JOURNAL OF ANTHROPOLOOICAL ARCHAEOLOGY 6, 297-335 (1987)

Household Possessions and Wealth in Agrarian States:

Implications for Archaeology

MICHAEL E. SMITH
Department of Sociology and Anthropology, Loyola University of Chicago,
6525 N. Sheridan Road, Chicago, Illinois 60626

Received October 28, 1985

This paper deals with the problem of measuring household wealth from do-
mestic artifactual remains. After an archaeologically useful definition and classiti-
cation of wealth is presented, quantitative anthropological, sociological, and his-
torical analyses demonstrating a strong association between household posses-
sions and wealth are reviewed. Then, ethnographic and historical data on
different functional categories of household goods are examined in order to assess
their relative ability to predict wealth. Once the nature of the relationship be-
tween household possessions and wealth in the systemic context is specified,
methodological problems influencing the detection of this relationship in the ar-
chaeological context are treated. It is concluded that household possessions are
quite good wealth indicators in the systemic context, and that proper attention to
a number of methodological issues allows archaeologists to take advantage of this
relationship to measure household wealth in past societies. Q 1987 Academic
Press, Inc.

INTRODUCTION

One of the most important recent developments in the archaeological

study of agrarian societies is the increasing .attention devoted to the
household as a unit of social analysis. The emergence of “household ar-
chaeology” as an explicit substantive and methodological focus within
the discipline (e.g., Wilk and Rathje 1982; Tourtellot 1983) is closely re-
lated to an increase in attention to households in related fields like socio-
cultural anthropology and social history (e.g., Yanagisako 1979; Netting
et al. 1984). Analysis at the level of the household has a number of ad-
vantages in studies of social and economic organization. First, the house-
hold is the basic unit of production and consumption in agrarian soci-
eties, and thus many of the activities involved in adaptation to the natural
and social environment take place at the household level. Second, a focus
on households usually permits a consideration of the entire range of so-
cial variation within a given society, since households can be selected
from various economic classes, residential areas, occupational groups,
297
0278-4165187 $3.00
Copyright 0 1987 by Academic Press, Inc.
All rights of reproduction in any form reserved.
298 MICHAEL E. SMITH

and so on. Third, many large-scale social, political, and economic institu-
tions and processes have impacts at the household level, and the study of
households allows the researcher to address such larger patterns through
their expression in a controlled social setting.
In the analysis of households in agrarian societies, one of the most
significant variables for many kinds of studies is household wealth.
Wealth can vary significantly across regions, across politicoeconomic
systems, and between sectors of particular large politicoeconomic
systems. More important for many anthropological and archaeological
problems, however, is the variation in household wealth within communi-
ties and within societies. At the household level, wealth is closely related
to a number of significant social and demographic variables, including
family size and structure, occupations of household members, and do-
mestic developmental cycle (e.g., Yanagisako 1979; Netting et al. 1984).
Investigations of household wealth can therefore lead to more adequate
models of local and regional social organization and of the local impact of
higher-level institutions like state governments, market systems, and the
like. Furthermore, the analysis of household wealth can contribute di-
rectly to the study of the social and behavioral implications of change
through time in both low-level (e.g., demographic) and high-level (e.g.,
political or economic) processes.
As archaeologists turn increasingly to the analysis of households, the
lack of a systematic approach to the material correlates of wealth is be-
coming obvious. Some progress has been made in using residential archi-
tecture to monitor household wealth, and studies of status patterning in
mortuary contexts are clearly relevant (this work is reviewed below).
However, one of the most promising lines of research has had almost no
systematic treatment-the use of material possessions (household arti-
facts) to study wealth. This paper contributes to the task of building a
methodology for the analysis of household wealth using portable arti-
facts. First, an archaeologically useful definition and classification of
wealth are provided. Then a number of case studies are presented to es-
tablish firmly that household possessions do indeed measure wealth in
agrarian states. The bulk of the paper is next devoted to a consideration
of different functional classes of household items in terms of their ability
to predict wealth in ethnographic, historical, and archaeological settings.
Finally, some of the methodological problems associated with the anal-
ysis of wealth from archaeological artifacts are discussed. The overall
conclusion of this paper is that household possessions clearly are good
indicators of household wealth in ethnographic and historical settings,
and that this relationship can be used archaeologically when the appro-
priate methods are applied.
 HOUSEHOLD POSSESSIONS AND WEALTH 299

Wealth: Definition and Classification

In a discussion of comparative social stratification, Haller (1970) points

out that wealth is a difficult concept, particularly when comparisons are
made among different economic systems. He defines wealth as “access
to goods and services” (1970:475f), and notes that no single variable (like
income, land holdings or possessions) is sufticient to adequately measure
household or individual wealth. While this definition is intuitively ap-
pealing and has the advantage of linking the concept of wealth to anthro-
pological definitions of social stratification,‘,’ a more concrete approach
is needed for archaeology. The standard definition of wealth used in the
field of economics (e.g., “anything that has market value,” Jones
(1980: 15)) can be extended to non-western economies by changing
“market value” to “value.” For example, Schneider defines wealth as
“the total of desirable (i.e., valuable) goods, both social and material,
possessed by someone or existing in a community” (Schneider 1974:256).
(Difficulties with the concept of value, whose operationalization is cru-
cial for archaeological studies of wealth, are discussed in the method-
ology section below.)
Jones (1980) presents a useful classification of types of wealth based
upon standard usage in economics, modified for historical data from an
agrarian state, late colonial America. Her scheme is portrayed graphi-
cally in Fig. 1 (see Jones 1980:14-27). Financial wealth refers to the bal-
ance between financial claims and liabilities, while physical wealth in-
cludes land, goods, and human capital. Because of the impossibility of
recovering data on financial wealth archaeologically, this category will
not be considered further here. Jones’ extensive tabulations of wealth
levels demonstrates that in colonial America, physical wealth is entirely
adequate for the analysis of total wealth patterns. Nonportable physical
wealth comprises land and buildings; this is the largest single component
of colonial wealth (55%; see Fig. 1). Portable physical wealth is divided
into human (slaves, servants, and other types of dependent labor) and
nonhuman forms. (Values for the former category are quite variable
within the thirteen colonies. As Jones’ tallies show, slaves and servants
range from 33.6% of total physical wealth in the south to 0.4% in the
northeast. The figures given in Fig. 1 and Table 1 represent mean values
for all thirteen colonies.) Nonhuman portable physical wealth is the most
relevant category for the present study, so a further breakdown of this
type is provided in Table 1. The distinction between producers’ and con-
sumers’ goods is generally useful, although it becomes problematic when

* See Notes section at end of paper for all footnotes.

300 MICHAEL E . SMITH

Total Wealth

Financial

Assets Liabilities Portable (~4) Non-portable (139)

Non-humonk)- Humon (49) Lond Gings

/ n
(see Table 1) Slaves Servants

FIG. 1. Classification of wealth in Colonial America, 1774 (from Jones 1980). Figures in
parentheses represent the mean value per household (in pounds sterling) in the 13 colonies
in 1774). Nonhuman wealth is further broken down in Table 1.

production for use takes place within the context of the household. The
separation of durable and perishables, an intuitive and theoretically im-
portant distinction, involves an arbitrary cutoff point; Jones (1980:92fl
separates these categories on the basis of a normal useful life of 1 year.
In studying household wealth archaeologically, we are interested in
finding methods to measure total household wealth using some subset of
the types listed in Fig. 1 and Table 1. While control of land typically
comprises the greatest proportion of total household wealth in agrarian
societies, there is no direct archaeological measure of landholding. Re-
mains of buildings, on the other hand, are typically recovered by archae-
ologists, and estimates of the value of residential structures is an impor-
tant method of studying prehistoric household wealth (see discussion
TABLE 1
TYPES OF NONHUMAN PORTABLE PHYSICAL WEALTH IN COLONIAL AMERICA, 1774”

Type Value6 Categories

Producers durable goods 31.7 A. Livestock
B. Farm equipment and implements
C. Nonagricultural producers’ capital
D. Domestic production equipment (textile, etc.)
Producers’ perishable goods 12.5 A. Crops
B. Producers’ materials (lumber, etc.)
Consumers’ durable goods 18.4 A. Clothing
B Furniture
C. Other household equipment and implements
Consumers’ perishable goods 1.5 A. Foodstuffs
B. Other supplies (firewood, candles, etc.)
0 Classification and data are from Jones (1980).
b Value figures represent the mean value per household (in pounds sterling) in the thirteen
Colonies in 1774.
 HOUSEHOLD POSSESSIONS AND WEALTH 301

below). Human wealth, defined as the access to the labor of non-house-

hold members, can be a highly significant component of total wealth in
agrarian states (Jones 1980; Polanyi 1968:93). While the indirect benefits
of such labor service may leave physical evidence (as in the construction
of larger houses for the wealthy), the nature of social control involved is
generally not recoverable archaeologically. The types of nonhuman por-
table physical wealth recovered in the archaeological record depend pri-
marily upon factors of preservation (durables will be favored over perish-
ables) and location (the locations of production, consumption, and depo-
sition activities). Thus, consumers’ goods will typically be recovered in
household contexts (in and around the house and in contexts of house-
hold refuse), while producers’ goods may be present in either household
contexts or in association with separate buildings away from the resi-
dence .

Archaeological Approaches to Stratification and Wealth

Archaeologists have traditionally approached the analysis of prehis-

toric stratification and wealth differentials with three kinds of data: resi-
dential architecture, burials, and household artifacts (see Haviland 1981;
Rathje 1983).*
(I) Residential architecture. This is probably the strongest and most
consistent expression of wealth levels in agrarian states (Blanton 1984).
Such societies exhibit a great range of variation in the size and quality of
housing, and these factors relate directly to the level of a household’s
access to goods and services (see Yang 1945:37-41; Lewis 1951:178ff,
Mack 1951; Blake 1981; Wilk 1983; McGuire 1983; Abrams 1984).
(2) Burials. Burials represent the class of archaeological remains that
have been most extensively analyzed in terms of wealth and status differ-
entials. Mortuary studies can shed considerable light on patterns of social
stratification in two realms: contextual and osteological. Contextual vari-
ables like burial form, orientation, and associated grave goods clearly
relate to sociocultural variability in prehistoric populations (e.g.,
Chapman et al. 1981; Bartel 1982), while osteological variables reflect
patterns of social stratification and wealth variation to the extent that
nutritional and health factors are affected by wealth in agrarian states.
Ethnographic sources reveal a strong correlation between dietary quality
and wealth in state societies (e.g., Lewis 1951:187@, DeWalt 1983; see
discussion below), and these differences often produce widespread clin-
ical malnutrition in the lower classes (e.g., DeWalt 1983). For archaeolog-
ical examples of this approach, see Haviland (1967), Klepinger (1984),
and Huss-Ashmore et al. (1982).
(3) Household artifacts. In comparison with architecture and burials,
302 MICHAEL E. SMITH

