Yasur-Lau Assaf. Household Archaeology in Ancient Israel and Beyond
Подождите немного. Документ загружается.
42 nimrod marom and sharon zuckerman
each consisting of several rooms or courtyards. A possible open area
was discerned between these structures. e whole area was residential
in nature. Characteristic of this phase are the sealed entryways and the
blocked openings discerned in the walls. is phase is attributed, on
the basis of its rich ceramic assemblage, to the last phase of the LBII
city of Hazor (i.e., stratum 1A in the stratigraphic scheme of Yadin)
(Zuckerman 2008).
Area S is thus an ideal candidate for the reconstruction of the devel-
opment of a domestic quarter within the context of the Canaanite city.
e project was designed as an interdisciplinary endeavor, aiming to
extract as much data as possible from the excavated structures and
associated open areas. During excavation, special emphasis was put on
the maximum retrieval of archaeological material. Floor deposits were
wet sieved using a otation machine, while other less secure depos-
its were sampled by partial dry sieving. e preliminary goal of the
excavation is the recovery of several categories of data that will be
crucial for the achievement of the research objectives. ese include
larger samples of well-stratied artifacts, such as ceramics, lithics, ani-
mal bones, and plant remains, all of which are deemed necessary for
determining the nature and scale of various daily activities, as well
as assessing their spatial distribution and temporal development. e
faunal assemblage is of prime importance in this context, and special
emphasis was therefore placed on understanding its formation pro-
cesses and taphonomic history.
Research Questions and Premises
Bone deposition can be divided into primary, secondary, and tertiary
episodes (Schier 1972, 1976, 1983; Meadow 1978). Primary deposi-
tion of bone reects refuse discarded at the place of animal processing
and consumption, and thereby provides the highest resolution infor-
mation on domestic activity areas. Animal bones in primary deposi-
tion are usually very small fragments, likely to have been trodden into
earthen oors or otherwise to have escaped periodic cleaning (Flad-
mark 1982; Schier 1987; LaMotta and Schier 1999). is fraction
of the bone assemblage cannot be recovered without high-resolution
recovery methods, such as wet sieving (James 1997). e density of
very small bone splinters and chips inside domestic spaces is expected
to vary, as places in which cooking and other bone processing activities
a case study from the lower city of hazor 43
took place should exhibit a higher density of such debitage (Hull 1987;
Metcalfe and Heath 1990; Bartram et al. 1991). e number of burnt
specimens can indicate the consumption of roasted vs. boiled esh
(Speth 2000), and the size and frequency of cancellous bone fragments
may likewise indicate bone fats (grease) rendering in particular areas
(Outram 2001).
However, since most identiable bone fragments are large (epiphy-
ses, teeth, large long-bone sha fragments), and may prove a sani-
tary disturbance and a hindrance to movement inside domestic space,
they cannot be assumed to be found in primary depositional contexts
(LaMotta and Schier 1999). Ethnoarchaeological and archaeological
studies conducted on refuse management show that larger bone frag-
ments are routinely removed from household oors to either the area
immediately adjacent to the house (the “to”) or to central dumps,
which are usually located near the primary consumption area (e.g.,
Hayden and Cannon 1983; Bartram et al. 1991; Martin and Russel
2000). Animal bones in secondary deposition are probably the best
target population to investigate domestic subsistence practices. Com-
munal dumps and bones recovered from to areas (alleys, accumu-
lations between house walls) present time-averaged “samples” of the
subsistence activities carried out in a domestic area, and are thus of
prime importance to the derivation of species, skeleton element abun-
dance, and demographic data. Being larger and less fragmented, bones
in secondary deposition are generally the specimens that yield more
zooarchaeological data (morphological, taxonomical, anatomical, and
mensural), which can be brought to bear on questions regarding socio-
economic variability between occupation periods (Hesse 1986; Zeder
1988; Bar-Oz et al. 2007).
Bones in tertiary deposition are accumulations brought as construc-
tion material (mudbricks or lls) to their archaeological context, or
otherwise removed from their original location. ese bone remains
are of little value to faunal analysis, as their original spatial and tem-
poral provenance is not known.
