Critical Analysis and Modeling of Mortuary Remains of the ...dmreed/Manuscripts/Reed2001.pdfCritical...
Transcript of Critical Analysis and Modeling of Mortuary Remains of the ...dmreed/Manuscripts/Reed2001.pdfCritical...
Critical Analysis and Modeling of Mortuary Remains of the Ancient Maya
Paper presented at the 66th annualSociety for American Archaeology meetings,
New Orleans, La.April 19th, 2001
David M. Reed ([email protected])
and
W. Scott Zeleznik ([email protected])
Please do not quote or cite without permission of the authors.
We are indebted to the late Jim Hatch for the collection and curation of a large portion of theunpublished field notes, maps, and reports pertaining to the mortuary remains. We are grateful to RebeccaStorey of the University of Houston and Stephen Whittington of the University of Maine for their hardwork analyzing many of the skeletal remains. We also acknowledge the support of William Sanders andDavid Webster of The Pennsylvania State University who directed Proyecto Arqueológico Copán II and laterexcavations that produced much of the data. David Reed was funded during the writing of this paper by theUniversity of Michigan’s Genome Science Training Program (grant T32 HG00040 from the NationalCenter for Human Genome Research of the National Institutes of Health).
ABSTRACT
Mortuary analysis is used for discerning social differentiation and elucidating models of socio- politicalorganization. Previous research among the Classic Maya has often been ambiguous and even misleading, asevidenced by studies where similar data yielded divergent conclusions. A comprehensive and unreporteddatabase (683 Copán individuals) forms the core of our multivariate, predictive model, incorporatingnumerous indicators of social differentiation (e.g., grave goods, grave space, position and treatment, age, sex,paleopathology, paleonutrition, structure association, and settlement pattern). We define links betweenbehavioral remains, inform competing models of socio-political organization, and develop a model foraddressing social differentiation in areas with limited data.
INTRODUCTION
Social organization - the study of its structure and evolution is a cornerstone of anthropology. Yet, itsreconstruction remains a vexing problem in archaeology. Mortuary analysis has long been used as a meansfor elucidating social and political organization, status, and wealth in past societies. This use results from itsunique position as one of the few opportunities in archaeology to examine both individual and aggregatelevel data. However, accurate measurement and representation of underlying dimensions such as individualsocial status is often merely assumed and studies that try to reconstruct categories of people such as classes orpolitical hierarchies frequently produce contradictory results. Before proceeding with mortuary studies, wemust confront several problems which we view in three analytical steps.
First, we have to be clear as to exactly what we are measuring through mortuary analysis. We mustquestion the dimensions that are being expressed through mortuary analysis. If the Maya had no intent ofdisplaying social status or wealth in the mortuary event then we will be hopelessly misguided in using thesedata to construct such a model. Similarly, if some aspects were meant to reflect the position or status of acategory of persons or a mythical person then it becomes more difficult to make claims about the status andmachinations of individuals. Furthermore, not all mortuary ritual is recoverable. Mortuary remains reflectsocial aspects of the associated individual and ritual activities independent of the interred person that reflectthe social context of death. Multiple explanatory frameworks are necessary to understand the nature of ritualbehavior, be it economic, symbolic, or other. The potential flaws in taking a socially uninformed approachto Maya mortuary analysis have been noted by Gillespie (2001), among others. We recognize there will besome confounding factors such as discordance between skeletal remains and epigraphic decipherment, butwe expect the majority of social assignments to fairly portray social dimensions.
Second, how do we measure the desired dimension? Do the data address the goals of the study and howdo we assess their validity and explanatory power? If we are interested in estimating the social status of theindividuals interred and from this exploring other dimensions of society such as differential access toresources then, the selection of data, types of analyses, and their interpretation can produce distinct resultsthat may or may not pertain to our desired goal.
Third, how does the knowledge garnered from a mortuary study fit into our prior reconstructions of thesociety? Mortuary analyses are of little use if they produce results that fit poorly with prior knowledge andestablished models of social and political organization.
For this presentation, we focus on the second analytical step. We believe that accurate, reliable, andplausible measures must be constructed before we can achieve our ultimate aim of reconstructing socialstatus. We explore ways of measuring status groups that hopefully bring us closer to developing a unified,middle-range theory of mortuary practices, social organization and status.
Before describing our data and analyses, one example of contradictory ranking schemes helps to illustratecritical problems. From the sites of Seibal and Altar de Sacrificios a sample of skeletons was duplicated for 41individuals between Gerry’s (1993) and Wright’s (1994) paleodietary studies. From highest to lowest socialrank, Wright assigned ranks, using site by site principal component analyses. Gerry assigned ranks generatedfrom a cluster analysis based on a pan-Maya scoring system. Direct association between their two schemes
2
results in only 6 of 41 cases of equivalent ranks being assigned by the authors, while 7 of 41 were extremely
different. Wright's approach assigned higher status to individuals than Gerry's method, more than 60% ofthe time. Thus, the social status assignments prove too dissimilar to rely on for direct comparison accordingto social status (figure 1).