household artifacts offer a number of advantages and disadvantages for

the archaeological study of wealth differentials. Their greatest advantage
consists of the fact that numerous studies have demonstrated that the size
and composition of household artifact inventories are strongly associated
with wealth levels in both industrial and agrarian states (see below). This
association stands in spite of the fact that the value of household posses-
sions usually comprises only a small proportion of total household wealth
(e.g., consumers’ durables comprise an average of 9.1% of nonhuman
physical wealth in Jones’ (1980: 103) sample of colonial households). Fur-
thermore, several observers have noted that artifact inventories change
more rapidly than architecture (Lewis 1973:259) or even furniture (Teller
1968) in response to changes in household situation. While this greater
sensitivity may be a drawback in studies of wealth in that inventories are
also highly responsive to change in nonwealth variables (e.g., household
size and composition), it can be an advantage in fine-scale studies of
household change where chronology is well controlled. Another advan-
tage of household artifacts for studying wealth levels that should not be
overlooked is their ubiquity and abundance in archaeological contexts.
Artifacts from virtually any set of residence-related contexts (house-
floors, kitchen middens, yard areas, discrete dumps, etc.) can be used to
investigate wealth levels if appropriate methods are applied.
Against these advantages to the use of household artifacts in the study
of social stratification are two serious disadvantages. First, while this
paper demonstrates the strong association between a number of compo-
nents of artifact inventories and wealth levels, inventories are also in-
fluenced by a series of other factors. For example, Schiffer et al (1981:83)
list the following as the major determinants of the kinds and quantities of
household goods cross-culturally: (1) stage in the domestic develop-
mental cycle; (2) degree of residential stability; (3) time elapsed since the
last change of residence; and (4) wealth or income. To these four factors
must be added at least two more: (5) household size and composition,
which are related to, though separate from, the first factor above; and (6)
the occupations of household members and the presence and nature of
specialized production/exchange activities in the residential area. The
second serious disadvantage of household artifacts concerns the nature
of archaeological formation processes. While household artifacts are
quite strong predictors of wealth while in use (systemic context), the
operation of a number of well-documented cultural formation processes
virtually guarantees that some of the wealth indicators recovered through
excavation (archaeological context) will be weaker predictors of wealth
levels (see Schiffer (1976) on systemic vs archaeological context). These
issues are discussed in the methodology section below.
 HOUSEHOLD POSSESSIONS AND WEALTH 303

DO HOUSEHOLD POSSESSIONS REFLECT WEALTH?

SOME EXAMPLES

Before turning to the analysis of how archaeologists can go about mea-

suring wealth from household goods, it should be firmly established that
household possessions in the systemic context do indeed monitor wealth
in agrarian states. Four examples clearly demonstrating this relationship
are presented in this section. These examples encompass a wide range of
cultural contexts and social science approaches, all showing that house-
hold possessions are useful predictors of wealth in state-level societies.
Additional empirical support of this finding is discussed in the remainder
of the paper.3
(1) The first type of evidence supporting the association between pos-
sessions and wealth consists of three studies in very different cultural
settings (Tahiti, Finland, and Mexico) that show how households conform
to Guttman scales when scored on material goods and furnishings. The
summary data are provided in Table 2. The fact that these household in-
ventories conform to Guttman scales (i.e., if a given household has one of
the listed items, then it also has all items listed below that item) merely
indicates that there is a consistent hierarchical ordering of material pos-
sessions within each community. However, the addition of two other con-
siderations permits the conclusions that these data reveal a strong associ-
ation between household wealth and material possessions. First, the
more restricted items (toward the top of each list) are generally more
costly than are the more prevalent items; and second, these costly and
restricted goods are highly valued in the three communities (Kay 1964;
Pelto 1973; DeWalt 1979). Therefore access to valuable material goods is
restricted and wealth variation is clearly and consistently expressed
when the distributions of particular goods are examined. Ethnographic
background data provided in the three studies indicate that the scores of
households on the material possessions scales are associated with inde-
pendent assessments of wealth. DeWalt (1979) is the most explicit about
this; the material possession scale is shown to have statistically signifi-
cant positive correlations with such variables as nonagricultural income
and value of agricultural equipment. In a related study, DeWalt (1983: 134)
shows positive correlations of this scale with dietary quality variables
including amount of meat and amount of total animal protein in the
household diet.
(2) Another ethnographic example confirming the relationship be-
tween household wealth and material possessions is Oscar Lewis’ study
(1969, 1973) of possessions in the Panederos vecindud of Mexico City in
the 1950s. Panaderos was a lower-class urban tenement inhabited by
304 MICHAEL E. SMITH

TABLE 2
GUTTMAN SCALE ANALYSES OF HOUSEHOLD POSSESSIONS
% of households
Place and citation Item with item
A. Papeete, Tahiti 1. Automobile 5.0
(Kay 1964) 2. Refrigerator 22.5
Number of households: 40 3. Kerosene or gas stove 27.5
Coefficient of 4. Motorcycle 35.0
reproducibility: .98 5. Radio 55.0
6. Bicycle 82.5
7. Primus stove 95.0
B. Sevettijiirvi, Finland 1. Television 5.8
(Pelt0 1973) 2. Refrigerator 10.1
Number of households: 69 3. Automobile 14.5
Coefficient of 4. Gas/electricity for cooking
reproducibility: .% and lighting 14.5
5. Oil heating 18.8
6. Gas/electricity for cooking
or lighting 31.9
7. Washing machine 34.8
8. Telephone 47.1
9. Snowmobile 82.6
10. Chain saw 84.0
C. Temazcalcingo, Mexico 1. Television 12.3
(DeWah 1979) 2. Stove 15.8
Number of households: 57 3. Wardrobe 24.6
Coefficient of 4. Sewing machine 24.6
reproducibility: .95 5. Raised hearth or stove 36.8
6. Bed 63.2
7. Radio 75.4
8. Iron 93.0

fourteen families living under conditions of extreme poverty. Lewis did a

complete inventory of all material possessions and assigned a monetary
value to each, A summary of this data, broken down by the material
categories employed by Lewis, is given in Table 3. It is clear that even
within a small segment of the urban lower class there is considerable
variation in the monetary value of household possessions; household
totals vary from $119.26 to $936.78 (expressed in dollars). Lewis
(1973:275) indicates that these values are associated with other wealth-
related factors like income (see also Lewis (1970)). The usefulness of in-
dividual categories of goods for assessing overall household wealth is
also indicated in Table 3; these data are discussed more fully below. For
the present, the major implication of the data in Table 3 is that household
 TABLE 3
HOUSEHOLD POSSESSIONS AND WEALTH IN A MEXICO CITY TENEMENTS

Total value, all 2 top and 2 bottom

14 households Mean extreme households Correlation*
nws of value per with household Use-
possession Dollars % household Wealthy Poor Wealthy + Poor totals life”
Household totals 4114d 100.0 294 721 120 6.0 - -
Furniture 1375 33.4 98 318 19 17.1 .883 4.5
Clothing 1127 27.4 81 114 26 4.3 538 0.8
Tools 327 7.9 23 103 3 34.0 .749 5.1
Household equipment 294 7.1 21 24 9 2.6 .323 2.1
g Bedding 279 6.8 20 41 12 3.5 .687 1.7
vI Kitchen utensils 230 5.6 16 21 9 2.4 .I62 2.5
Jewelry 126 3.1 9 30 <l 40.0 .611 0.8
Toys 122 3.0 9 44 0 - .384 no data
Religious items 108 2.6 8 13 35 0.4 - .288 5.1
Ornaments 92 2.2 7 13 5 2.4 .165 no data
Animals 16 0.4 1 <l 0 - -364 no data
Plants 10 0.3 <l <l <l 1.2 .077 no data
Medicine 8 0.2 <I <l I 0.5 - .055 no data
4 Data are taken from Lewis (1973).
* Correlations are Pearson’s r.
c Use-life represents the mean length of possession (in years) for items of each category.
d All figures are in U.S. dollars as presented by Lewis (1973).
306 MICHAEL E. SMITH

wealth differences are clearly expressed in the value of material posses-

sions found in the dwelling place.
(3) A third set of examples of the positive association between wealth
and household possessions consists of several studies done in the United
States in the 1930s through 1950s by sociologists. These investigators
were looking for a quick and easy method to measure “socioeconomic
status” in order to study the relationship between this factor and other
variables of interest. Direct measurement of status was difficult and time
consuming, involving a combination of objective scales (like income, oc-
cupation, or education) and subjective judgments. It was found that
scales based upon the material culture of the dwelling effectively mea-
sured socioeconomic status or “level of living” and could thus be em-
ployed as a substitute measure in their studies. This finding was made
independently by Chapin (1935) in an urban setting (see also Guttman
(1942)) and by Sewell(l940) in a rural setting (see also Belcher and Sharp
(1952)). Chapin was able to measure status solely on the basis of obser-
vations in urban living rooms (using variables like furniture type, floor
coverings, and reading material), while Sewell used a wider range of ob-
servations (from house construction to plumbing to furniture) in his char-
acterization of Oklahoma farm families. In these and later related studies
(e.g., Chapman 1955; Laumann and House 1970; Belcher 1972), house-
hold possessions form an important constituent of the status scales, and
in all cases the material scales are shown to correlate highly with income
and other wealth-related measures.
(4) A final example of the ability of household possessions to monitor
total wealth is provided by Jones’ (1980) massive study of colonial Amer-
ican wealth in 1774. The data presented in Table 4 here indicate that when
various social categories are arranged in order from high to low wealth
(total physical wealth), the corresponding mean values of consumers’
goods follow approximately the same order (the correlation between
these sets of figures, ignoring the subclasses under merchants and
farmers, is .885). Within the merchant and farmer categories, subclasses
defined by wealth level (see footnote b to Table 4) also exhibit this rela-
tionship. In summary, these data suggest that the value of a household’s
consumers’ goods is a good predictor of a household’s total physical
wealth. (It should be noted that the aggregate nature of these data is not
ideal for firmly establishing this conclusion. It would be useful to analyze
the relationship between consumers’ goods and total wealth using the
raw data as published in Jones (1978).)
There are several interesting characteristics of the data presented in
Table 4. First, a decline occurs in the percentage of total wealth repre-
sented by consumers’ goods from low to high wealth categories. As
 HOUSEHOLD POSSESSIONS AND WEALTH 307

TABLE 4
CONSUMERS’ GOODS AND PHYSICAL WEALTH FOR SELECTED SOCIAL CATEGORIES IN
COLONIAL AMERICA, 1774a

Value in pounds sterling

Total Consumers’
Social category physical wealth goods Percentage
Esquires, gentlemen, officers 572.4 40.6 6.5
All merchants 497.1 11.9
High-wealthb (797.7) (2;) (9.2)
Middle-wealth (235.9) (42.4) (18.9)
Professionals 341.0 35.2 9.0
All farmersC 262.3 17.5 6.5
High-wealth (676.3) (31.4) (4.6)
Middle-wealth (171.6) (15.9) (9.2)
Low-wealth (34.5) (8.2) (23.2)
Shopkeepers, innkeepers 204.3 26.2 12.3
Artisans 122.5 13.5 11.6
Mariners, laborers 62.0 14.0 21.9
o These data, representing mean household figures for all thirteen colonies, are from
Jones (1980: Tables 7.5-7.31, pp. 224-247).
b The merchant and farmer categories are subdivided into high-, middle-, and low-wealth
groups on the basis of total physical wealth of: over 400 pounds, 100 to 400 pounds, and
under 100 pounds, respectively.
c The farmers category includes only full-time farmers without significant nonagricultural
income.

