in their real world, geographic locations.    I used an Exploratory Data Analysis (EDA) approach and methods as a foundation to the quantitative analyses carried out in this analysis (see Chapter III).   I use stem-and-leaf graphs and box-plots as core analyses in the EDA toolkit (Drennan 1996; Tukey 1977). These analyses resulted in the identification of a number of settlement groups which I coded by type and projected on the regional maps presented below.  As proxies for social relations, housepit settlement groups identified in this analysis reveal both the landscape of travel and communication networks.    4.3 Analyzing Primary Water Travel Route Distances  The first step in defining housepit settlement groups focused on primary water route travel distances for all 112 settlements in the region.  This analysis readily quantified the ‘impressionistic’ recognition of downriver and upriver groups noted above.  These groups appear in the structure of these data when shown as stem-and-leaf and box-plots (Figures 4.2 and 4.3).  The main point of Figures 4.2 and 4.3 is to show that the total group of housepit settlements can be initially sub-divided into two basic groups based on primary travel distance   115values with those less than 11 km and those greater than 63 km respectively representing downriver and upriver groups of settlements.  The shape of the overall group is downwardly skewed by five housepit settlements with travel distance values under 11 km corresponding (sites DgRs-1, DgRs-14, DhRs-2, DhRs-275, and DhRt-T1) -- representing five extreme outliers within this group.  These five settlements represent the ‘downriver group.’  The remaining set of values represents the ‘upriver group.’  The downriver and upriver groups are separated by a spatial gap or ‘break’ of 53 km within which there are no recorded housepit settlements. Primary Water Route  Travel Distance (km) All Settlements (n=112)  Frequency   Stem & Leaf   (Dist. in km)      5.0        Extremes    (=<11 km)      1.0            6.  3      2.0            7.  89      6.0            8.  346999    16.0            9.  4888888888888888    18.0          10.  000000000477777777      3.0          11.  448      9.0          12.  014444467      3.0          13.  477      7.0          14.  0335679    17.0          15.  01122344555556778      7.0          16.  1444567    16.0          17.  1223455566667788      2.0          18.  01                       Stem width: 10.0    Figures 4.2 (left) and 4.3 (right).  The structure of primary water route  travel distances (km) shown in a stem-and-leaf plot and box plot, respectively.  The primary water travel distances of upriver group settlements range from 63-181 km.  The stem-and-leaf plot identifies at least two additional groups of upriver settlements with peak measurements between 90-100 km and 150-170 km.  Overall, the spatial distribution of downriver (n=5) and upriver (n=107) settlements form the first two groups identified in this analysis (Figure 4.4).  I cut the downriver group of settlements from this study in the process of   116narrowing my set of data and scope of analysis within the region.  The upriver group forms a spatially clustered and robust set of settlements on which I focus the remaining analyses.  Figure 4.4.  Spatial distribution of downriver and upriver housepit settlement groups (n=112).  Three additional sites (DhRo-25, DjRl-3, and DkRn-5) were culled from the upriver group of settlements.  The cases were removed from this analysis for the following reasons.  The validity of the housepit settlement at DhRo-25, the case with the lowest (i.e., westernmost) ‘primary water route’ value (63 km) of the upriver group located in the Stave River Valley, is highly suspect and lacking confirmation.  This site was removed from consideration due to its questionable nature.  DjRl-3 and DkRn-5 -- respectively situated in the mid- to upper-Harrison Lake area -- were removed due to their very distant location f