Providing we accept the premise that houses were routinely cleaned
when occupied, large bone fragments found above and on house oors
may reect primary discard from post-abandonment use of a build-
ing by human or animal visitors (squatters). Alternatively, house
oor assemblages may reect post-abandonment use of a house as a
dump—in which case they are assumed to be constituted of secondary
deposition from nearby houses (Cameron and Tomka 1993; LaMotta
44 nimrod marom and sharon zuckerman
and Schier 1999). A criterion for dierentiating use of abandoned
houses as a communal dump from a squatter community may be the
comparison of the above-oor and oor deposits from a house to
known dumps and middens in the vicinity, especially in terms of spe-
cies and body-part diversity (i.e., the number of taxa represented in
the assemblages). Deposition by squatters would probably show fewer
(if any) specimens from exotic or imported taxa (e.g., Nilotic sh), and
an overall lower diversity than a time-averaged accumulation of food
refuse from part of a large urban center. Use of an abandoned house as
a dump in an active urban area would show a pattern similar to nearby
middens in terms of animal diversity and taphonomic signature. e
microfaunal assemblage may also hint at which of the two options is
likely to be true: many noncommensal microvertebrate remains may
indicate deposition to have been the result of post-abandonment use
of a house by squatters; presence of commensals (e.g., house mice,
Mus musculus domesticus) may indicate continued human presence
at the house’s vicinity and thus the use of an empty house as a dump
(see, also, Hesse 1978, 1979; Hesse and Wapnish 1985). Microfaunal
remains can also yield important environmental information. Since
these animals have limited ranges and specic habitat preferences,
they are ne indicators of the immediate environment of the archaeo-
logical context in which they were found (e.g., Piper and O’Connor
2001; Weissbrod et al. 2005).
A primary research question is which taxa and of what age and sex
groups were eaten by the residents of the domestic quarters; which
body parts were consumed most oen; and how these foodways varied
spatially and diachronically. Demography and anatomical representa-
tion of the various livestock taxa are considered means to reconstruct-
ing ancient economy (Payne 1973; Redding 1981; Zeder and Hesse
2000; Vigne and Helmer 2007). Diachronic or spatial uctuations in
the frequencies of gourmet (i.e., meat-bearing) vs. non-meat-bearing
portions of livestock taxa may indicate changes in auence (Crabtree
1990; e.g., Schulz and Gust 1983; Jzereef 1988; Schmitt and Lupo 2008;
Marom et al. in press). Demographic shis from a young adult male
dominated assemblage to one in which older animals and especially
more females are represented may denote changes toward a less con-
sumerist and more autarkic economy, indicating a regional socioeco-
nomic shi toward lesser integration. An increase in the proportions
of livestock animals over game and low variability in the demographic
parameters of the population and skeleton element representation
a case study from the lower city of hazor 45
indicate consumers in the distribution chain of animal products, and
thus higher socioeconomic status (Zeder 1991). Trade relations can
be partly reconstructed using the presence of nonlocal animals in the
faunal assemblage (Van Neer and Ervynck 2004; van Neer et al. 2004;
Raban-Gerstel et al. 2008). ese imported animal commodities can
also serve as a measure of wealth: imported goods are usually limited
to the auent (Ervynck et al. 2003; Veen 2003).
Another important research question is diachronic and spatial pat-
terning in butchery practices of dierent taxa. Intensive butchery,
which includes the utilization of bone marrow and even grease, indi-
cates heightened eorts to utilize bone fats (Outram 2001). Utiliza-
tion of a wider selection of skeleton elements, especially those poor
in esh (feet, heads), may indicate subsistence stress or low socioeco-
nomic status (Schulz and Gust 1983; Izjereef 1989). Such information
can be derived from bones in secondary deposition (for interstrata
comparisons) and, to a lesser degree, from bones in primary deposi-
tion (for interhousehold comparisons; see below). Butchery patterns as
evident by cut marks on bone surfaces can illuminate dierential treat-
ment of carcasses from dierent animals (Lyman 1994; Lyman 2005),
as well as the intensity of consumption of various animal portions
(Domínguez-Rodrigo 1999). Inter- and intrastrata changes in butch-
ery intensity may be good indicators of auence. As primary animal
products (meat) become scarce, butchery would intensify to extract
animal fat from bone marrow (Bar-Oz and Munro 2007) and even
grease (Outram 2001, 2003)—practices which require some eort.