Jones (1980:207) points out, this pattern is similar to that described by

Engel’s Law in which the percentage of income spent for food declines as
income rises (see Douglas and Isherwood (1979:97) on Engel’s Law). The
two categories which deviate from this pattern are merchants (whose
consumers’ goods are a higher percentage of wealth than would be ex-
pected), and farmers (whose value tends to be lower than expected). A
similar finding is reported by Hayden and Cannon (1984:190), who show
that in contemporary highland Maya villages, increases in household in-
come are invested differently by merchants and farmers. Merchants are
more likely to invest in household items (“luxury items”), while farmers
are more likely to invest in agricultural capital like land or livestock (see
Douglas and Isherwood (197950) for discussion of wealth-related con-
sumption differences between farmers and others in modem settings).
Because of these patterns, archaeologists should attempt to determine
the occupations of household members as part of any study of household
wealth.
On the basis of these anthropological, sociological, and historical
studies, it is clear that the number and type of possessions are strongly
308 MICHAEL E. SMITH

associated with total household wealth in a variety of cultural settings.

The task of the following section is to determine what kinds of household
goods are the best predictors of wealth.4

TYPES OF GOODS THAT REFLECT WEALTH

Functional Classification

The most useful approach to the classification of household posses-

sions as wealth indicators is one based upon the function of material
items. Different functional categories relate to wealth in different ways,
and most historians and ethnographers dealing with wealth and material
culture typically employ functional classifications (e.g., Jones 1980;
Lewis 1973; Douglas and Isherwood 1979; see also South 1977:27Of). Be-
cause ceramics tend to be the most abundant class of artifact in agrarian
societies, archaeological analyses of wealth often depend heavily upon
ceramic data; archaeologists should thus base their studies on a func-
tional classification of ceramic artifacts (see Hally 1986; Smith 1985).5
This paper employs a functional classification modified after that used by
Lewis in his study of modern Mexican households (see Table 3 above).
1. Furniture. Numerous studies have demonstrated a strong positive
association between household wealth and the quality and amount of fur-
niture. In modern Mexico, furniture is one of the strongest predictors of
wealth as shown by Lewis’ inventory analysis (1973; see Table 3 above)
and DeWalt’s Guttman scale analysis (1979; see Table 2 above). Jones
(1980:229) notes a clear association between furniture and wealth for co-
lonial America, Chapin (1935:383ff), Guttman (1942), and Sewell
(1940:40) employ items of furniture as wealth indicators in their sociolog-
ical level-of-living scales, and Castro et al. (1981:413) discuss a cross-cul-
tural relationship between furniture and wealth in Third World rural set-
tings today. This clear role of furniture as a predictor of household wealth
probably derives from a combination of utilitarian and symbolic factors.
Furniture tends to be one of the most expensive categories of households
goods (at least in rural or nonindustrial settings), and because it com-
prises a relatively high proportion of total physical wealth (see Table 3), it
serves as a good indicator of total wealth, if only for numerical reasons.
Furthermore, the semipublic context of much household furniture gives it
a significant symbolic role in expressing wealth and status to nonhouse-
hold members (the symbolic role of material items in relation to wealth is
discussed with serving ware below).
In spite of the strong association between furniture and wealth in
modern settings, there are three reasons why this category is not particu-
 HOUSEHOLD POSSESSIONS AND WEALTH 309

larly useful as an indicator of wealth in archaeological settings. First,

most preindustrial furniture is made of perishable materials (primarily
wood) which does not often survive in the archaeological record. Second,
the high value and long use-life of many items of furniture (see Table 3)
cause them to be curated, reused, and scavenged at a high rate (Schiffer
et al. 1981), thus inhibiting their deposition into the archaeological record
in direct association with their primary location of use. Third, prior to the
modem era, most nonelite households in agrarian states simply had very
little furniture in their homes (Braudel 1981:283ff), thereby limiting the
usefulness of this category to wealth assessmentwithin the elite class or
to separation of elite from nonelite contexts.
2. Clothing. As in the case of furniture, clothing tends to form a rela-
tively large component of total household wealth in many societies
(27.4% of the value of household possessions in Panaderos-see Table 3
above; 20.9% of the value of consumers’ durables in the thirteen Amer-
ican colonies-Jones (1980: 103)), and it is positively associated with
wealth (Lewis 1951:57, 1973; Jones 1980:330@,Castro et al. 1981:413)for
what are probably similar reasons. The extremely rare preservation of
clothing in the archaeological record effectively rules it out as a useful
indicator of household wealth, however.
3. Tools and household equipment. This category includes all utili-
tarian equipment and items found in household contexts except for food
preparation and serving items (which are treated as separate categories
below). Although Lewis (1973) found that tools (specialized production
items) were one of the best indicators of household wealth in Eanaderos
(see Table 3 above), the use of such items in predicting household wealth
under preindustrial conditions is problematic. Ethnographic studies show
that a number of household items are good wealth indicators (e.g., televi-
sions and radios, sewing machines, washing machines, etc.-see Table 2
above), but in all cases these are costly industrial items. Douglas and
Isherwood (1979:95-113) discuss the role of such goods in relations to
wealth and status in modern societies; however, it is difficult to imagine
preindustrial equivalents of such items.
Four types of utilitarian goods may be distinguished: household main-
tenance items, tools used in domestic production for use, agricultural im-
plements, and tools for specialized artisan production (for exchange).
(1) Household maintenance items are the basic equipment needed for
cleaning, repair, and maintenance. Because the need for these items is
relatively constant across households, several archaeologists have sug-
gested that they are not good wealth indicators (Hayden and Cannon
1984:40-95, 189; Drennan 1983).
(2) Tools used in domestic production for use, like household mainte-
nance items, have a widespread distribution and do not appear to have a
310 MICHAEL E. SMITH

strong association with wealth in preindustrial settings. Cross-culturaliy,

textile manufacture is one of the most common types of domestic produc-
tion activity; in addition to textile production tools, archaeologists are
likely to recover evidence for production of lithic, wood, and sometimes
ceramic items for domestic consumption in household contexts. Elevated
frequencies of such items might correlate with larger household size,
which in turn is often associated with wealth (see above), but this is an
unexplored and tenuous link. In highly stratified societies where nonelite
households provide labor service and/or goods for the elite, there could
easily be a negative association betwen the abundance of such domestic
production tools and wealth (i.e., nonelite households with greater ser-
vice/tax obligations would have evidence of a higher rate of domestic
production).
(3) Agricultural implements. Outside of modern settings where farmers
have access to industrial farm equipment (see Castro et al., 1981:404),
there is little information on the relationship between a household’s agri-
cultural implements and its level of wealth. It appears unlikely that such
items would be good wealth indicators.
(4) Tools for specialized artisan production have a negative association
with household wealth in many non-elite settings. Arnold (1985: 171- 198)
reviews cross-cultural data revealing that in many parts of the world,
pottery specialization in peasant households is related to land shortages
and poverty, and potters in these societies tend to be of lower socioeco-
nomic status than their farming neighbors. Cook (1982:74ff, 129ff) dis-
cusses this situation for peasant artisans in general and provides parallel
data on specialized stoneworkers in rural Oaxaca, Mexico. Within an
elite class, however, specialized production tools may have a positive
association with wealth if there is a tendency for wealthier households to
have attached craft specialists. In summary, tools and household equip-
ment are not good wealth indicators, and two subcategories-tools used
in domestic production for use and artisans’ tools-may have a negative
association with household wealth in certain settings.
4. Food preparation items. Food preparation items reflect wealth to
the extent that variations in diet and/or techniques of food preparation
are associated with wealth level in a society. Numerous ethnographic and
historical studies have demonstrated a strong correlation between dietary
quality and wealth in agrarian state societies. For example, Lewis
(1951: 189) reports that in Tepoztlan, Mexico, wealthy households con-
sume more food, a wider variety of foods, and consume desirable foods
at a higher frequency than poor households. DeWalt (1983) finds a similar
pattern in Temazcalcingo, Mexico, and provides detailed quantified data
on diet and its relationship to a wide range of sociocultural variables (in-
cluding household possessions-see discussion of Table 2 above). The
 HOUSEHOLD POSSESSIONS AND WEALTH 311

wealth-related differences in dietary quality (primarily the amount of pro-

tein and dietary diversity) are sufficient in this area to produce a situation
of clinical undernutrition in poorer households (DeWalt 1983:115-120,
141-147). Castro et al. (1981:414-416) report cross-cultural data for the
association of dietary quality and wealth; additional examples may be
found in Yang (1945:22ff), Braudel(1981:183ff), and Otto (1977, 1984). As
noted above, the most direct way to investigate this situation archaeo-
logically is through osteological remains. However, wealth-related di-
etary differences are also reflected in household material culture invento-
ries,
One way in which wealth-related dietary differences may be expressed
materially is through distinctions in the type of cooking vessels employed
for different foods. There is some evidence for this pattern from ethno-
graphic studies (Lewis 1970:432; Sewell 1940:40), and Otto (1984:48, 60,
112) provides sketchy but suggestive archaeological and historical data
from the Cannon’s Point plantation. At Teotihuacan, where wealth- or
status-related artifactual patterning has been explored in a preliminary
fashion, there are suggestions of differential food preparation techniques
(perhaps associated with dietary differences) between low- and interme-
diate/high-status residences on the basis of differences in the frequencies
of San Martin Orange ware and plain ollas in surface collections of the
Teotihuacan mapping project (Cowgill et al. 1984:166; Smith 1975:71-75).
As archaeologists begin to associate specific ceramic types with specific
food types through an increased concern with ceramic vessel function
(see Hally 1986; Braun 1983), it will become clearer whether or not this
sort of pattern will be of use for wealth analysis.
Within most ethnographically documented settings, however, the basic
repertoire of food types and the cookware used to prepare them is rela-
tively constant across households of different wealth levels (e.g., Lewis
1951:189, 1973:271f, Deal 1983), although there may be an association
between the quality of workmanship of cookware and wealth. Wealth-re-
lated dietary differences tend to be expressed in the total quantity of food
consumed and in the relative proportions of different food types, so ar-
chaeologists need to look more at quantitative artifact differences or at
different methods of preparation for a basic food set. For example, Nelson
(1981: 121) found that the aggregate volumes of several functional ceramic
categories in a highland Maya village are associated with several wealth-
related variables, including amount of farmland planted and “social
rank.” A number of ethnographers have noted that in rural Mexico,
wealthier households have raised cooking facilities while their poorer
neighbors use the traditional three-stone hearth at floor level (see Table 2
above; Lewis 1951:183). A similar pattern may have characterized
Classic Teotihuacan, where at least two types of cooking facilities have
been identified: simple floor-level hearths and ceramic braziers known as
312 MICHAEL E. SMITH

