UC Riverside
UC Riverside Previously Published Works
Title
4 Community Labor and Laboring Communities within the Tiwanaku State (C.E. 500-
1100)
Permalink
https://escholarship.org/uc/item/1wt7237p
Journal
Archeological Papers of the American Anthropological Association, 28(1)
ISSN
1551-823X
Author
Becker, Sara K
Publication Date
2017-07-01
DOI
10.1111/apaa.12087
Peer reviewed
eScholarship.org Powered by the California Digital Library
University of California
4
Community Labor and Laboring Communities
within the Tiwanaku State (C.E. 500–1100)
Sara K. Becker
University of California, Riverside
ABSTRACT
Understanding how work was managed and who participated in state-level societies can help elucidate daily
activities as well as community development within an emerging complex society. Tiwanaku, with multiethnic
neighborhoods in the Titicaca Basin, Bolivia and colonies near present-day Moquegua, Peru, provides a comparison
of labor between groups. Specific skeletal evidence of activity (i.e., musculoskeletal stress markers and osteoarthritis)
was evaluated to infer how habitual activity varied within this state. Labor rates show that laborers did not work
at the behest of elites and results suggest instead, that people worked as reciprocal laborers in a guild-like system.
[Organized labor, Bolivia, Peru, Musculoskeletal stress markers, Entheses, Osteoar thritis, Practice theory]
T
he organization of labor as part of resource manage-
ment is one way to understand the development of
complex societies. People i n the past worked at various jobs,
creating communities based around tasks, such as craft pro-
duction or farming, as well as building homes and home-
lands for themselves (e.g., Brumfiel 1991; Costin 2004;
Costin and Earle 1989; Crumley 1987, 2007; Crumley et al.
1987; D’Altroy 1992; D’Altroy and Earle 1985; Earle 1997;
Kunen and Hughbanks 2003; Levy 2006; Moseley 1975).
Often, these workers are defined archaeologically through
the product of their labor, such as monumental architecture,
ceramics, or lithic tools. While this evidence does provide
information about people’s daily life, additional knowledge
can be gained from a bioarchaeological methodology that
uses the evidence of labor and activity on human skeletal
remains, complimenting an artifactual approach, and en-
gaging with the actual individuals who lived this lifeway.
Remembering that these people were once a community is
also essential. At the very least, a community involves some
kind of shared background where group members recognize
each other as different from others (i.e., “us” versus “them”)
(cf., Barth 1966; Goldstein 2000a; Gupta and Ferguson
1992; Isbell 2000; Reycraft 2005; Yaeger and Canuto 2000).
How to evaluate group membership can become compli-
cated when skeletal remains are the focus, as bioarchae-
ologists may face challenges associated with an incom-
plete burial record due to issues like skeletal preser vation,
sample representativeness, or choice of excavation location
(e.g., Cook and Buikstra 1979; DeWitte and Stojanowski
2015; Gowland 2006; Halcrow and Tayles 2008; Hoppa and
Vaupel 2002; Roberts and Mays 2010; Sofaer Derevenski
1994, 1997; Waldron 1994; Wood et al. 1992; Wright and
Yoder 2003). In addition, questions posed by Canuto and
Yaeger (2000) in The Archaeology of Communities on how
to define past communities still stand, and must be reen-
gaged from a nuanced perspective on how we can define
“community” from skeletal remains and burial populations
(see Chapter 2 of this volume by Kakaliouras for a review).
Of the theoretical approaches to community Yaeger and
Canuto (2000:3) describe in their introductory chapter, prac-
tice theory provides a useful way to address group labor and
civic membership, as people’s lives can become inscribed on
their physical bodies via their regular daily habits (Bourdieu
1977; Budden and Sofaer 2009; Merleau-Ponty 2013; Sofaer
2006). Through the repeated practice of laboring, the house-
hold tasks executed and the occupations people perform can
Volume editors: Sara L. Juengst and Sara K. Becker, Volume 28: The Bioarchaeology of Community
ARCHEOLOGICAL PAPERS OF THE AMERICAN ANTHROPOLOGICAL ASSOCIATION, Vol. 28, pp. 38–53, ISSN 1551-823X,
online ISSN 1551-8248.
C
2017 by the American Anthropological Association. All rights reserved. DOI: 10.1111/apaa.12087.