Also, a greater diversity of animal portions from more locally avail-
able taxa, including wild game and fowl, would probably be present at
the assemblage (Zeder 1988).
e sex and age distribution of the animals represented at the site
can be used to test specic hypotheses (Marom and Bar-Oz 2009)
regarding ancient herd management strategies. Such strategies include
optimization of culling to increase work, caloric, or wool yields from
domestic ocks providing animal products to the site (e.g., Payne
1973; Redding 1981).
e picture of past foodways gained by zooarchaeologists is greatly
biased by deposition and post-depositional loss of information
brought about by various taphonomic agents (Lyman 1994; Reitz and
Wing 1999). ese include, for example, the deleterious eects of bone
fragmentation by humans (Brooks et al. 1977; Marshall and Pilgram
1991; Enloe 1993; Pickering and Egeland 2006; Bar-Oz and Munro
46 nimrod marom and sharon zuckerman
2007), carnivore gnawing (Binford 1981; Lyman 1994; Blumenchine
et al. 1996), subaerial weathering (Behrensmeyer 1978), and trampling
(Olsen and Shipman 1988). Another goal of faunal analysis is reveal-
ing to what degree various taphonomic actors aected the assemblage,
and thus how reliable cultural inferences based on faunal data is (Bar-
Oz and Munro 2004). Of equal importance is the determination of the
bias caused by bone recovery and analysis procedures (James 1997;
Marean et al. 2004). is problem is exacerbated in large-scale excava-
tions, where the recovery of all faunal material is not practical.
On-Site Faunal Recovery Procedures: A Case Study from Hazor
e recovery of faunal remains from large urban sites is problematic,
since practical limitations prohibit the complete recovery of all bone
material. erefore, at Hazor a two-tier recovery system is used to col-
lect bone specimens. e rst tier consists of recovery by hand by the
excavators; and the second of wet sieving (Fig. 3).
Recovery of faunal remains by the excavators without sieving was
shown to be a serious source of bias (Meadow 1980; Turner 1989;
James 1997; Zohar and Belmaker 2005). To reduce this bias, the hand-
picked sample is used to derive zooarchaeological information per-
taining only to larger mammals: sheep, goats, pigs, cows, and larger
game animals, the bones of which are more massive and likelier to be
picked up by excavators.
Wet sieving was carried out using a 1 mm aperture mesh. e buck-
ets for wet sieving contained sediments from which no bone mate-
rial was extracted by the excavators. Following sieving and drying, the
sieved fraction was sorted to isolate micromammal, sh, and mam-
mal bones. e samples for wet sieving were taken daily, in vertical
intervals of approximately 10 cm. When digging through extensive
lls, mudbrick collapse, or topsoil, ve buckets were sampled from
each square. e horizontal pattern of sampling alternated daily
between an X-shaped pattern, where one bucket is lled from each
corner of the square and one from its center, and a cross-shaped pat-
tern where four buckets were lled from the mid-side of the excava-
tion square and an additional bucket was taken from its center. is
superimposition of an X and a cross formed a “Union Jack” pattern
of sampling in each square, thereby making an allowance for spa-
tial variability in nds’ density and composition, while keeping the
a case study from the lower city of hazor 47
Figure 3. A owchart presenting a scheme of the faunal analysis procedures employed at the Lower City of Tel Hazor. See text
for explanations.
Fossil assemblage sampling using...
Recovery by hand
Context: Secondary
Analytical methods: Diagnostic zones, bone surface modications
Derived information: Herd demography, body-part representation, taxonomic frequencies, butchery, taphonomy
Researchquestions: Social status and identity, gross diachronic and spatial trends in taxonomic, demographic and
anatomical parameters
Recovery by wet sieving
Recovery level: Lo
w, mainly medium and large mammal bones
Recovery level: High, representative sample of...
Bird remains
Fish remains
Research question: Trade and dietary spectrum
Research question: Fowling, dietary spectrum
Derived information: taxonomic frequencies, body-part representation, taphonomic observations
Context: Primary and secondary
Context: Primary and secondary
Micro-vertebrate remains
Research question: Depositional context of th
e associated faunal assemblage, micro-environment
Context: mainly primary deposition
Macro-mammal remains
Research questions: High resolution contextual comparisons; activity areas; control over recovery and anatical biases in
the hand-recovered sample
Context: Small fragments on oor levels may be in primary deposition
Derived informaton: Size-class (caprine/cattle size) and rough body part (limb, he
ad, axis) relative frequency
determinations per contextual unit; Density of bone nds of each type per volume unit.