“three-prong burners” (Millon 1981:203). These braziers are statistically

associated with intermediate- and high-status residences and are un-
common in low-status residences (Smith 1975:71-75), while the distribu-
tion of hearths has not been reported. It is quite possible that these two
types of cooking facility may serve as wealth indicators for Teotihuacan
residences.
Hayden and Cannon (1984:74, 81) document a situation in which food
preparation items have a negative association with wealth. Metates
(grindstones) in their sample have a weak negative correlation with in-
come, which they interpret as due to “high income households either
buying already ground maize or finding alternative mechanized means to
grind their maize” (Hayden and Cannon 1984:81). Similarly, at Teoti-
huacan preliminary analyses indicate that manos and metates are less
common in high-status houses than in low- or intermediate-status houses
(Smith 1975:73-75). The most parsimonious behavioral interpretation
would be that lower status households were responsible for grinding
maize for the elite. As suggested for tools and household equipment
above, the nature of labor service arrangements in stratified societies has
important implications for the use of material items as wealth indicators.
Conquest-period ethnohistoric documents from Morelos, Mexico, de-
scribe a situation where commoner women were responsible for grinding
maize (and weaving textiles) for noble households, but they lived nearby
and went to the elite house to do the work (Smith n.d.). In this case,
evidence of such activity (manos and metates) would have a positive as-
sociation with wealth (since elite households probably consumed more
maize than commoner households). In summary, food preparation items
may be good wealth indicators in many situations, but more comparative
work is needed in this area. To the archaeologists’ benefit, many objects
involved in food preparation (particularly storage and cooking vessels)
are made of ceramics and stone in preindustrial settings, and their rela-
tively low value makes them less susceptible to curation, scavenging, and
other processes that hinder direct deposition into the archaeological
record in association with their domestic context of use.
5. Serving ware. This category of items is a strong predictor of house-
hold wealth in many ethnographic and historical settings, and appears to
represent the most useful class of archaeological artifact for assessing
wealth. The conceptual framework for understanding the relationship of
serving ware to wealth is provided by Douglas and Isherwood (1979) in
their treatment of household consumption patterns. They focus on ways
in which people use household goods to communicate information. One
of the primary roles for such communication in nearly all societies is the
conveyance of signals concerning wealth and status to individuals of
other households. Serving ware has a relatively public domain compared
 HOUSEHOLD POSSESSIONS AND WEALTH 313

to other household items and is therefore a common vehicle for the com-
munication of information (Wobst 1977:328-330); for this reason, serving
items are likely to be used as “markers” (Douglas and Isherwood
1979:74-76) of wealth and status.
An important setting for the symbolic use of serving ware as a marker
of wealth and status is the household consumption ritual, a social gath-
ering within the context of the household that includes the consumption
of food, drink, and other goods (Douglas and Isherwood 1979:114-127,
15Off). Household consumption rituals in modem United States society
include dinner parties and cocktail parties; examples from other cultures
include the Latin American family fiesta, cooperative work parties, and a
host of culturally specific household social occasions associated with
life-course rituals like baptisms, comings-of-age, weddings, funerals, and
so on. These are all the occasion for non-household members to gather
and be served food and drink in vessels that convey information on the
wealth and status of the host houshold. Consumption rituals can become
quite elaborate affairs, as in the case of the Northwest Coast potlatch or
elite banquets in state-level societies.
According to Douglas and Isherwood (1979:114- 170), within any given
cultural setting, wealth and status are closely related to the frequency,
scale, and nature of household consumption rituals. Wealthier house-
holds tend to (1) host more frequent rituals; (2) host more elaborate
rituals with more guests and more food and drink; and (3) use serving
ware that is more prestigious (i.e., more valuable and symbolically signif-
icant) for both periodic consumption rituals and everyday meals. The re-
lationship of wealth to the frequency and scale of consumption rituals in
rural settings is described briefly by Castro et al. (1981:414). Lewis re-
marks specifically on the material expression of consumption ritual scale
in Tepoztlan as follows:
Wealthier families generally own enough plates, pots, glasses, spoons, chairs, etc. to be
able to serve fiesta meals to a large number of people without having to borrow as do
many other Tepoztlan families (Lewis 1951:183)

Thus wealth should be expressed in the number of serving vessels pos-

sessed by a household and in the value or costliness of those vessels.
Perhaps the best evidence of the strong positive association between
the amount and value of serving vessels and household wealth comes
from historical and archaeological studies of early American ceramics.
During the eighteenth and nineteenth centuries, ceramic vessels were the
primary type of serving vessel for all social classes. The tea ceremony
was a ceramics-based household consumption ritual expressing wealth
and status that was practiced by a large segment of the population, both
rich and poor (South 1977:40-43, 231; Otto 1984:166). Household inven-
314 MICHAEL E . SMITH

tories and archaeological studies show that as wealth increases: (1)

households tend to have larger ceramic inventories; (2) a greater propor-
tion of households possesses expensive wares (like Chinese porcelain)
and they tend to own a greater number of such vessels; and (3) the value
and size of tea sets increase (Teller 1968; Stone et al. 1973; South
1977:40-43; Otto 1984:64-69, 89f, 112, 166f; Miller 1980).
This strong association between serving ware and wealth is at the base
of archaeologists’ traditional use of elaborate painted pottery to infer
wealth and status (e.g., Smith 197554, Bawden 1982:175), simply be-
cause painted vessels are often serving vessels. Fancy decorated ware is
usually of considerably higher value than plain serving ware (in terms of
both energetic and symbolic value- see discussion of value below), and
a number of archaeologists have posited an association between such
painted pottery and elite groups in ancient states (e.g., Smith and Heath-
Smith 1980:31; Blanton et al. 1981:246ff). A complication arises, how-
ever, when valuable serving ware (or other goods) is not limited by law to
the elite. Many less wealthy households may possess valuable items
identical to those used by the elite, and thus it is not completely correct
to refer to these goods as “elite wares” (see Fried1 (1964) and Clark
(1986) on lower-class emulation of elites). In such a situation, it should be
possible to distinguish less wealthy from wealthier household inventories
on two grounds: (1) wealthy households will have a higher frequency of
valuable wares, and (2) their valuable serving pieces will exhibit more
use-wear. The latter consideration derives from the observation that less
wealthy households that own very valuable serving ware tend to use it
much less often than do wealthy households who have the same ware
(Miller 1980:13; Douglas and Isherwood 1979: passim). Studies of ce-
ramic use-wear, a new and promising field, are thus directly relevant to
investigations of household wealth that employ ceramic serving vessels
(see Grifftths (1978:77) for an example). It should be noted that some
attrition of systemic inventories must be expected for the most and least
valuable kinds of serving vessels in many cases. Perishable items (e.g.,
wooden bowls, gourds) are often used as less valuable serving vessels
(Jones 1980:330; Deal 1983:136ff), and the costliest wares will have
higher rates of curation and scavenging than less valuable wares.
6. Religious items. In evaluating the relationship between household
religious items and wealth, a distinction must be made between public
and domestic religion. In agrarian states wealthier households tend to
have a higher level of participation in public religious observations than
do poor households, and this results in a positive association between
wealth and religious paraphernalia involved in such ritual. Poorer house-
holds, on the other hand, often show a greater concern with private fa-
milial ritual, and the material expressions of such ritual therefore show a
 HOUSEHOLD POSSESSIONS AND WEALTH 315

negative association with wealth. The Mesoamerican civil-religious

cargo system provides a good example of the linkage of public religious
observation to wealth in agrarian societies. In areas where the cargo
system still plays an important irole in community integration, wealthier
households clearly participate at a higher level, sponsoring more and
larger fiestas than poorer households (Cancian 1965; Hayden and Cannon
1984:152- 173). Furthermore, Hayden and Cannon show that the level of
participation of a household in civil and religious cargos is reflected in the
frequency of associated material items in the residential setting (1984:96-
106, 164-171; see also Deal (1983:146)), although the statistical associa-
tion is not high (for reasons which they discuss). In areas of Mesoamerica
where the cargo system rituals no longer play an important role in com-
munity solidarity and integration, however, the association between par-
ticipation and wealth is not found (e.g. Lewis 1951:272f); in fact, DeWalt
reports a negative correlation between cargo participation and wealth in
Temascalcingo (1979: 123).
When a private familial ritual is considered, poor households in central
Mexico show a greater participation and a greater frequency of relevant
material items (most of which are associated with household altars). In
Panaderos for example (see Table 3 above), the value of a household’s
religious items exhibits a (very weak) negative correlation with its total
value of possessions,and the two poorest households contained religious
items worth nearly three times as much as those of the two wealthiest
households. In Tepoztlan, Lewis (1951:214-218) found that although a
typical rich household spends somewhat more annually than a typical
poor household on both religious items and other religious fiesta ex-
penses, such religious expenses comprised a much larger proportion of
the total household budget for the poor family (3.2% compared to 0.8% of
total annual expenses). Lewis also compared households in Casa Grande
(a middle-income residential complex) with the much poorer households
of Panaderos, and found a greater intensity of household religious obser-
vation among the latter group (1970:437-439). DeWalt (1979:123) reports
that his variable “religiosity” (a scale based on church attendance and
frequency of communion) has a negative correlation with wealth. If this
religiosity also extends to household ritual participation (which is likely),
then the pattern is similar to that found by Lewis in Tepoztlan and
Mexico City.
The positive association between wealth and involvement in public re-
ligious observance has also been documented for traditional China. An-
cestor worship was carried out in two types of context-the household
and the lineage ancestral hall-differentiated by ideology, rituals, and
material objects involved (Feedman 1958:81ff). The scale of worship at
the ancestral halls was clearly related to lineage wealth (which in turn is
316 MICHAEL E . SMITH

associated with household wealth of lineage members); poor lineages

might not have an ancestral hall, while the quality of construction of halls
and the frequency and scope of ritual activity were expressions of wealth
(Freedman 1958:48ff, Yang 1945: 137ff). Worship at ancestral halls may be
considered public in the sense that (1) many separate households were
involved; (2) the location of activity was not in a domestic setting; and (3)
the rituals served the function of extrahousehold social solidarity. I have
not been able to find any information on the association between do-
mestic ancestral worship (or other domestic ritual) and wealth in tradi-
tional Chinese society. One example of the suggested relationship for a
quite different cultural setting is provided by Laumann and House
(1970:Fig. 1) in their study of contemporary United States living rooms.
They examined the relationship of particular items to “status” (which in
turn is related to wealth), and found that two of the types of goods lowest
in status were religious objects and religious paintings.
These findings on the disjunction between public and private religion in
their association with wealth reinforces the emphasis placed on artifact
function above. Archaeologists need to establish first that a given class of
objects served a religious or ritual function, and second, objects used in
domestic ritual should be separated from those used in public rituals.
This will probably be a difficult task in most archaeological settings, but
the issue must be dealt with if religious items are to be used in the anal-
ysis of wealth.
7. Nonutilitarian luxuries. Luxuries and necessities are defined by
economists as follows:
Necessities are defined as those goods which are bought in the same quantities regard-
less of changes in prices or in supply . . . Luxuries, by contrast, are a completely
heterogeneous class defined as those goods on which the individual will quickly cut
down, in response to a drop in income. (Douglas and Isherwood 1979:97)

Douglas and Isherwood proceed to divide luxuries into two types:

An anthropological definition allows of two kinds of luxuries-one the mere rank sig-
triflers, such as the best china for the family christening; the other, the newest techno-
logical aids, the innovating capital equipment, which relieve the pressure on available
time, space and energy. (Douglas and Isherwood 1979: 112).