Community Labor and Laboring Communities 39
set them apart from others within the larger society. These
jobs may have been done at the behest of leaders of their
society for maintenance of their civilization, or for members
of their peer group as part of their social or familial require-
ments. Whatever the reason, this agent-oriented approach
considers the people performing these activities as part of
a past community. Moreover, Yaeger and Canuto (2000:5-
6) note that a within-region but supra-household pattern,
along with a limited time frame of cultures studied, makes
a good and flexible way to discuss community archaeologi-
cally while also avoiding reification and essentialization of
this concept.
Using these i deas on societal formation, labor, and com-
munity, this chapter focuses on the Tiwanaku civilization.
This culture formed a state-level society around C.E. 500
in the Andean highlands region of the Lake Titicaca Basin,
and expanded (ca. C.E. 500–650) into a lower elevation
colony near present-day Moquegua, Peru (Figure 4.1) be-
fore its collapse in both areas around C.E. 1100
1
.Myre-
search addresses patterns of habitual labor observable on
the bones of people who lived in the Tiwanaku state using
specific skeletal evidence of activity (i.e., musculoskeletal
stress markers and osteoarthritis) in order to understand what
life was like for people working within this culture. The Ti-
wanaku heartland and hinterland provide an ideal opportu-
nity to compare activity between individuals from these two
areas. In addition, it also provides the opportunity to exam-
ine the formation of smaller laboring “communities” within
its variety of multiethnic neighborhoods
2
(Becker 2013;
Berryman 2011; Blom and Janusek 2004; Couture 2003;
Couture et al. 2008; Couture and Sampeck 2003; Goldstein
1993a, 2000b, 2005; Janusek 1999, 2003, 2005; Janusek and
Blom 2006; Valli
`
eres 2010, 2012). My goals involve show-
ing how the bioarchaeological evidence of labor can define
different working communities at various levels, and to dis-
cuss how each fits within this emerging complex society.
Tiwanaku’s Background and Cultural Context
Archaeological excavations have shown that the main
heartland or core of the Tiwanaku state emerged around
C.E. 500 in the high, flat plains of the Lake Titicaca Basin,
Bolivia with the main city of Tiwanaku emerging as an
important population center with growing cultural and
political influence among the Titicaca Basin’s residents
(Kolata 1986, 1993a). Within the city, distinct neighbor-
hoods (i.e., barrios) developed around the municipality’s
center, archaeologically noted as home to various peoples,
such as elites, stone tool manufacturers, potters, weavers, or
herders (Couture et al. 2008; Couture and Sampeck 2003;
Geisso 2011; Janusek 1999, 2005, 2008; Rivera 1994;
Valli
`
eres 2012). Initially, these barrios were thought to
be focused on supporting elite settlements, with influence
declining the further away one was from the “center” of
elite power. This idea was described as a “concentric cline
of the sacred that diminished in intensity from the city
core to its far peripheries ....Inhabitants of the Tiwanaku
occupied physical space in accordance with their relative
social and ritual status” (Kolata 1993a:93-94; 2003). Kolata
(1997:253) also suggested that the Tiwanaku city’s whole
purpose was for servicing elites and their aristocratic
lineages, and that Tiwanaku urbanites and craftspeople
serving the aristocracy enjoyed high status living.
More recent excavations and analyses of the Tiwanaku
culture instead suggest that independent households or
larger artisan collectives performed craft production au-
tonomously or semiautonomously, especially in their social
and exchange relationships (Bermann 1994; Goldstein 2005;
Janusek 1999, 2004, 2008; Rivera 1994, 2003). Bermann
(1994) and Janusek (1999) note that regular household ac-
tivities and their associated artifacts (e.g., food processing
lithics, ordinary textiles, and utilitarian hoes for agriculture)
occurred in areas of focused craft production, suggesting
household living more than specialist elite-production en-
claves. Goldstein (2005:77) described this style of labor or-
ganization, combining urban and craft living, as “embedded
in Tiwanaku’s diverse and segmentary social substructure
and not dictated by the demands of patrician sponsors.”