DATA
Our data of 683 burials were compiled from published material and the field reports and notes ofexcavations directed by Penn State researchers. More than forty features of each burial were recorded,although the completeness of the information for each feature varies considerably. These features were codedinto 51 variables for statistical analysis. We briefly describe some of the more significant variables and theirpatterns here.
First, we coded a set of variables related to the location of the burial in the Copán Valley. Burials arenumerous for all ranks of sites in both rural and urban areas (figures 2, 3, 4). Clearly, there is apreponderance of burials from the highest ranking urban Type 4 sites, but this distribution only slightlyexaggerates the distribution of population throughout the valley. Not surprisingly, burials from Type 1 sitestend to be located farther from the Main Group, while the range and median distance for burials from othertypes of sites are similar (figures 5).
Second, we used a modified version of the typology employed by Welsh (1988) and Diamanti (1991)for categorizing grave space and construction of 584 burials in the sample. Six types of graves have beenidentified at Copán, a range of types similar to those found elsewhere in the Maya area. With the exceptionof burials lacking any kind of treatment, the distribution is monotonically decreasing as is commensuratewith the assumed increasing complexity and labor involved in creating the grave space (figure 6).
Third, we coded a set variables based on the osteological and paleopathological results. In large part, dueto the work of Whittington (1990) and Storey (1992, 1997), the relative age could be determined for 636 ofthe burials and sex could be determined for 282 adult burials. We find a remarkably normal ratio of 51%women to 49% men in the sample (figure 7). The age distribution shows an inordinate number of childrenand infants, as described by Storey (1992), but otherwise appears relatively normal. The distribution of sexby relative age, shows significantly more men than women in the young adult category and more womenthan men in the older adult category (figure 8). Also coded were a number of other skeletal features such astooth filing, inlays, cranial deformation, paleopathological inferences, missing bones, and carbon andnitrogen stable isotope ratios. These features hold potential significance for individual cases, but they arerelatively infrequent or inconsistently reported and do not lend themselves to the statistical analysispresented here.
Fourth, we coded a set of contextual and locational features for each burial. Our analysis confirms thepoint that space in and around residential structures was important for the Maya. The majority of burialswere placed either in the fill of platforms and structures or in the fill of plazas directly in front of structures.However, there were a substantial number of burials placed along the rear and sides of structures, often inmidden debris (figures 9, 10). The overwhelming majority of burials date to the Late Classic period (figure11).
Fifth, we recorded a set variables dealing with the position of the burial in the grave. These variablesshow little correlation with other possible indicators of status. The orientation of the burials is equitablydistributed among the four cardinal directions and relates primarily to the orientation of the associatedstructure (figure 12). The majority of the burials were flexed, which is related most closely to temporal phase(figure 13). Extended burials are more prevalent prior to the Late Classic and disarticulated burials are moreprevalent in the Terminal and Postclassic periods (figure 14). As expected, extended burials are more likelyto be found in tombs. There is a clear preference for burial of people face-up instead of face down, but thereis no preference for left or right sides (figure 15).
Finally, we coded grave goods according to two factors - the parent material of the object and its possibleuse. The placement of the object in the grave was also noted. The limited variety of artifacts in our sampleprecluded a more complex categorization such as that employed by Welsh (1988). Most ceramics from
3
graves tend to be fine wares and ceremonial wares (figures 16, 17). Imported ceramics are rare (figure 18), as
is the case with imported ceramics in the general Copán domestic artifact assemblage. Non-ceramic gravegoods consisted primarily of jade, bone, shell, obsidian, and ground stone. Jade is the most common of theseitems (figure 19); bone, shell and stone are rare (figure 20). We then divided the artifacts into utilitarian andornamental artifacts (figure 21). Interestingly, they are found in roughly similar quantities but withcontrasting distributions. Burials with many utilitarian artifacts tend to have few or no ornamental artifactsand vice-versa. Finally, we recorded the presence or absence of glyphs and animal or human representationson the artifacts in the grave on the pretext that access to such symbols could be restricted. Both are rare and,not surprisingly, closely associated with fine wares and jade objects (figure 22).
ANALYSIS AND RESULTS
The typical approach to mortuary studies is to select a set variables, either a priori or through a datareduction technique such as factor, principal component, cluster, or correspondence analysis. The product isa model that describes or utilizes these clusters. The typology used to construct the original variables, theselection of variables, and the manner in which the variables are reduced and combined becomes extremelyimportant.