Most discussions of luxuries by anthropologists have concerned the

second type-industrially produced technological aids. Hayden and
Cannon, for example, define luxuries as “all items which could be con-
sidered to indicate an expenditure beyond standards for most people in
the village” (1984:125), and their list of luxury items in Maya villages
(1984: 107) is dominated by technological aids and other costly industrial
imports (they report a positive but weak and inconsistent relationship
between luxury item possession and wealth (1984:112, 128)). The
 HOUSEHOLD POSSESSIONS AND WEALTH 317

Guttman scale inventory studies listed in Table 2 above are also domi-
nated by technological luxuries like automobiles, televisions, and refrig-
erators. In short, most ethnographic studies of luxury goods and wealth
focus on items without counterparts in preindustrial settings, and are
therefore of limited usefulness to archaeologists.
For the present analysis, it is useful to single out nonutilitarian luxu-
ries. This category eliminates industrial items as well as high-value utili-
tarian goods (like serving or food preparation ware) that are discussed
separately above. There is considerable evidence linking nonutilitarian
luxuries to household wealth. Lewis’ data from Panaderos (Table 3
above) show that jewelry has one of the strongest associations with
wealth of any category. In Laumann and House’s (1970: 192) study of
modem living rooms, luxuries like sculpture and abstract paintings are at
the high end of their (wealth-related) status scale. Archaeological ex-
amples of the association between nonutilitarian luxuries and wealthy
residences are common. For example, almenas (stone and ceramic roof
makers of high value) are associated with high-status residences at Teoti-
huacan (Smith 1975:75); imported shell, jade, and other valuable objects
are markers of wealthy households at Tikal (Haviland 1981:104), and
silver is associated with wealthy residential contexts at Galindo in Peru
(Bawden 1982:175).
Although the luxuries under discussion are termed nonutilitarian (in
the sense that they do not have useful daily functions in a household
context), they often play an important role in the political systems of
agrarian states, serving as “political capital” to help regulate political
and social interaction between segments of the ruling elite class. As a
result, control over the production and exchange of such items is an im-
portant part of the power base of local elites. This role was particularly
important in the Aztec state (Blanton and Feinman 1984:676; Smith 1986;
Brumfiel n.d.), and a similar pattern is suggested by D’Altroy and Earle
(1985) for the Inca. Luxury items employed in this manner are often
sumptuary goods, and their distribution in household contexts is there-
fore a good indicator of (1) elite vs nonelite residences, and (2) wealth
variation within the elite class. In spite of their great value as wealth
indicators in the systemic context, luxuries are of lesser use archaeologi-
cally. This is partly because such items are often made of perishable ma-
terials, but more importantly their high value produces very low rates of
loss and discard and a high rate of reuse and scavenging. When these
items do enter the archaeological record, however, they are good wealth
indicators. Useful theoretical discussions of luxuries and their relation-
ship to wealth and other social factors are found in Douglas and Isher-
wood (1979), Braudel (1981), and Appadurai (1986).
318 MICHAEL E. SMITH

Summary

The above data indicate that there is a wide variation in the usefulness
of different functional categories of goods for measuring the wealth level
of households in agrarian societies. These findings are summarized in an
impressionistic manner in Table 5, where usefulness as wealth indicators
in systemic context is assessed in three categories-within the elite,
within the commoner class, and between elite and commoner house-
holds. Based upon the available ethnographic and historical sources,
clothing and serving ware appear to be the strongest and most consistent
wealth indicators in agrarian societies, followed by furniture and non-
utilitarian luxuries. These are not absolute judgments, however, since
various functional classes will reflect wealth differently in different cul-
tural settings. For archaeologists, two primary types of factors detract
from the usefulness of various categories for goods for wealth analysis-
lack of preservation in the archaeological record and the operation of
cultural formation processes like curation, reuse, and scavenging that
place limitations on the archaeological context as a direct expression of
the systemic context of goods. These detractions are minimal for food
preparation items and serving ware (Table 5), and as a result, the latter
category becomes the most effective functional class of measuring wealth
in archaeological situations. Rather than focusing all attention on this cat-
egory, however, archaeologists are advised to examine possible wealth-
related variation in a number of functional classes, and to combine this

TABLE 5
SUMMARYASSESSMENTOFTEIEUSEFULNESSOFWEALTH INDICATORS
Usefulness within:
Systemic context’
Functional Within Elite vs Within Archaeologicalb
Category elite commoners commoners context
Furniture ***e *** * -
Clothing *** *** *** -
Tools and household equipment * * * *
Food preparation equipment * * * ***
Serving ware *** *** *** ***
Religious items * *** * *
Nonutilitarian luxuries *** *** * *

a These judgments are based upon the data presented in the text.
b Archaeological usefulness is based upon a combination of preservation factors and the
effects of cultural formation processes as discussed in the text.
c ***, Consistently useful category; *, potentially useful in some cases; -, not useful.
 HOUSEHOLD POSSESSIONS AND WEALTH 319

approach with an examination of properties of overall household invento-

ries.

HOUSEHOLD INVENTORY CHARACTERISTICS AND WEALTH

Ethnographic and historical studies show that in addition to variation

in specific functional categories of goods, household wealth is also ex-
pressed in a number of properties of domestic material culture that cut
across function. Three such properties are examined here-quantity, di-
versity, and origin.

1. Quantity

Wealthier households tend to have more material goods than do poorer

households. Lewis reports this finding for both Tepoztlan (1951: 182,
213-219) and Panaderos (1973:269). For the latter settlement, the number
of goods in a house has a positive correlation with the total value of
household goods (which measures wealth as discussed above). Kottak
notes this relationship in Arembepe, Brazil (1983:93), and Schiffer et al.
(1981:78f) report that inventory size in modern Tucson is more closely
related to income than to household size. Because wealthier households
possess more goods and consume more food (see above) than do poor
households, they produce more domestic refuse per day (Phillips, et al.
1984:14% 148), a finding of direct relevance for the archaeological study
of household wealth.

2. Diversity

Not only do wealthier households possess more goods than do poor

households, they also possess a greater variety or diversity of material
goods (Lewis 1973:269; Douglas and Isherwood 1979:143f, 191f). Several
archaeologists have noted this relationship between wealth and diversity
(Smith 1975:74; Haviland 1981:103f), and Hayden and his associates
found that diversity measures are often better wealth predictors than are
frequency measures (Hayden and Cannon 1984:208f; Deal 1983:252ff,
293ffJ. Because of the strong infhrence of sample size on the level of di-
versity, a method like Kintigh’s (1984) simulation approach to the quanti-
fication of diversity should be employed in place of the simple measures
proposed by Hayden and Cannon (1984) and Deal (1983).
320 MICHAEL E. SMITH

3. Place of Origin
Material objects of exotic origin are good wealth indicators for a
number of reasons. To start with, such items tend to have a high value
because of the energetics involved in their procurement and their resul-
tant scarcity within a given economic system (Polanyi 1968: 185ff;
Drennan 1984:33-see the discussion of value below). For this reason,
objects used as currency in nonwestern economies are often imported
goods (Schneider 1974: 164f). However, the value of imported items often
goes beyond the simple factors of energetics and scarcity. Helms (1979),
for example, documents a situation in which exotic trade goods are en-
dowed with ritual power in lower Central American chiefdoms. Chiefs
compete with one another for control over the trade and circulation of
these items, and they therefore serve as good markers of the wealthier
chiefly group as contrasted with other households.
A more common link between exotic goods and wealth is based upon
the observation that wealthier households tend to have geographically
wider networks of social interaction and identification than do poorer
households (Douglas and Isherwood 1979: 16off; Smith 1986). These long-
distance interaction systems have two types of direct expression in
household material culture. First, there will be a higher frequency of ex-
otic goods, as well as goods from farther away, in households that partici-
pate in larger spatial networks. Second, such households will have a
higher frequency of goods expressing stylistic interaction covering wider
horizons. These goods may be locally produced, but exhibit decoration
that is part of a geographically extensive stylistic network.
Douglas and Isherwood (1979:16Off) discuss their relationship of wealth
to the spatial range of social interaction in general terms. DeWalt
(1979:223-229) presents a factor analysis of households in Temascal-
cingo, Mexico, in which variables representing income and material pos-
sessions (see discussion of Table 2 above) both load highly on a factor
interpreted as “extralocal orientation.” Smith and Heath-Smith (1980)
and Rice (1983) both interpret examples of widespread decorative ce-
ramic styles in Postclassic Mesoamerica as evidence of elite-oriented
communication networks (rather than as evidence of migrations or
simply stylistic diffusion as prior interpretations stressed). Other archae-
ologists who have used exotic goods to measure wealth include Wright
(1972), Haviland (1981:104) and Rathje and McGuire (1982:712). Addi-
tional ethnographic evidence for this association is provided by Lewis
(1973:257), Kramer (1982:74), and Deal (1983:370).

METHODOLOGICAL ISSUES
The material presented above demonstrates that within the systemic
 HOUSEHOLD POSSESSIONS AND WEALTH 321

context, household wealth in agrarian societies is clearly reflected in ma-

terial culture inventories. In cases where investigators have reported lim-
ited or weak associations between wealth and household items, the ex-
planation probably lies in the limited degree of wealth variation within the
context studies or in a lack of a comprehensive treatment of household
items (see note 3). However, before archaeologists can take advantage of
these findings and use household artifacts to measure wealth in prehis-
toric settings, a number of methodological problems must be dealth with.
Some of these have been alluded to in the text above, and four such
issues are singled out for treatment here: measures of value, confounding
variables, cultural formation processes, and the problem of household
and community comparisons.

1. Measuring Value

Economists and economic anthropologists have traditionally ap-

proached the question of value from two perspectives. The Marxist per-
spective holds that a commodity’s value is determined by the amount of
labor involved in its production (O’Laughlin 1975:352f), while the neo-
classical perspective holds that value (or price in monetized economies)
is determined by the interplay of supply and demand (e.g., Schneider
1974:43-84; see Godelier (1977:146ff) on the interplay of labor and scar-
city in determining value in a nonmonetary economy). Douglas and
Isherwood (1979: 114- 127), expressing dissatisfaction with prior ap-
proaches to value, take a different tack and link the value of a good to the
periodicity and social significance of its consumption. They describe this
relationship (in terms of necessities versus luxuries) as follows:
The cultural aspect of necessities is revealed in their service in low-esteem, high-
frequency events, while luxuries tend to serve essentially for low-frequency events
that are higher esteemed . . . In our experience with goods, periodicity in con-
sumption use marks rank and creates quality goods. (Douglas and Isherwood
1979: 116)

Of these three approaches to measuring value-labor, scarcity, and

periodicity-the former is most useful for archaeologists. Use of scarcity
to measure value would be misleading with archaeological data because
the operation of a number of cultural formation processes virtually guar-
antees that the relative scarcity of valuable items recovered archaeologi-
cally will not match their relative scarcity in the systemic context (see
discussion below). Some notion of periodic@ of consumption can be
gained from artifact use-wear studies (see discussion of serving ware
above), but this is not feasible for large-scale analyses based upon frag-
mentary artifacts (e.g., sherds). The labor involved in the production of
322 MICHAEL E. SMITH

different types of goods, however, can be determined by archaeologists

through careful use of analogy. Labor input is best measured by the time
involved in production (Gross 1984), and if exotic goods of known origin
are present, transport costs (in labor-time) can be added to production
costs. Archaeologists have begun to use such “energetic” measures of
value (see for example, Feinman et al. (1981) or Sanders and Santley
(1983) on ceramic production, and McGuire (1983) or Abrams (1984) on
architecture), and continuing ethnographic (Gross 1984) and ethnoar-
chaeological (Mandeville 1974) studies will help refine this approach in
the future.
Some consideration should also be given to the matter of symbolic
value, which may or may not derive from the energetic (or scarcity) value
of an item. For example, elaborate serving ware is shown above to be
strongly associated with wealth for symbolic reasons that surpass the
simple energetic value of individual items. Because many material goods
with a high symbolic value also have a significant stylistic component,
the degree of stylistic expression of an object can serve as a crude mea-
sure of its symbolic value. Therefore, while decorated serving ware will
be assigned a high value because of the labor involved in its production, it
should probably also be given an independent rating for symbolic value
when household wealth comparisons are carried out.