Analytical methods: Identication to rough taxonomic and body-part categories of ALL bone fragments from wet-sieving;
nds’ density estimations; some taphonomic observations
48 nimrod marom and sharon zuckerman
wet sieving and consequent sorting of the sieved material logistically
manageable (Fig. 4). When the excavation reached above oor lev-
els, the sampling method changed to include a varying number of
buckets from each locus per 10 cm vertical increment. e samples
were taken to provide a balanced horizontal representation of all
parts of each of the sampled loci. Mean heights were taken for each
sampling unit, and the volume of sediment (in liters, before sieving)
was recorded.
Zooarchaeological Analytical Procedures
Processing the hand-picked sample includes the identication of cer-
tain skeleton elements, known as “diagnostic zones,” following the
protocol detailed by Davis (1992). ese zones are easily ascribed to
taxon and element, and provide the basic set of aging data through
teeth eruption and wear (Grant 1982), and state of epiphyseal fusion
(Silver 1969). Measurements of these elements are later used for
determining the sex ratios in the archaeological sample (e.g., Wilson
et al. 1982; Greeneld 2006). e cost-eciency of using diagnostic
zones is high, and it enables the derivation of maximal zooarchaeolog-
ical information in a short time. Diagnostic zones are additive values
that represent real counts, which makes them amenable to statistical
Figure 4. e sampling scheme for wet sieving above oor levels. X marks
a bucketful of sediments from which bone material was not removed by the
excavators in a 5 m × 5 m excavation square. e sampling scheme alter-
nates between the St. Andrew’s cross (le) and the St. George cross (right) in
10–15 cm vertical increments.
a case study from the lower city of hazor 49
analyses of frequencies, and free from problems of aggregation (unlike
minimum number of individuals/elements counts) (Grayson 1984;
Lyman 2008). In addition, diagnostic zones are small and far apart on
the skeleton, which allows the assumption of independence of counts,
necessary to analyses of frequencies, slightly easier to make.
Although cost-ecient, the diagnostic zones methods have been
criticized in the last decade for being liable to biases stemming from
density-mediated attritional processes (Marean and Frey 1997; Picker-
ing et al. 2003; Marean et al. 2004). However, empirical studies con-
ducted in later prehistoric sites (Bar-Oz and Dayan 2002; Marom and
Bar-Oz 2007) have shown the dierence between the various analytical
methods to be small. Recent studies at Tel Rehov (Marom et al. 2009)
showed a high frequencies of high-value, low-density bones, in spite
of the density-mediated attrition that was demonstrated to have taken
place at the site. A similar study of skeleton element abundance at Tel
Dor (Raban-Gerstel et al. 2008) showed similar underrepresentation
of heads and feet, which is contrary to the prediction of the “sha-
critique” for a site where consistent recovery and analysis of long-bone
sha fragments did not take place.
Processing the wet-sieved sample units includes picking through the
sieved sediments to isolate micromammal and sh bones, which are
passed on to respective experts aer their broad taxonomic designation
was recorded. All larger mammal bones are separated and counted to
calculate nds’ density per sample. Mammal bones are then sorted into
a “large mammal” (cow/horse-sized) and “medium mammal” (sheep/
goat-sized) categories, and further into long bone, axial, and cranial
fragments. e number of burnt bones is also recorded (Table 1).
e number of head, limb, and axis skeleton bones in the sieved
sample can be contrasted with the skeleton element breakdown
derived from the analysis of the identied remains per mammal size
category (medium/large) to estimate the eects of analytical and recov-
ery biases. e smaller osseous nds in the various sample units and
their division to broad taxonomic and anatomic categories can be used
to estimate what garbage had been disposed of in primary deposition
in each domestic unit. e density of nds may show activity areas
where carcass processing took place within structures, or, alternatively,
which places escaped routine cleaning (corners, under or behind
furniture, etc.).
All bones over 2 cm in length, recovered by hand or by wet sieving,
were scanned under magnication (×3) using an oblique light source
50 nimrod marom and sharon zuckerman
to detect bone surface modications (Blumenchine et al. 1996). ese
include carnivore gnawing, burning, weathering, percussion marks,
cut marks, and root etching. Fracture morphology was recorded for
diagnostic bones that retained diaphyseal fragments (Villa and Mahieu
1991), to provide evidence for fresh bone fracturing (presumably for
marrow extraction) versus fracturing of dry bones due to post-depo-
sitional processes (trampling, sediment weight, weathering). Tapho-
nomic information was listed in the database on a per-locus basis,
enabling interlocus taphonomic comparisons in the frequency of vari-
ous taphonomic indicators.