We believe that exploratory statistical techniques, such as principal components and correspondenceanalysis, are an important first step in understanding the structure of the data, but they are not necessarilythe best means for selecting variables to include or condense in a model. For example, the simple pairwisecorrelation of the variables in the database shows that there is an high degree of positive correlation betweengrave goods, grave type, locations in structures, structure fill and non-midden contexts, and age. There islittle correlation among the variables and valley- wide location variables such as site type and urban or rurallocation, and there is virtually no correlation between body position variables and the remainder of thevariables (table 1).
Principal components analysis provides further insights into the data (table 2). Grave type and gravegoods including ceremonial ceramics, ornamental artifacts, fine wares, shell and glyph representations loadhighly on the first principal component. Context and structure specific location variables load highly on thesecond and third components. Temporal phase and extended and flexed burials define a fourth component.A fifth component is defined by the valley-wide location variables of site type, urban or rural location, anddistance to the Main Group. Utilitarian grave goods and obsidian and stone materials generally used inmaking utilitarian artifacts load highly on a sixth component. Correspondence analysis, which is oftenconsidered better suited for count data, shows a similar pattern. Variables plotted on the first twodimensions show that the above set of grave goods form a cluster; the context and location variablesdescribing individuals buried in structures and their fill and in locations other than middens and behindstructures form another tight cluster; grave type and age stand apart (figure 23).
However, such exploratory techniques cannot be the end, because their goal is to reduce and partitionthe data using all available data. Instead, we need a technique that helps us to identify one underlyingdimension – one latent variable – that is not entirely dependent on an a priori selection of a few variables.For instance, does the relationship between time phase and extended and flexed burials relay anyinformation about social status? Does the relationship between site type, rural or urban location, anddistance to the Main Group really say anything more than the fact that urban sites tend to be larger andconcentrated in the center of the Copán polity? How do we know which variables are relevant?
We turned to reliability analysis, a measurement modeling technique designed to identify a latentdimension and for constructing a scale or index. Reliability analysis is based on the structure of thecorrelation matrix, but its aim is to determine if the correlational structure allows a set of the variables to besensibly additive. In other words, does there exist a set of variables in the data set that can go together toform a scale? If not, we know that either we have made a poor choice of variables or that there is no singledimension that defines the data.
Several variables in our dataset do combine to define an underlying dimension and a robust scale. Theycan be grouped according to the degree to which they positively contribute to the scale and their inherent
4
similarities. A large alpha value of 0.75 was obtained and can be thought of as an overall correlation. The
alpha if item deleted then represents how the overall correlation would be reduced if the particular variablewere dropped from the scale (table 3).
The first item is grave type. The second group consists of principal grave goods – fine ware ceramics,ceremonial ceramics, ornamental objects, and jade. The third group consists of two sets of interrelatedcontext and location variables. Burials found in structure fill and within structures and conversely burialsnot coming from middens or the backs of structures are important. Age forms a fourth category. The finalcategory is a set of what might be termed “lesser grave goods” that marginally contribute to the scale, due totheir rarity. These artifacts include, glyph and effigy representations, imported ceramics, utilitarian itemsand stone, shell, and bone artifacts. The relationship between variables can be shown in another way bypresenting the degree to which each variable correlates with the resultant scale (table 4).
Now it is up to us as anthropologists to determine if the scale makes sense and measures what weintended it to – social status. Intuitively, we believe so, but we need to return to the actual burials andexamine their distribution across the various variables both included and not included in the scale to findout. Indeed, what may be more surprising is the absence of certain variables in the scale such as site type,urban and rural location, sex, or any of the burial position characteristics.
To accomplish this, each burial was given a mortuary score based on the sum of the standardized scoreon each variable in the scale. We found that 623 of the burials had sufficient data to be given scores. Thislatent variable can be explored further by rank ordering the scores and noting how classes of individuals aredistributed along the scale (figure 24). The most striking feature of the distribution is its linearity. Weconclude that when a complex of factors or variables is considered, it becomes difficult to demarcate specificstatus groups or classes in the society. There appears to have been considerable variability in access to gravegoods such as jade, or fine ware ceramics for example.
For demonstrative purposes, we fitted a trend line. Red crosses indicate exceptional cases – thoseindividuals that depart from the trend line at the tails. What factors make these individuals unique? Asdefined, 75 trend trailers fall below the line at the lower end. Children and infants comprise 95% of theseindividuals, and the remaining 5% are female adults. Over 90% had no grave treatment or were found in pitgraves. However, we find that they are distributed across all locations and associated site types in similarproportion to the overall sample. The upper end of the continuum holds 43 individuals that lie above thetrend line. Conversely, 95% of these individuals are adults and 68% are males, which is significantly greaterthan the total sample. Also, more than half come from tombs. But, these individuals are found at all typesof sites except for the smallest Type 1 sites and in locations of the valley roughly equal to the proportions ofthe overall sample.