2. Confounding Variables

The kinds and quantities of goods present in households in systemic

context are affected by a number of socioeconomic factors besides
wealth. Among the more significant of these, Schiffer et al (1981:83) list
the following three.
(I) Stage in the domestic developmental cycle. The wealth level of an
individual household tends to increase throughout its normal develop-
mental cycle (as more physical wealth is accumulated), and this factor
must be controlled when making comparisons among households (Jones
1980:166ff). Most types of artifactual remains from households do not
pertain to a single moment in time, however (except for de facto refuse
left at the house’s abandonment). Primary and secondary refuse and pro-
visional discard should generally include material from one or more com-
plete domestic developmental cycles, and the length of time involved will
serve as a control for this confounding variable.
(2) Degree of residential stability. This factor measures the frequency
with which a household changes residence. Archaeologists are interested
in the inverse perspective-how many different households have lived at
a particular dwelling site? Residential stability may be higher in preindus-
trial communities than in the setting of Schiffer et al.‘s study (modern
 HOUSEHOLD POSSESSIONS AND WEALTH 323

‘Ihcson), and this variable therefore would not be expected to have signif-
icant adverse effects on archaeological studies of the socioeconomic im-
plications of household artifact inventories.
(3) Time elapsed since the last move. This variable, which has a posi-
tive association with inventory size in the study of Schiffer et al.
(1981:78f), is of considerable importance to archaeologists. As the span
of occupation of a residence increases, not only does the total amount of
refuse increase, but the types of goods deposited change (for example,
rarer categories have a greater chance of being discarded or lost over a
long span of occupation than during a short interval-see Rathje and
Schiffer (1982:119)). Rathje (1983) proposes a method of controlling for
this variable by standardizing wealth measurements (e.g., frequencies of
luxuries in household deposits) in terms of utilitarian items whose fre-
quencies are strongly influenced by the span of occupation and relatively
independent of household wealth (he suggests using the frequencies of
flint tools, grain-grinding tools, and total sherds for standardizing mea-
surements).
In addition to these confounding variables isolated by Schiffer et al.
(1981), wealth measurements may also be influenced by factors of house-
hold size and structure, occupations of household members, or ethnic
group affiliation. Household size may infhrence wealth indicators like the
quantity and diversity of goods, or quantity of serving ware, but this may
not be a serious problem since household size has a positive association
with wealth in most preindustrial settings (Yanagisako 1979; Netting
1982). The occupations of household members (e.g., farmers vs mer-
chants) and the presence of specialized production activities within the
home are other factors which may confuse wealth assessment. It should
be possible to identify these situations in many archaeological cases, and
their relationship to wealth is discussed in the text above. Finally, eth-
nicity can affect household inventories, particularly in the realm of food
remains and ceramics (McGuire 1982). Intracommunity ethnic differen-
tiation in agrarian states is often correlated with wealth and power dis-
tinctions (Cohen 1978), and aracheologists have only begun to address
this issue (e.g., McGuire 1982). The disentanglement of ethnicity and
wealth thus presents a methodological problem that needs resolution. In
summary, some of the non-wealth related factors influencing household
inventories in the systemic context will have minimal effects on the ar-
chaeological study of household wealth, others can be identified and con-
trolled for by the archaeologist, and some factors like ethnicity will
present problems.

3. Cultural Formation Processes

Before any kind of social or behavioral inferences can be made from
324 MICHAEL E. SMITH

household remains, the nature of household deposits and refuse must be

determined for each study setting. A considerable amount of artifactual
variability derives from differences between such types of remains as pri-
mary refuse, secondary refuse, de facto refuse, provisional discard, etc.
(Schiffer 1976, 1983; Rathje and Schiffer 1982:105-125; Hayden and
Cannon 1983), and this variability must be controlled (or at least mini-
mized) before meaningful socioeconomic analysis can be undertaken.
Different types of deposits will probably reflect wealth differently, and it
is probably not valid to compare household wealth as derived, for ex-
ample, from de facto refuse in one case and secondary refuse in another.
Beyond the general necessity of documenting formation processes prior
to undertaking any kind of socioeconomic analysis, two specific kinds of
problems must be dealt with in the study of household wealth.
First, attention needs to be devoted to the question of “where the gar-
bage goes” (Hayden and Cannon 1983). Agrarian societies are sedentary
societies, and as such most of their domestic trash is deposited as sec-
ondary refuse (i.e., away from its location of use). Therefore archaeolo-
gists need to be able to associate specific secondary refuse deposits with
individual households. This becomes increasingly difficult as settlements
get larger and denser, because more secondary refuse gets removed from
residential areas for reasons of hygiene and convenience (Rathje and
Schiffer 1982: 116, 289). Dense urban settlements, where wealth variation
is often of great interest to social scientists, present the greatest obstacles
in this regard. However, even when settlements are relatively dispersed,
a good deal of household refuse may be removed from the vicinity of the
household, and refuse not removed in this manner may be scattered
widely around the houselot rather than placed in discrete dense middens
for the convenience of the archaeologist (Hayden and Cannon 1983; Deal
1983: 185ff).
The second major problem that cultural formation processes present
for wealth analysis is that a number of household goods that may be im-
portant wealth indicators in the systemic context will not enter the ar-
chaeological context in association with their household of primary use.
There are at least two components of this problem. First, valuable items
enter the archaeological record at a lower rate than items of little value
because they have higher levels of curation, reuse, and scavenging
(Rathje and Schiffer 1982: 11% 123; Hayden and Cannon 1983; Schiffer et
al. 1981; Hayden and Cannon 1984:26). This is particularly a problem
with luxuries as discussed above, although Rathje (1983) suggests that
jewelry and other items of personal adornment will still be more common
in refuse from wealthier households than poorer households due to their
occurrence in greater frequencies and their higher rates of use. Second,
some wealth indicators may be present in either higher or lower fre-
 HOUSEHOLD POSSESSIONS AND WEALTH 325

quencies than is to be expected in individual households due to other

cultural practices. For example, expensive utilitarian items are often bor-
rowed among households (Kramer 1982:741; Hayden and Cannon
1984:188); also, the nature of labor service arrangements in highly strati-
fied societies may lead to higher or lower concentrations of specific
wealth-related items (as discussed under tools and household equipment
above). The way around these obstacles to wealth analysis is no different
than for any other kind of social or behavioral study-the archaeologist
needs to ascertain the nature of household archaeological deposits as best
as possible, and to use this knowledge to guide assessmentsof individual
household wealth and comparisons among households.

4. Comparisons between Households and Communities

It would not be useful to make programmatic statements about how

archaeologists should go about measuring and comparing wealth among
households. One’s measures and techniques depend heavily upon factors
like the research setting (e.g., peasant villages vs urban centers), the na-
ture of the data (e.g., housefloor artifacts vs midden debris), the size of
the data base (e.g., four households, or 40, or 400), and perhaps most
important, the particular research problem at hand. Nevertheless, some
suggestions can be offered about promising approaches to the measure-
ment and comparative treatment of wealth from household artifacts.
First, it is advisable to use as many different functional categories of arti-
facts as possible. Data presented above show that while some categories
(e.g., serving ware) may express wealth more strongly and consistently
than others, many other functional classes can be effective wealth indi-
cators in some settings. Second, the greater the variety of specific mea-
sures employed, the more likely it will be that useful information on
household wealth will be generated. Simple frequency and diversity mea-
sures for different functional classesand subclassesare a good start; vari-
ables or scales based upon energetic or other assessmentsof value will
provide additional insights.
When artifactual data are the primary source of wealth information, it
is probably useful to begin with an inductive approach. The various
wealth-related scales developed for households can be subjected to a
factor analysis or perhaps cluster analysis followed by discriminant anal-
ysis of the clusters (see Cowgill et al. 1984). This should yield information
on the strongest and most useful variables for measuring wealth, and
these can be retained for further use. It may prove possible to construct a
single multivariate wealth scale (perhaps based upon factor scores) which
can be employed to give a unidimensional ranking of households. If inde-
326 MICHAEL E . SMITH

pendent assessments of wealth like architectural remains are available,

these can be employed together with the artifactual data (e.g., Bawden
1982) or else the two sets of data can be counterposed. For example,
wealth and status at Teotihuacan were investigated in one study by
starting with a threefold classification of residences (based upon size and
architectural quality) into low, intermediate, and high status. These cate-
gories then served as groups for a discriminant analysis whose goal was
to see whether or not the architectural groups could be distinguished by
artifacts, and if so, which artifacts were the strongest groups discrimi-
nators (Smith 1975).
If data for a sizable number of households are available and one is
interested in the nature and extent of wealth variation (inequality) within
a community or other population, it is possible to use Lorenz curves and
Gini indices (Jones 1980: 163- 169; 424f; McGuire 1983: 127- 131). The
Lorenz curve summarizes the concentration of wealth among households
(or other units) graphically, and the Gini coefficient expresses in a single
measure the degree of wealth concentration within the population. These
techniques are useful for summarizing wealth inequality in a simple intu-
itive manner, and for comparing inequality among diverse social settings.
To apply them to archaeological household data, there are two require-
ments: (1) a unidimensional wealth scale is needed (or perhaps several
scales for parallel analyses); and (2) there must be comparable measure-
ments for a sizable number of households selected through probability
sampling techniques. This section has mentioned only a few of the many
approaches that can be usefully applied to the problem of comparing
wealth levels and patterns within and between communities; as stated
above, one’s particular measures and methods must be determined sepa-
rately in each case.

CONCLUSIONS
This paper represents an attempt to stimulate the systematic compara-
tive study of the relationship between household possessions and wealth
in agrarian states and its application to archaeological data. While the
treatment of ethnographic, historical, and archaeological case studies is
by no means complete or comprehensive, the sources cited above do
clearly demonstrate a consistent and strong association between the
kinds and amounts of goods found in households in systemic context and
the level of household wealth. This general relationship holds in a number
of diverse sociocultural settings and has been documented quantitatively
by anthropologists, sociologists, and historians. Apart from a few obser-
vations on the relationship of quantity, diversity, and overall value of
goods to wealth, however, the material expression of household wealth is
 HOUSEHOLD POSSESSIONS AND WEALTH 327

not a simple affair. Different functional classes of goods reflect wealth

differently, and the nature of the relationship varies between and within
socioeconomic systems. What is needed now is a more comprehensive
and systematic examination of the published literature (along the lines
proposed by Blanton 1984) and the expansion of relevant ethnoarchaeo-
logical studies (such as Hayden and Cannon (1984) and Deal (1983)) to
explore more fully the range of variation and the principles at work be-
hind that variation.
In moving from the systemic context to the archaeological context
(Schiffer 1976), the association between household possessions and
wealth becomes weaker. However, when cultural formation processes
and other methodological issues are addressed it is clear that household
artifacts can still be used to monitor wealth in archaeological settings.
They may not represent the strongest class of wealth indicators (residen-
tial architecture probably reflects wealth more directly and consistently
(Blanton 1984)), but household artifacts are certainly adequate in them-
selves for measuring wealth and can provide additional insights when
used in conjunction with architectural remains. As archaeological studies
of households become more common, the use of artifactural measures of
wealth will assume greater significance and a more systemic method-
ology is required. This paper suggests some of the potentials of this ap-
proach, but considerable comparative and substantive work remains to
be done.