Janusek (1999) attributes these “embedded” craftspeople,
who were not attached to elites b ut also not strictly inde-
pendent, as a way the Tiwanaku state dealt with political
integration without forcing assimilation or loss of corpo-
rate identity. As such, the closest parallel to these embed-
ded neighborhoods may be the Western notion of a labor
guild where work was small-scale, and social capital built
through craft production seen as for the good of the larger
society (Epstein 1998; Jovinelly and Netelkos 2007; Kieser
1989; Ogilvie 2004; Vardi 1988). In addition, the members
of each Western guild community identified with her or his
work (e.g., masons, goldsmiths, woodworkers, weavers, pot-
ters), even adopting the trade as a surname for identification.
Thus, it is likely these Tiwanaku neighborhoods would have
been responsible to the larger community for the production
of various goods in a reciprocal environment that was not
elite-driven. Instead, crafts would have been for the gen-
eral public, while the crafting process also reified each local
community’s barrio identity (Janusek 1999:125).
In addition to crafting, during Tiwanaku times and in
association with the urban environment, pastoral and agri-
cultural production increased, likely to support the grow-
ing population
3
(approximately 20,000–40,000 people). The
city of Tiwanaku established control over local agricultural
40 Sara K. Becker
Figure 4.1. Map of heartland core Tiwanaku area in the Lake Titicaca Basin of Bolivia around
the present-day town of Tiwanaku and the hinterland colony region near the present-day town of
Moquegua, Peru.
production centers (i.e., raised-field agricultural beds) in
the nearby Katari Valley with increasing local control of
trade routes and an emphasis on an agro-pastoral lifeway
(Janusek 2008:20). Bandy (2001:204) interpreted many of
these changes as a successful strategy involving a system
of labor management. This system increased political and
ideological control with greater levels of ceremonialism and
large-scale feasting, so that by C.E. 500, “Tiwanaku was a
city [that] had become capable of dominating the entire Titi-
caca Basin politically, economically, and militarily” (Bandy
2001:204).
After the advent of the state in C.E. 500, Tiwanaku-
style material culture was also found increasingly farther
away from the heartland area in the warmer, lower-elevation
hinterlands. Prior to this expansion, there is very little ev-
idence for control over lowland areas, just trade exchanges
(Goldstein 1989, 2000a, 2005; Goldstein and Owen 2001).
Archaeologists (Albarrac
´
ın-Jord
´
an 1999; Goldstein 1989,
2005; Janusek 2004, 2008; Kolata 1993a, 1993b) generally
agree that this expansion to lower elevation areas was a
political one. The Tiwanaku peoples had a wish for luxury
items, such as maize or coca, which can only be abundantly
grown at lower elevations and in warmer climes. Goldstein
(1989:251) noted that sometime within C.E. 500–650,
Tiwanaku peoples arrived in lowland valleys, such as
the Moquegua Valley of Peru, “suddenly and in force,”
bringing Tiwanaku-style material culture with them
4
.In
this region, colonization was primarily focused on riverine
agro-pastoral production in three different areas (i.e.,
Omo, Chen Chen, and Rio Muerto). The control of these
important agricultural lands would have secured the maize
supply beyond levels that could have been traded for in this
pre-market economy, especially as chicha (fermented corn
beer) was important for ritual feasting to both heartland and
hinterland peoples
5
(Berryman 2011; Goldstein 2005).
During C.E. 800–1100, increased construction around
the city of Tiwanaku occur red alongside mass produced
Tiwanaku-style ceramics (Janusek and Kolata 2004) and
intensified agricultural production in the Katari Valley
(Bermann 1994; Janusek 2004, 2008; Janusek and
Kolata 2004). Janusek (2008:192-193) noted that “raised-
field farming became the signature productive regime of the
Lake Titicaca Basin.” Other agro-pastoral activities (e.g.,
herding, fishing, and rain-fed farming) would have been
lower status tasks as the main push was on raised-field
crops. These agricultural goods funded the cyclical feasting
that helped Tiwanaku’s residents negotiate power relations
(Janusek 2008:193).
The change in agriculture intensification may have
had a direct impact on lower elevation colonies. After
Community Labor and Laboring Communities 41
C.E. 900, a destruction and rejection of Tiwanaku-style
material culture in the Moquegua Valley coincided with
the Tiwanaku state losing control of this region (Goldstein
1993b:42). In addition, the focus on agricultural intensifi-
cation in t he highlands had eventual negative consequences
in the Titicaca Basin. The region underwent a long-term
drought that started around C.E. 1000 and could have been
a factor in the collapse of Tiwanaku about 100 years later
(Binford et al. 1997; Erickson 1999, 2006; Kolata et al.