We briefly compare our suggested mortuary scale against a common ranking scheme. This figure showsour scores plotted by site type (figure 25). There is a great deal of overlap between site types along thecontinuum. The larger Type 2, 3 and 4 sites show the full range of variability. Only the smallest Type 1sites lack the full range having no individuals scoring above 0.3. Type 2 and larger sites held a large range ofstatus individuals. This point is emphasized in the following figure (26) that shows individual sites orderedby the mean score of the burials from the site. Other than Type 1 sites, the highest score or size of range is abetter predictor of site type than median value due to the great range of variability found at larger sites.
CONCLUSIONS
In conclusion, we believe that the above measured dimension does relate to social status as filteredthrough mortuary practice.
Second, many studies use a single measure or small set of a priori determined indicators to infer status.We find that a small set of factors (e.g., jade, ceramics, site size, or grave type alone) fails to defineadequately the complexity of Maya mortuary practices and the patterns of social and political organizationthat they may portray. Nevertheless, there exist sets of variables that were consistently used in mortuarycontexts, including a varied combination of principal grave goods, the location and placement of the burialwith respect to residential structures, the burial space, and the relative age or life position of the individual.
5
We believe they are reflective of both the specific individual in the cases of age and certain ornamental grave
goods, and the larger social group in the cases of burial space, burial location and context, and other gravegoods. Mortuary data represents one of the few opportunities in archaeology to examine both individualand aggregate level data. Gillespie (2001) has cogently reminded us that mortuary remains may reflect theinteraction between structure and individual, or agent. The interaction of agency and structure helpsexplain why it is difficult to identify any one item, characteristic, or definitive assemblage as indicative ofstatus. Indeed, we see both individual and structural factors interacting in mortuary ritual and theirreflection is a complex status system.
Third, many studies dealing with mortuary data and other tangible remains of Maya society propose afew delineated ranks or status groups. Our data show that, while such groups may have been recognized inMaya society, practically it is difficult to make such an assumption from mortuary data. It is only natural towant to condense the world into a small, neat set of categories and types. Unfortunately the real world,neither past nor present, is so tidy. We question the validity of models that propose sharply delineated socialgroups. The degree of variability in the distribution of individuals along the continuum of the scaledemonstrates the lack of a deterministic nature to the status system. Various sets of grave goods, spatiallocation, grave space, and relative age were somewhat fluid and could have been combined in different waysto demarcate the status of the individual and the social group responsible for the interment.
Fourth, we infer from the great range of mortuary scores at Type 2 and larger sites that these sites andplaza groups at Copán should best be thought of as extended households. Elites, retainers and slaves mayhave maintained fictive kin ties, as was the case in Maya households at the time of contact, but we believesince both the lowest and the highest status individuals are found at these larger sites that they are composedof both kin groups and non-related retainers.
Finally, the next analytical step is to incorporate these findings into general models of social organizationproposed for Copán and the Maya. From a statistical standpoint – having a continuous variable of statusaffords us many more analytical opportunities than a categorical variable. From a theoretical standpoint, weleave you with these questions. Does the variability in grave good associations (such as the contrastingdistributions of utilitarian and ornamental artifacts or the varied use of jade and ceramics) point to theinfluence of other, more subtle dimensions in mortuary practice? What might these models predictindependently and how then can they inform each other? How can this approach be used for other sites andregions?
6
REFERENCES CITED
Diamanti, Melissa1991 Domestic Organization at Copán: Reconstruction of Elite Maya Households through Ethnographic
Models. Ph.D. Dissertation, Department of Anthropology. The Pennsylvania State University,University Park, Pa.
Gerry, John P.1993 Diet and Status Among Classic Maya: an Isotopic Perspective. Ph.D. Dissertation, Department of
Anthropology. Harvard University, Cambridge, Ma.Gillespie, Susan D.
2001 Personhood, Agency, and Mortuary Ritual: A Case Study from the Ancient Maya. Journal ofAnthropological Archaeology 20:73-112.
Storey, Rebecca1992 The Children of Copán: Issues in Paleopathology and Paleodemography. Ancient Mesoamerica
3:161-167.1997 Individual Frailty, Children of Privilege, and Stress in Late Classic Copán. In Bones of the Maya:
Studies of Ancient Skeletons. Edited by Stephen Whittington and David Reed. SmithsonianInstitution Press, Washington, DC.
Welsh, W.B.M.1988 An Analysis of Lowland Maya Burials. BAR International Series, No. 409. Oxford University
Press, London.Whittington, Stephen L.
1990 Characteristics of Disease and Demography in Low-Status Maya from Classic Period Copán,Honduras. Ph.D. Dissertation, Department of Anthropology. The Pennsylvania StateUniversity, University Park, Pa.