ACKNOWLEDGMENTS

This paper is a revised version of a presentation (“The Archaeology of Social Stratifica-

tion: Measuring Wealth from Household Artifacts”) given at the 50th Annual Meeting of the
Society for American Archaeology (Denver, CO, 1985). Some of the ideas discussed here
were first presented at a talk in the Department of Anthropology, Northwestern University,
in May 1984. I am grateful for useful discussions with James Brown, Michael Blake, Cyn-
thia Heath-Smith, and Richard Blanton; the comments of an anonymous reviewer were also
helpful. Brian Hayden and Richard Blanton kindly sent me unpublished materials related to
the theme of this paper.

NOTES

l This paper employs the common neo-Marxian anthropological definition of social strati-
fication as the division of a society into vertically ranked categories or classes of individuals
distinguished on the basis of differential access to basic resources (Fried 1967:185-191;
Adams 1975:243-265; Smith 1976). Stratification is closely related to the notion of in-
equality, and all stratification systems involve social inequalties, but the two concepts are
not identical. As discussed by M. G. Smith (1966), inequality is a universal feature of human
societies, whereas social stratification exists only where inequality is pronounced and insti-
tutionalized. Defined in this manner, social stratification is certainly not a universal at-
328 MICHAEL E. SMITH

tribute of human societies as some writers have maintained (Davis and Moore 1945), but
rather is limited to state-level societies. While the findings of this paper probably have
relevance to most agrarian societies, the primary focus is on wealth variation within
agrarian states.
* Before proceeding, some words should be said about the social context of wealth varia-
tion. While some inequalities of wealth among households are found in nearly all societies
(e.g., Kay 1964; Smith 1%6; McGuire 1983) they become extreme and pervasive only in
states. Within state-level societies, wealth differentials form an important component of the
overall system of social stratification (see footnote 1). In a socially stratified society, major
discontinuities in the distributions of wealth, status, and power (see Weber (1946) for the
classic treatment of stratification in terms of these variables) indicate the existence of social
classes. While some writers prefer to limit the term class to modem industrial society (e.g.,
Failers 1973). I favor a broader approach as exemplified by Lenski (1966) or Carol Smith
(1976). For Lenski, a class is “an aggregate of persons in a society who stand in a similar
position with respect to some form of power, privilege or prestige” (1966:74-75). In most
agrarian states, there are two classes-a small dominant class. or elite, and everyone else.
Measurable wealth variation in state societies is not limited to class divisions, however.
Considerable wealth differences within elite classes have long been recognized (e.g., Finley
1973:53ff, 99ff; Lenski 1966:219ff), and recent ethnographic and historical studies of peasant
classes, often assumed to be relatively egalitarian in wealth (Wolf 1%6), has revealed con-
siderable variation in household wealth levels (e.g., Lewis 1951; Mintz 1973; Berdan 1982).
It is even possible for some members of a lower class to be wealthier than some members of
a higher class when status or legal distinctions define classes (see Lenski (3966:770 or
Finley (1973:45) on legally defined classes or estates).
3 Two recent ethnoarchaeological investigations have yielded results indicating a very
weak association between household possessions and wealth (Hayden and Cannon
1984:129, passim; Kramer 1982:71-75). However, these findings are not particularly sur-
prising given the low level of wealth variability within the communities studied (as Hayden
and Cannon suggest (1984:202)). Furthermore, Kramer did not carry out an inventory anal-
ysis but only tallied the presence/absence of a few types of household goods.
4 A number of recent anthropological treatments of commodities, consumption, and
value have considerable relevance for the topic of this paper (e.g., Appadurai 1986; Belk
1983; Douglas and Isherwood 1979). Most directly useful is the latter work, which will be
discussed in the pages that follow.
5 It is important to use a genuine functional classification rather than try to go directly
from traditional archaeological typologies, based upon material and surface appearance, to
an analysis of wealth. Consider the absurdity of an ethnographer using a classification anal-
ogous to traditional ceramic typologies in an analysis of household wealth: the hypothetical
ethnographer would employ categories like glassware (including water glasses, mirrors, and
light bulbs), unpainted wood (bookshelves, children’s blocks, rolling pins), painted wood
(pencils, picture frames, piano benches), and composite plastic and metal objects (radios,
frying pans, and hair dryers). While some information on household wealth could be
gleaned from such a classification, it is clear that a priori functional categories are far supe-
rior. (Examples of archaeological studies that have attempted to measure wealth using
preexisting artifactual typologies include Smith (1975). Bawden (1982), and Kowalewski and
Finsten (1983).)

REFERENCES CITED
Abrams, Elliott M.
1984 Systemsof labor organization in Late Classic Copan, Honduras: The ener-
 HOUSEHOLD POSSESSIONS AND WEALTH 329

getics of construction. Ph.D. dissertation, Department of Anthropology, Penn-

sylvania State University.
Adams, Richard N.
1975 Energy and structure: A theory of social power. Univ. of Texas Press, Austin.
Appadurai, Arjun, editor
1986 Introduction: Commodities and the politics of value. In The social life ofthings:
Commodities in cultural perspective, pp. 3-63. Cambridge Univ. Press, New
York.
Arnold, Dean E.
1985 Ceramic theory and cultural process. Cambridge Univ. Press, New York.
Bartel, Brad
1983 A historical review of ethnographic and archaeological analyses of mortuary
practice. Journal of Anthropological Archaeology 1:32-X
Bawden, Garth
1982 Community organization reflected by the household: A study of Precolumbian
social dynamics. Journal of Field Archaeology 9165-182.
Belcher, John
1972 A cross-cultural household level-of-living scale. Rural Sociology 37:208-220.
Belcher, John C., and E. Sharp
1952 A short scale for measuring farm family level of living. Oklahoma Agricultural
Experimental Station Technical Bulletin, No. 46, Stillwater.
Belk, Russell
1983 Wordly possessions: Issues and criticisms. Advances in Consumer Research
l&514-519.
Berdan, Frances E
1982 The Aztecs of central Mexico: An imperial society. Holt Rinehart, New York.
Blake, Michael
1981 Detecting social organization with household architecture. Paper presented at
the 17th Mesa Redonda, Sociedad Mexicana de Antropologia, San Cristobal de
las Casas, Chiapas, Mexico.
Blanton, Richard E.
1984 Cross-cultural comparison of inequality and level of living among peasants.
Proposal submitted to the National Science Foundation.
Blanton, Richard E., and Gary Feinman
1984 The Mesoamerican world system. American Anthropologist 86:673-682.
Blaton, Richard E., S. A. Kowalewski, G. Feinman, and J. Appel
1981 Ancient Mesoamerica: A comparison of change in three regions. Cambridge
Univ. Press, New York.
Braudel, Femand
1981 The structures of everyday life: The limits of the possible. Harper & Row, New
York.
Braun, David P.
1983 Pots as tools. In Archaeological hammers and theories, edited by James E.
Moore and Arthur S. Keene, pp. 107-134. Academic Press, New York.
Brumtiel, Elizabeth M.
nd. Elite and utilitarian crafts in the Aztec state. In Production, exchange, and
complex societies, edited by Elizabeth M. Brumfiel and Timothy K. Earle.
Cambridge Univ. Press, New York, in press.
Cancian, Frank
1965 Economics and prestige in a Maya community: The religious cargo system in
Zinacantan. Standford Univ. Press, Stanford.
330 MICHAEL E. SMITH

Castro, Alfonso, N. Hakansson, and D. Brokensha

1981 Indicators of rural inequality. World Development 9401-427.
Chapin, F. Stuart
1935 Contemporary American institutions: A sociological analysis. Harper &
Brothers, New York.
Chapman, Dennis
1955 The home and social status. Routledge & Kegan Paul, London.
Chapman, Robert, Ian Kinnes, and Klavs Randsborg, editors
1981 The archaeology ofdeath. Cambridge Univ. Pres, New York.
Clark, Graham
1986 Symbols of excellence: Precious materials as expressions of status. Cambridge
Univ. Press, New York.
Cohen, Ronald
1978 Ethnicity: Problem and focus in anthropology. Annual Review of Anthropology
7:379-403.
Cook, Scott
1982 Zapotec stoneworkers: The dynamics of rural simple commodity production in
modern Mexican capitalism. University Press of America, Washington, DC.
Cowgill, George L., Jeffrey H. Altschul, and Rebecca S. Sload
1984 Spatial analysis of Teotihuacan, a Mesoamerican metropolis. In Zntrasite spatial
analysis in archaeology, edited by Harold J. Hietala, pp. 154- 195. Cambridge
Univ. Press, New York.
D’Altroy, Terence N., and Timothy K. Earle
1985 Staple finance, wealth finance, and storage in the Inka political economy. Cur-
rent Anthropology 26187-206.
Davis, Kingsley, and Wilbert E. Moore
1945 Some principles of stratification. American Sociological Review 10~242-249.
Deal, Michael
1983 Pottery ethnoarchaeology among the Tzeltal Maya. Ph.D. dissertation, Depart-
ment of Archaeology, Simon Fraser University.
DeWalt, Billie R.
1979 Modernization in a Mexican ejido. Cambridge Univ. Press, New York.
DeWalt, Kathleen M.
1983 Nutritional strategies and agricultural change in a Mexican community. UMI
Research Press, Ann Arbor.
Douglas, Mary, and Baron Isherwood
1979 The world of goods: Toward an anthropology of consumption. Basic Books,
New York.
Drennan, Robert D.
1983 Comment on Kowalewski and Finsten (1983). Current Anthropology 24:429-
431.
1984 Long-distance movement of goods in the Mesoamerican Formative and Classic.
American Antiquity 49~27-43.
Fallers, Lloyd A.
1973 Inequality: Social stratification reconsidered. Univ. of Chicago Press, Chicago.
Feinman, Gary M., Steadman Upham, and Kent G. Lightfoot
1981 The production step measure: An ordinal index of labor input in ceramic manu-
facture. American Antiquity 46871-885.
Finley, M. I.
1973 The ancient economy. Univ. of California Press, Berkeley.
Freedman, Maurice
1958 Lineage organization in southeastern China. Athlone, London.
 HOUSEHOLD POSSESSIONS AND WEALTH 331