Data Processing
Comparison of skeleton elements and taxonomic abundances across
strata and architectural units will be performed using standard multi-
variate statistical techniques: correspondence and cluster analyses. Age
determination will rely on both epiphyseal fusion and tooth wear data;
mixture analysis and morphological attributes of horn cores and pubis
bones will be used to determine sex ratios (Greeneld 2006).
Discussion
e basic assumption underlying the excavation of Area S in the Lower
City of Hazor is that cultural processes of rise and decline, centraliza-
tion and decentralization, economic growth and impoverishment of a
city and its inhabitants can all be detected through a careful and prob-
lem-oriented excavation of the households and domestic units. e
Table 1. An example of the format used to describe the unidentied fraction of wet sieved and
sorted samples
Sample # Date Area Square Locus Basket Height Volume
(litres)
Sample spots Type
6 3/9/2008 S T20 L08-2001 2000/09 225.9 24 diagonal, NE-SW wet-sieving, 1 mm
7 5/9/2008 S A20 L08-2008 2000/47 225.53 18 diagonal, NE-SW wet-sieving, 1 mm
8 5/9/2008 S S20 L08-2006 2000/50 226.03 22 diagonal, NW-SE wet-sieving, 1 mm
9 5/9/2008 S T1 L08-2000 2000/51 225.87 20 diagonal, NW-SE wet-sieving, 1 mm
10 5/9/2008 S U1 L08-2004 2000/52 225.64 7 diagonal, NW-SE wet-sieving, 1 mm
11 5/9/2008 S T20 L08-2010 2000/49 225.78 21 diagonal, NW-SE wet-sieving, 1 mm
12 5/9/2008 S U20 L08-2002 2000/48 225.69 23 diagonal, NW-SE wet-sieving, 1 mm
13 8/9/2008 S A20 L08-2008 2000/70 225.46 34 cross wet-sieving, 1 mm
14 8/9/2008 S U20 L08-2012 2000/71 225.61 32 cross wet-sieving, 1 mm
a case study from the lower city of hazor 51
nature and scale of the activities performed by the household (food
and cra production, distribution, consumption, social reproduction)
are expected to change according to larger processes inuencing its
historical and social wider contexts (Wilk and Rathje 1982; Watten-
maker 1998; Robin 2003). A meticulous study of all aspects of material
culture can illuminate issues of mundane life and daily activities of the
urban commoners of Hazor. A better understanding of these facets of
the city’s life will enable us to tackle questions concerning the larger
processes of development and decline aecting the site.
Here we would like to highlight one relevant example of such a
process—the gradual economic decline, leading to a situation of politi-
cal and social strife, witnessed in Canaan during the Late Bronze Age.
is process is well attested both by written documents and in the
archaeological record of the period (Liebowitz 1987; Bienkowski 1989;
Knapp 1989; Bunimovitz 1995). Material indices of these trajectories
might include architectural development and internal alterations in
contemporary structures, changes in pottery vessel form and technology,
variation and stability in the availability and consumption of dierent
foodstus. All these aspects of the archaeological record are assumed
to be sensitive to situations of economic and social stress. At Hazor,
features of “crisis architecture,” “termination rituals,” and deteriora-
tion in other aspects of the LB assemblage were detected in the con-
text of monumental temples and public structures and are interpreted
as the archaeological correlates of this process (Zuckerman 2007a).
According to this reconstruction, these internal conicts and eco-
nomic stresses led, eventually, to the conagration of the monumental
structures in the Lower City and on the acropolis, in a violent destruc-
tion campaign by the city’s inhabitants aimed at the symbols of elite
power.
Table 1 (cont.)
Sediment Context NSP Large mammal Medium mammal Bird Fish Rodent Burnt
gray clay under top soil 12 0 0 0 6 0 0 0 0 0 2
gray clay ll 29 0 0 0 7 0 1 0 3 1 5
gray clay ll 35 0 0 0 6 2 0 0 1 0 2
gray clay ll 38 0 0 0 19 1 0 0 0 0 5
gray clay ll 15 0 0 0 8 1 0 0 1 1 4
gray clay ll 47 0 0 0 15 2 0 0 1 0 10
gray clay ll 25 0 0 0 5 0 0 0 1 0 5
gray clay ll 28 0 0 0 14 1 0 0 2 0 4
gray clay ll 41 0 0 0 28 2 0 0 1 1 2