Wright, Lori E.1994 The Sacrifice of the Earth? Diet Health and Inequality in the Pasión Maya Lowlands. Ph.D.
Dissertation, Department of Anthropology. University of Chicago, Chicago, Il.
7
Table 1. C
orrelation Matrix of A
ll Variables w
ith Grave T
ype, Fine Ceram
ics, Jade, and Age.
Variable N
ames
Grave
Type
FineC
eramics
JadeA
geV
ariable Nam
esG
raveT
ypeFineC
eramics
JadeA
ge
GR
AV
E T
YP
E.4397 (584)
.4477 (584).3925 (557)
DE
MO
GR
AP
HIC
SG
RA
VE
GO
OD
SA
ge.3925 (557)
.1967 (590).1982 (590)
Fine Ceram
ics.4397 (584)
.4162 (623).1967 (590)
Sex.1478 (257)
.1966 (267).0880 (267)
-.079 (279)
Cerem
onial Ceram
ics.4130 (584)
.8481 (623).4222 (623)
.1770 (590)SP
AT
IAL and T
EM
PO
RA
L LOC
AT
ION
Ceram
ics Total
.4285 (584).9571 (623)
.3959 (623).1830 (590)
Site Type
.0370 (559).0253 (593)
-.018 (593)-.087 (620)
Imported C
eramics
.1191 (584).2379 (623)
.0037 (623).0171 (590)
Site Location (Urban)
-.049 (578).0330 (617)
-.005 (617)-.089 (635)
Jade.4477 (584)
.4162 (623).1982 (590)
Dist. to M
ain Group
.0151 (552).1118 (586)
.0674 (586).0209 (612)
Ornam
ents.4159 (584)
.4393 (623).6380 (623)
.1788 (590)A
ssociated Structure-.038 (562)
-.059 (593)-.064 (593)
-.005 (614)U
tilitarian Items
.1845 (584).3116 (623)
.2082 (623).0975 (590)
Tem
poral Phase
.1318 (569)-.059 (597)
.0412 (597).0549 (626)
Bone
.1083 (584).2701 (623)
.1120 (623).0601 (590)
PO
SITIO
N and P
LAC
EM
EN
T in G
RA
VE
Shell.1441 (584)
.2232 (623).1924 (623)
.0229 (590)D
isart. Position
-.057 (398)-.034 (398)
-.064 (398)-.061 (386)
Stone.1725 (584)
.1383 (623).1779 (623)
.0998 (590)E
xtended Position
.0734 (398).2519 (398)
.2219 (398)-.060 (386)
Obsidian
.1076 (584).1440 (623)
.0926 (623).0166 (590)
Flexed Position
-.007 (398)-.160 (398)
-.116 (398).0844 (386)
Effigy R
epresentation.2644 (584)
.4088 (623).1322 (623)
.0792 (590)B
ack Side.1740 (280)
.1143 (280).1103 (280)
-.147 (279)G
lyph Representation
.2589 (584).4370 (623)
.2984 (623).0782 (590)
Front Side-.047 (280)
.0420 (280)-.032 (280)
.0204 (279)
CO
NT
EX
T and ST
RU
CT
UR
E LO
CA
TIO
NLeft Side
-.118 (280)-.072 (280)
-.093 (280).0296 (279)
Front Location.0467 (537)
-.034 (548).0039 (548)
.0347 (548)R
ight Side-.088 (280)
-.064 (280)-.038 (280)
.0736 (279)P
laza Location.2076 (537)
.1811 (548).1183 (548)
.0185 (548)E
ast Orientation
-.053 (348).0180 (349)
-.127 (349).0472 (342)
Not R
ear Location.246 (537)
.142 (548).124 (548)
.128 (548)N
orth Orientation
.0880 (348)-.041 (349)
.1006 (349)-0.15 (342)
Side Location-.058 (537)
-.093 (548)-.102 (548)
-.089 (548)South O
rientation.0214 (348)
.0901 (349).0548 (349)
-.053 (342)Structure Location
.0894 (537).1227 (548)
.0467 (548).0971 (548)
West O
rientation-.062 (348)
-.042 (349)-.003 (349)
-.048 (342)N
ot Midden C
ontext.279 (520)
.160 (531).111 (531)
.255 (531)P
rim./Second. B
urial.1160 (544)
.0702 (559)-.024 (559)
.0513 (533)P
laza Fill Context
.1237 (520).0044 (531)
.0212 (531).0443 (531)
Multiple B
urial.1129 (554)
.0983 (567)-.007 (567)
-.029 (545)
Structure Fill Context
.1405 (520).1297 (531)
.0828 (531).1361 (531)
correlation (n) p £ 0.001
p £ 0.01 p £ 0.05
Table 2. Principal Component
ComponentComposition
Variable C 1 C 2 C 3 C 4 C 5 C 6
Ceremonial Ceramics 0.87879
Fine Ceramics 0.87784
Total Ceramics 0.86818
Jade 0.74234
Grave Type 0.60729
Ornaments 0.60380 0.38586
Glyph Representation 0.48077 0.38320
Component 1
Grave Type andPrincipal
Grave Goods
Shell 0.42551
Non-Midden 0.89674
Non-Rear Structure 0.87203
Interior Structure 0.34025 0.90715
Structure Fill 0.33664 0.90493
Front Structure -0.34619 -0.62303
Plaza Fill -0.3575 -0.86899
Rear Structure -0.87203
Components 2 & 3
Context andLocation
Midden Context -0.89674
Flexed Position 0.87037
Temporal Phase 0.6193Component 4Body Position
and Phase Extended Position -0.86128
Site Location (Urban) 0.92482
Site Type 0.54815Component 5Site Location
Dist. to Main Group -0.93136
Stone 0.74259
Utilitarian 0.65713Component 6
Lesser Grave GoodsObsidian 0.62525
Loading >0.5 Loading > 0.3 Negative Loading < -0.3
Table 3. Reliability Analysis Results.