Fried, Morton
1967 The evolution of political society: An essay in political anthropology. Random
House, New York.
Friedl, Emestine
1964 Lagging emulation in post-peasant society. American Anthropologist 66:569-
586.
Godelier, Maurice
1977 Perspectives in Marxist anthropology, translated by Robert Brain. Cambridge
Univ. Press, New York.
Griffiths, Dorothy M.
1978 Use-marks on historic ceramics: A preliminary study. Historical Archaeology
1268-81.
Gross, Daniel R.
1984 Time allocation: A tool for the study of cultural behavior. Annual Review of
Anthropology 13:519-558.
Guttman, Louis
1942 A revision of Chapin’s social status scale. American Sociological Review
7~362-369.
Haller, Archibald
1970 Changes in the structure of status systems. Rural Sociology 35:469-487.
Hally, David J.
1986 The identification of vessel function: A case study from northwest Georgia.
American Antiquity 51:267-295.
Haviland, William A.
1967 Stature at Tikal, Guatemala: Implications for Classic Maya demography and
social organization. American Antiquity 32:316-325.
1981 Dower houses and minor centers at Tikal, Guatemala: An investigation into the
identification of valid units in settlement hierarchies. In Lowland Maya settle-
ment patterns, edited by Wendy Ashmore, pp. 89-117. Univ. of New Mexico
Press, Albuquerque.
Hayden, Brian, and Aubrey Cannon
1983 Where the garbage goes: Refuse disposal in the Maya Highlands. Journal of
Anthropological Archaeology 2:117-163.
1984 The structure of material systems: Ethnoarchaeology in the Maya Highlands.
Society for American Archaeology, Papers, No. 3.
Helms, Mary W.
1979 Ancient Panama: Chiefs in search of power. Univ. of Texas Press, Austin.
Huss-Ashmore, Rebecca, Alan H. Goodman, and George J. Armelagos
1982 Nutritional inference from paleopathology. Advances in Archaeological Method
and Theory 5395-474.
Jones, Alice H.
1978 American colonial wealth: Documents and methods, 2nd ed., 3 Vols. Amo
Press, New York.
1980 Wealth of a nation to be: The American colonies on the eve of the revolution.
Columbia Univ. Press, New York.
Kay, Paul
1964 A Guttman scale model of Tahitian consumer behavior. Southwestern Journal
of Anthropology 20~160-167.
Kintigh, Keith W.
1984 Measuring archaeological diversity by comparison with simulated assemblages.
American Antiquity 49~44-54.
Klepinger, Linda L.
332 MICHAEL E. SMITH

1984 Nutritional assessment from bone. Annual Review of Anthropology 13:75-96.

Kottak, Conrad P.
1983 Assault on paradise: Social change in a Brazilian village. Random House, New
York.
KowaIewski, Stephen A., and Laura Finsten
1983 The economic systems of ancient Oaxaca: A regional perspective. Current An-
thropology 24:413-442.
Kramer, Carol
1982 Village ethnoarchaeology: Rural Iran in archaeological perspective. Academic
Press, New York.
Laumann, Edward O., and James S. House
1970 Living room styles and social attributes: The patterning of material artifacts in a
modem urban community. In The logic of social hierarchies, edited by Edward
0. Laumann and James S. House, pp. 189-203. Markham, Chicago.
Lenski, Gerhard E.
1966 Power and privilege: A theory of social stratification. McGraw-Hill, New
York.
Lewis, Oscar
1951 Life in a Mexican village: Tepoztlan restudied. Univ. of Illinois Press, Urbana.
1969 The Possessions of the poor. Scientific American 221(4): 114- 124.
1970 The Culture of the vecindad in Mexico City: Two case studies. In Anthropolog-
ical Essays, pp. 427-440. Random House, New York.
1973 The material culture of the poor: A case study of a Mexico City vecindad. In
Variation in anthropology, edited by Donald W. Lathrap and Jody Douglas, pp.
253-277. Illinois Archaeological Survey, Urbana.
Mack, Raymond W.
1951 Housing as an index of social class. Social Forces 29391-400.
Mandeville, Margaret
1974 A study of contemporary ceramic techniques at ‘Itda and a possible archaeolog-
ical application. In Studies of ancient Tollan: A report of the University of Mis-
souri Tula Archaeological Project, edited by Richard A. Diehl, pp. 122-129.
Department of Anthropology, University of Missouri-Columbia, Columbia.
McGuire, Randall H.
1982 The study of ethnicity in historical archaeology. Journal of Anthropological Ar-
chaeology 1:152-178.
1983 Breaking down cultural complexity: Inequality and heterogeneity. Advances in
Archaeological Method and Theory 6:91- 142.
Miller, G. L.
1980 Classification and economic scaling of 19th century ceramics. Historical Ar-
chaeology 14: l-41,
Millon, Rene
1981 Teotihuacan: City, state and civilization. In Handbook of Middle American In-
dians, Supplement No. 1, Archaeology, edited by Jeremy A. Sabloff, pp.
198-243. Univ. of Texas Press, Austin.
Mintz, Sidney
1973 A note on the definition of peasantries. Journal of Peasant Studies 1:91- 106.
Nelson, Ben A.
1981 Ethnoarchaeology and paleodemography: A test of ‘Ibmer and Lofgren’s hy-
pothesis. Journal of Anthropological Research 37:107- 129.
Netting, Robert McC.
 HOUSEHOLD POSSESSIONS AND WEALTH 333

1982 Some home truths on household size and wealth. American Behavioral Scien-
?is?s 25:641-662.
Netting, Robert McC., Richard R. Wilk, and Eric J. Amould, editors
1984 Households: Comparative and historical studies of the domestic group. Univ.
of California Press, Berkeley.
O’Laughlin, Bridgit
1975 Marxist approaches in anthropology. Annual Review of Anthropology 4:341-
370.
Otto, John Solomon
1977 Artifacts and status differences: A comparison of ceramics from planter, over-
seer and slave sites on an antebellum plantation. In Research Strategies in His-
torical Archaeology, edited by Stanley South, pp. 91-118. Academic Press,
New York.
1984 Cannon’s point plantation, 1794-1860: Living conditions and status patterns in
the Old South. Academic Press, New York.
Pelto, Pertti J.
1973 The snowmobile revolution: Technology and social change in the arctic. Cum-
mings, Menlo Park.
Phillips, David A., Jr., Ivan Restrepo, and William L. Rathje
1984 El Proyecto Basura. American Behavioral Scientist 28:139-153.
Polanyi, Karl
1968 Primitive, archaic and modern economices: Essays of Karl Polanyi. Beacon
Press, Boston.
Rathje, William L.
1983 To the salt of the earth: Some comments on household archaeology among the
Maya. In Prehistoric settlement patterns: Essays in honor of Gordon R. Willey,
edited by Evon 2. Vogt and Richard M. Leventhal, pp, 23-34. Univ. of New
Mexico Press, Albuquerque.
Rathje, William L., and Randall H. McGuire
1982 Rich men . . . poor men. American Behavioral Scientist 25:705-716.
Rathje, William L., and Michael B. Schiffer
1982 Archaeology. Harcourt Brace Jovanovich, New York.
Rice, Prudence M.
1983 Serpents and styles of Peten Postclassic pottery. American Anthropologist
85866-880.
Sanders, William T., and Robert S. Santley
1983 A tale of three cities: Energetics and urbanization in pre-hispanic central
Mexico. In Prehistoric settlement patterns: Essays in honor of Gordon R.
Willey, edited by Evon Z. Vogt and Richard Leventhal, pp. 243-291. Univ. of
New Mexico Press, Albuquerque.
Schiffer, Michael B .
1976 Behavioral archaeology. Academic Press, New York.
1983 Toward the identification of formation processes. American Antiquity 48:675-
706.
Schiffer, Michael B., Theodore E. Downing, and Michael McCarthy
1981 Waste not, want not: An ethnoarchaeological study of reuse in ‘Ibcson, Arizona.
In Modern material culture: The archaeology of us, edited by Richard Gould
and Michael B. Schiffer, pp. 67-86. Academic Press, New York.
Schneider, Harold K.
1974 Economic man: The anthropology of economics. The Free Press, New York.
334 MICHAEL E. SMITH

Sewell, Wiliam H.
1940 The construction and standardization of a scale for the measurement of the
socioeconomic status of Oklahoma farm families. Oklahoma Agriculturul Ex-
periment Station Technical Bulletin, No. 9, Stillwater.
Smith, Carol A.
1976 Exchange systems and the spatial distribution of elites: The organization of
stratification in agrarian societies. In Regional analysis, Vol. 2, Social Systems,
edited by Carol A. Smith, pp. 309-374. Academic Press, New York.
Smith, M. G.
1966 Pre-industrial stratification systems. In Social structure and mobility in eco-
nomic development, edited by Neil J. Smelser and Seymour M. Lipset, pp.
141- 176. Aldine, Chicago.
Smith, Marion F., Jr.
1985 Toward an economic interpretation of ceramics: Relating vessel size and shape
to use. In Decoding prehistoric ceramics, edited by Ben A. Nelson, pp.
254-329. Southern Illinois Univ. Press, Carbondale.
Smith, Michael E.
1975 Temples, residences and artifacts at Classic Teotihuacan. Unpublished senior
honors’ thesis, Department of Anthropology, Brandeis University.
1986 The role of social stratification in the Aztec empire: A view from the provinces.
American Anthropologist 88:70-91.
n.d. Economic and social organization in the provinces of the Aztec empire: Cuauh-
nahuac and Huaxtepec. To appear in Pattern and Process in Ancient Meso-
america, edited by Jeffrey R. Parsons, Richard A. Diehl, and Robert S. Santley
(volume in preparation).
Smith, Michael E., and Cynthia M. Heath-Smith
1980 Waves of influence in Postclassic Mesoamerica? A critique of the Mixteca-
Puebla concept. Anthropology 4(2): 15-50.
South, Stanley
1977 Method and theory in historical archaeology. Academic Press, New York.
Stone, Gary Wheeler, J. Glenn Little III, and Stephen Israel
1973 Ceramics from the John Hicks site, 1723-1743: The material culture. In Ce-
ramics in America, edited by Ian M. G. Quimby, pp. 103-109. The University
Press of Virginia, Charlottesville.
Teller, Barbera G.
1968 Ceramics in Providence, 1750-1800: An inventory survey. Antiques 94:570-
577.
Tourtellot, Gair
1983 An assessment of Clasic Maya household composition. In Prehistoric settle-
ment patterns: Essays in honor of Gordon R. Wiley, edited by Evon 2. Vogt and
Richard Leventhal, pp. 35-54. Univ. of New Mexico Press, Albuquerque.
Weber, Max
1946 From Max Weber: Essays in sociology, edited and translated by H. H. Gerth
and C. Wright Mills. Oxford Univ. Press, New York.
Wilk, Richard R.
1983 Little house in the jungle: The causes of variation in house size among modem
Kekchi Maya. Journal of Anthropological Archaeology 2:99- 116.
Wilk, Richard R., and William L. Rathje, editors
1982 Archaeology of the household: Building a prehistory of domestic life. American
Behavioral Scientist 25(6).
Wobst, Martin H.
 HOUSEHOLD POSSESSIONS AND WEALTH 335

1977 Stylistic behavior and information exchange. In Papers for the director: Re-
search essays in honor of James B. GrifIin, edited by Charles E. Cleland, pp.
317-342. University of Michigan, Museum of Anthropology, Anthropological
Papers, No. 61.
Wolf, Eric
1966 Peasants. Prentice-Hall, Englewood Cliffs.
Wright, Henry T.
1972 A consideration of interregional exchange in Greater Mesopotamia, 4000-3000
B.C. In Social exchange and interaction, edited by Edwin N. Wilmsen, pp.
9% 104. University of Michigan, Museum of Anthropology, Anthropological
Papers, No. 46.
Yanagisako, Sylvia
1979 Family and household: The analysis of domestic groups. Annual Review of An-
thropology 8:161-206.
Yang, Martin C.
1945 A Chinese village: Taitou, Chantung province. Columbia Univ. Press, New
York.