2000; Moseley 1997; Ortloff and Kolata 1992). Any major
construction projects were discontinued by C.E. 1000, and
around this time, monuments associated with elites and elite
ancestors were ritualistically defaced and buried. After C.E.
1100, populations shifted from large, urban centers to small,
hilltop fortress settlements (pukaras) (Albarrac
´
ın-Jord
´
an
1992; Arkush 2011, 2012; Stanish 2003; Zovar 2012).
Materials and Methods Used to Study Labor
and Activity
To examine activity diff erences within different areas
of Tiwanaku society, I compared 1,235 adults from the two
areas: the heartland in Bolivia, which had 452 individuals,
and the hinterland colony in Peru, which had 783 people. I
evaluated all individuals for two skeletal measures of phys-
ical activity: musculoskeletal stress markers (sometimes re-
ferred to as entheses) and osteoarthritis. Because bones and
muscles work in conjunction with each other while tasks are
performed, my primary interest was in patterns and levels of
activity in order to understand the social structure of labor
in the various laboring communities of the Tiwanaku state.
In order to do this, I looked at labor in the Tiwanaku state
from four different spatial perspectives: (1) heartland ver-
sus hinterland colony; (2) heartland Tiwanaku Valley versus
the Katari Valley; (3) between each of the three hinterland
colonial settlements; and (4) within each highland valley
(i.e., within the Katari Valley, and within the Tiwanaku
Valley). The fourth objective was especially important in
this research per the previously reported multiethnic com-
munities of laborers and possible elite peoples, which could
provide bioarchaeological evidence of neighborhood-based
work groups.
In order to estimate labor, I first evaluated the evidence
of musculoskeletal stress markers within Tiwanaku skeletal
populations. Prior medical and bioarchaeological research
(e.g., Bridges 1989; Churchill and Morris 1998; Yu et al.
2011) has shown that certain tasks, like farming, show an
increase in muscle mass over an individual’s lifetime. Since
muscles work like bony levers for the underlying skeleton,
and where the muscles attach to bone as a person increases
muscle mass, so too can the connection points on bone grow
and strengthen. The attachment points, or musculoskeletal
stress markers, can help identify directional movement in
kinds of activities people did as well as levels of physical
labor such as workload. Overall, I looked at 37 muscle at-
tachment points and sorted them into five groups according
to location on the body: upper arm (i.e., shoulder move-
ment), lower arm (i.e., forearm movement), mid-body (i.e.,
hip movement), lower body (i.e., knee movement), and feet
(i.e., ankle and foot movement). For each point, a score of
present or absent was assigned.
Osteoarthritis (OA) was the second activity indicator I
used. Osteoarthritis shows injuries helpful in determining
repetitive movement as it can measure the same motion
used over and over again, such as grinding grain or weaving
textiles. I looked at 24 joint surfaces within seven joints:
shoulder, elbow, wrist, sacroiliac, hip, knee, and ankle. For
each individual, the multiple surfaces within each of the
seven joints were noted as present or absent for the evidence
of osteoarthritis.
Data were analyzed using generalized estimating equa-
tions (GEE), a population-averaged method accounting for
correlation among measures within subjects (Agresti 2007;
Ghislatta and Spini 2004). GEE works well for this type of
data because it models estimates of population parameters
that are calculated using individually recorded data points,
allowing for the largest possible sample size. However, each
of these data points remains linked to the individual, thus
preserving individual level information (Ghislatta and Spini
2004). The GEE procedure retains the categorical dependent
variable while keeping the data points linked (for example,
for each of the different joint surfaces), and does not bias the
data even though there are multiple data points within each
joint. It also accommodates variables that are not normally
distributed, small sample sizes, and randomly missing or
unobservable variables, which is especially useful in bioar-
chaeological studies, and social science research in general
(Becker 2012, 2013; Gagnon and Wiesen 2013; Nikita 2014,
2015). GEE can also evaluate any number of nominal or
quantitative predictor variables that cannot be assessed us-
ing bivariate analysis, such as controlling for age-at-death
and sex, as has been previously performed for these datasets
(Becker 2013). All data were evaluated for significance at
.05 level using the chi-square statistic.
Results and Discussion of Laboring
Communities
Comparisons between the Heartland and Hinterland
When looking at the muscle marker scores between
the heartland and hinterland, four out of five areas are