Variable Type Variable Scale Meanif Deleted
ScaleVariance if
Deleted
CorrectedItem TotalCorrelation
Alpha ifItem
DeletedGrave Type Grave Type 3.2979 3.5700 .4579 .7296
Ornaments 3.4713 3.5552 .4535 .7294Fine Ceramics 3.5321 3.8234 .4771 .7396Jade 3.5358 3.8046 .3580 .7409
PrincipalGrave Goods
Ceremonial Ceramics 3.5460 3.8553 .4234 .7420Structure Fill Context 3.1987 2.8947 .5448 .7176Non-Midden Context 2.7554 3.1920 .5098 .7198Structure Location 3.1703 2.9029 .5289 .7208
Location andContext
Non-Rear Location 2.7877 3.1486 .4975 .7218
Age Age 3.1241 3.5610 .2946 .7442
Effigy Representation 3.5145 3.7343 .2489 .7459Glyph Representation 3.5691 3.9237 .2475 .7477Utilitarian Items 3.5158 3.8502 .2371 .7464Imported Ceramics 3.5657 3.9670 .2426 .7495Shell 3.5651 3.9672 .1818 .7502Stone 3.5659 3.9702 .1889 .7502
Lesser GraveGoods
Bone 3.5525 3.9434 .1501 .7506Alpha = .7506
Table 4. Correlation of Variables with Mortuary Scale.
Variable Names Correlationwith Scale
Variable Names Correlationwith Scale
Structure Fill Context .6825 (531) Temporal Phase .1005 (597)Structure Location .6749 (548) Multiple Burial .0919 (567)Not Midden Context .6327 (531) Sex .0919 (267)Not Rear Location .6155 (548) Plaza Location .0840 (548)Grave Type .5464 (584) Site Location (Urban) .0806 (617)Ornaments .5436 (623) Site Type .0801 (593)Fine Ceramics .5067 (623) Back Side .0775 (280)Ceremonial Ceramics .4548 (623) Prim./Second. Burial .0610 (559)Jade .4289 (623) Extended Position .0506 (398)Age .4102 (590) South Orientation .0261 (349)Effigy Representation .3406 (623) Disart. Position .0224 (398)Utilitarian Items .3294 (623) East Orientation .0083 (349)Glyph Representation .3098 (623) Left Side -.002 (280)Bone .2474 (623) West Orientation -.010 (349)Imported Ceramics .2472 (623) Associated Structure -.020 (593)Stone .2462 (623) North Orientation -.040 (349)Shell .2356 (623) Flexed Postion -.055 (398)Obsidian .1733 (623) Front Location -.071 (548)Distance to Main Group .1536 (586) Front Side -.076 (280)
Right Side -.098 (280)Plaza Fill Context -.171 (531)Side Location -.220 (548)
Variables ordered by strength of correlation from top to bottom and left to right p ££££ 0.001 p ££££ 0.05
Table 5. Characteristics of Burials at Upper and Lower Ends of Continuum.
Variable Type Variable Name Lower Endn=75
Upper Endn=43
Total Samplen=683
Sex Female 5 (100%) 7 (32%) 144 (51%)Male 0 (0%) 15 (68%) 138 (49%)
Age Older Adult 0 (0%) 5 (13%) 57 (9%)Middle-Aged Adult 1 (1%) 13 (33%) 117 (18%)Adult 3 (4%) 19 (47%) 138 (21%)Young Adult 2 (3%) 1 (2%) 71 (11%)Child and Adolescent 28 (41%) 2 (5%) 172 (27%)Infant 34 (50%) 0 (0%) 90 (14%)
Grave Type Complex Tomb 0 (0%) 9 (21%) 12 (2%)Simple Tomb 0 (0%) 15 (35%) 37 (6%)Crypt 0 (0%) 7 (16%) 50 (9%)Cist 5 (9%) 9 (21%) 117 (20%)Pit 14 (26%) 3 (7%) 253 (43%)No Grave 35 (65%) 0 (0%) 115 (20%)
Site Type Type 4 Site 43 (65%) 23 (58%) 409 (63%)Type 3 Site 9 (14%) 10 (25%) 82 (13%)Type 2 Site 12 (18%) 7 (17%) 98 (15%)Type 1 Site 2 (3%) 0 (0%) 64 (10%)
Site Location Urban - Main Group 9 (12%) 12 (28%) 62 (9%)Urban - Bosque 0 (0%) 3 (7%) 26 (4%)Urban - Sepulturas 56 (76%) 22 (51%) 483 (72%)Pocket 7 (9%) 6 (14%) 78 (12%)Rural 2 (3%) 0 (0%) 24 (4%)
Temporal Phase After Late Classic 0 (0%) 5 (12%) 79 (12%)Late Classic 58 (85%) 36 (84%) 472 (72%)Before Late Classic 10 (15%) 2 (5%) 106 (16%)
Figure 1
BBB
BB
BBB
BB
BB
BB
BB
BBB
BBB
BBB
BB
BB
BBB
Wright Gerry
Altar de Sacrificios
BBBBB
BBBBB
BBBB
BB
Absolutedifference
low
est
hig
hes
t
low
est
hig
hes
t
sam
e1
dif
fere
nce
2 d
iffe
ren
ces
3 d
iffe
ren
ces
BBBB
BBBB
BB
BBBBBBBBBBB
BB
BBBB
BBBB
BB
BB
BB
BBBBBBBBBBB
BBBBBB
Wright Gerry
Seibal
BBBBBBBBB
BBBBBBBBBBBBBB
BBBB
Absolutedifference
low
est
hig
hes
t
low
est
hig
hes
t
sam
e1
dif
fere
nce
2 d
iffe
ren
ces
3 d
iffe
ren
ces
�
������
�
��
���
������������
��������������������
�������������������
����������������������������������������������������������������������������������������������������
��
�
�
���
�
�����
�
��
�
��
�����
��
�
�
�
�
�
����� ���������
�����������
���������
����������������������������������
���
��
�������������������������
�������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������
���
��
�
�
� � � � � � �� ���� ���
��� ������������������
����������� �����
�� �� ��
�����
����������
����� ������
�����
�!���!�"
��#��
��#��
���$����
���$����
"����"���"�����%�� �%���
�%��� �������������"���&
���������������
�'�&���&
��%����%��"��%����%��
��%����%�"��%�� ��%���
��%���
��(��
'��)
��*����
���$��� ��� ���� �
���&
�&)$���
��+��
�����
�
+�,�����
-.���/��0��� ���1� 2�����������3� �)���
Figure 3
Non-Md.Single Md.
Type 1Type 2
Type 3Type 4
Main Group Missing
Site Type
0
100
200
300
400
Associated Site TypeN
um
ber
of
Bu
rial
s
Figure 4
RuralPocket
Urban-BosqueMain Group
Urban-SepulturasMissing
0
100
200
300
400
500
Location in Valley
Nu
mb
er o
f B
uri
als
Figure 5
5843
545
2961
3546
137
1004
3416
469
1079
3467
240
1154
Type 1 Type 2 Type 3 Type 4
Site Type
0
2000
3000
4000
5000
6000 Farthest
NearestMedian
Distance of Burials to Main Group by Site Type
Dis
tan
ce (
met
ers)
1000
Figure 6
NonePit
CistCrypt
Tomb-SimpleTomb-Complex
Missing
Grave Type
0
100
150
200
250
300
Distribution of Grave TypesN
um
ber
of
Bu
rial
s
50
Figure 7
Female144
51.1%
Male138
48.9%
SexMissing = 401
Figure 8
AdolescentYoung Adult
Adult-Unspec.Middle-Aged
Older Adult
Age Categories
0
10
20
30
40
50
60
female
male
Frequency of Relative Age by SexN
um
ber
of
Bu
rial
s
Figure 9
Superstr.Substr.
FrontRear
SidePlaza
No Assoc.Missing
Burial Location
0
50
100
150
200
Burial Location with Respect to Structure
Nu
mb
er o
f B
uri
als
Figure 10
FeatureCollapse
MiddenFill-Plaza
Fill-Substr.Fill-Superstr.
Missing
Context
0
50
100
150
200
250
Context in which Burial was Placed
Nu
mb
er o
f B
uri
als
Figure 11
PreclassicEarly Classic
Middle ClassicMid./Late
Late ClassicTerm./Post.
Missing
Time Phase
0
100
200
300
400
500
Time Phase Assigned to BurialsN
um
ber
of
Bu
rial
s
Figure 12
NE15
4.3%
E78
22.3%
SE15
4.3%
S80
22.9%SW9
2.6%
W55
15.8%
NW14
4.0%
N83
23.8%
DirectionFrequencyPercent of Total
Frequency of Burial Missing = 334
Figure 13
BundledDisarticulated
ExtendedFlexed
Missing
Position
0
50
100
150
200
250
300
350
Frequency of Body Positions in Graves
Nu
mb
er o
f B
uri
als
Figure 14
0%
flexed extended disarticulated
Body Position
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Early Classic
Middle Classic
Mid./Late
Late Classic
Term./Post.
Percentage of Body Positions by Time Phase
Per
cen
t o
f B
uri
als
Figure 15
Front Sitting Right Left Back Missing
Side
0
100
200
300
400
Frequency of Side of Body in Grave
Nu
mb
er o
f B
uri
als
Figure 16
439
105
31 1 2 7 0 1 2 0 0 2
6
01
23
45
67
89
1011
12Missing
Absolute Number
0
100
200
300
400
500
Fine Ware Ceramics
Nu
mb
er o
f B
uri
als
Figure 17
518
6
1 1 9 4 2 0 0 1
6
01
23
45
67
89
Missing
Absolute Number
0
100
200
300
400
500
Ceremonial Ceramics
Nu
mb
er o
f B
uri
als
Figure 18
587
32 0 0 1 1
6
0 1 2 3 4 5 6 Missing
Absolute Number
0
100
200
300
400
500
600
Imported Ceramics
Nu
mb
er o
f B
uri
als
Figure 19
44
NoneFew/Small
ModerateMany/Large
Missing
Relative Quantity and Size
0
100
200
300
400
500
600
700
Jade
Obsidian
Jade and Obsidian Artifacts
Nu
mb
er o
f B
uri
als
Figure 20
0 1 2 3 4 5 6 7 8 9 Missing0
100
200
300
400
500
600
Bone
Shell
Stone
Bone, Shell, and Stone Artifacts
Figure 21
None Few Moderate Many Missing
Relative Quantity and Quality
0
100
200
300
400
500
600
Ornament
Utilitarian
Ornaments and Utilitarian Artifacts
Nu
mb
er o
f B
uri
als
Figure 22
Absent Present Missing0
100
200
300
400
500
600
700
Glyph
Effigy
Glyph and Effigy RepresentationsN
um
ber
of
Bu
rial
s
Figure 23
AB
Ce
nM
Sf
EfCf
Gl
Gr
I
J
nR
Ls
Or
Sh
St Ut
-0.5 0 0.5 1 1.5 2 2.50
Axis 1
-1
-0.5
0
0.5
1
1.5
Correspondence AnalysisScale Variables, n=494
Ax
is 2
Age
Grave Type
Context andLocation
Grave Goods
Variable KeyAge ABone Artifacts BCeremonial Ceramics CeContext Not Midden nMContext Structure Fill SfEffigy EfFine Wares CfGlyph GlGrave Type GrImported Ceramics IJade Artifacts JLocation Not Rear nRLocation Structure LsOrnamental Artifacts OrShell Artifacts ShStone Artifacts StUtilitarian Artifacts Ut
Figure 24
Burials Ordered by Mean Mortuary Score
n = 623
Trend LineR 2 = 0.9503
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Individual Burials
Mea
n S
core
Figure 25
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
mortuary scale
Main Group
Type 4
Type 3
Type 2
Type 1
Figure 26
Site [Type]
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0
0.2
0.4
0.6
0.8
High
LowAvg.
Range
Sites with 3 or More Burials Ordered by Average Mortuary Score
Mo
rtu
ary
Sco
re
10L-
2[4]
9M-2
7[2]
9M-1
8[3]
8L-1
2[4]
9N-8
-I[4]
des
tro
yed
8L-1
0[3]
9N-8
-C[4
]9N
-8-K
[4]
3O-8
[2]
9N-8
-D[4
]11
L-12
/13[
3]
Gro
up
5[4
]9N
-5[1
]
7M-8
[2]
9M-2
2A[3
]9N
-8-E
[4]
9N-8
-A[4
]9N
-8 T
ot.
[4]
11L-
8[2]
3O-7
[1]
11E-
2[3]
9M-2
4[1]
9M-2
2B[2
]9N
-8-H
[4]
10L-
3[2]
9N-8
-F[4
]9N
-8-B
[4]
10E-
6[4]
8N-1
1[4]
El B
osq
ue
Mai
n G
r.[5
]
99a/
18-2
[1]
34a/
12-2
[1] Ran
ge o
f Sc
ore
s