Table of Contents

Tables

Table 1 Summary of the sample of cultures
Table 2 Validity and reliability of private property measures
Table 3 Correlates of private ownership
Table 4 Summary of overlap in samples in five studies
Table 5 Summary of robust correlates of private ownership

Abstract

This is the fifth study of archived holocultural data in a program of cross-cultural research to identify the social institutions and behavioral norms that coincide with private property rights. Using Simmons' 1945 data base of 71 societies, a composite index of private rights for the aged, in conjunction with two other indices of private ownership, correlated (p<.0019) with 46 of Simmons' total 217 variables. Replications across all five studies show private ownership to be a positive correlate with agricultural subsistence, social stratification, social control based on law and religion, large populations and permanent settlements, patriarchal family norms, support of the aged, especially aged men, and economic practices of trade, money, debt, metallurgy, and war. Such results can be used to critique theories of property, for example, those of John Locke and George Mead.

Introduction

The ownership of property is one of the most enduring and widespread topics of social science discourse. From Greek antiquity to the present-day, across the disciplines of political science, anthropology, law, geography, history, sociology, and psychology, debate has focused on the foundation causes of private property, on its natural history and evolution, on its social functions and consequences, and on its justice (Rudmin, 1988; Rudmin, Belk & Furby, 1987; Schlatter, 1951). Much of this debate, beginning with Aristotle's 4th century B.C. survey of 158 circum-Mediterranian societies (Jaffa, 1963), has sought to muster cross-cultural evidence, especially from primitive or "natural" societies. Sometimes this evidence has been mythical, as in the Roman and early Christian accounts of communal ownership during the Golden Age or in the Garden of Eden (Avila, 1983; Schlatter, 1954; Tawney, 1926). Sometimes this has been hypothetical or surmised, as in Renaissance accounts by John of Paris, Fortescue, Hobbes, Locke and others of the origins of private property (Hobbes, 1655/1839; Locke, 1690/1952; Schlatter, 1951). Sometimes ethnographic evidence has been carefully selected to make or refute particular political-economic arguments, a common practice in the writings of such 19th century scholars as Morgan (1878), Marx (1972), Engels (1884/1920), Veblen (1899), and a host of comparative sociologists (Rudmin, 1988; 1995).

Hobhouse, Wheeler and Ginsberg in their 1915 book, The Material Culture and Social Institutions of the Simpler Peoples, challenged these naive and abusive uses of culture in political economic debates:

"Theories of social evolution are readily formed with the aid of some preconceived ideas and a few judiciously selected corroborative facts. The data offered to the theorist by the voluminous results of anthropological inquiry on the one hand, and by the immense record of the history of civilization on the other, are so vast and so various that it must be an unskilled selector who is unable, by giving prominence to the instances which agree and by ignoring those which conflict with his views, to make out a plausible case in support of some general notion of human progress. On the other hand, if theories are easily made, they are also easily confuted by a less friendly use of the same data." (Hobhouse, Wheeler & Ginsberg, 1915, p.1)

Hobhouse et al. argued that large numbers of cultures from around the world should be examined by statistical methods if claims about relationships between political economic variables are to be valid. Their own study coded categorical data from ethnographic and travelogue documents for 642 societies in order to classify societies into 7 stages of subsistence economy, from primitive hunters to advanced agriculturalists, so that graphic correlations might be made of stages of subsistence economy with other socio-economic variables, including private property. Hobhouse, Wheeler and Ginsberg (1915) were thus pioneers in applying sampling, quantification, and statistical inference to debates about private property.

Now, eight decades later, the present program of research seeks to continue their project by re-analyzing archived cross-cultural data bases in order to determine which social institutions and behavioral norms correlate with private ownership. Strict standards of social science methodology are adhered to, including independence of sampled societies, demonstrated validity and reliability of measures, appropriate statistical analyses, conservative significance criteria, and replication. To date, four data bases have been examined:

* Simmons' 1937 sample of 71 societies and 109 variables (Rudmin, 1992a),

* Swanson's 1960 sample of 50 societies and 39 variables (Rudmin, 1992b),

* Murdock's 1967 atlas of 862 societies and 44 variables (Rudmin, 1995), and

* Zelman's 1974 sample of 60 societies and 145 variables (Rudmin, 1996).

Replicated results from these studies will be presented and discussed later. McClelland's 1961 data base of 45 tribal societies has not been included in this list because his measure of individual ownership showed near-zero correlations with other established measures of private ownership and because his sample of cultures included many related societies (Rudmin, 1993). Correlations with private property based on McClelland's data, for example, in Textor's (1967) Cross-Cultural Summary, should be cited with caution.

The purpose of the present study is to examine the variables and data compiled in 1945 by Simmons as an expansion of his 1937 study. In his 1945 book, The Role of the Aged in Primitive Society, Simmons identified 56 variables pertaining to the situation of elderly people in society. Using ethnographic sources, he coded four-point ordinal data for these variables. Data were tabled separately for aged men and aged women. One of the new variables was "Property Rights Including Slaves". This measure of ownership practices is thus available to seek for new correlates of private ownership and to replicate the correlates already identified in the prior four studies.

This study is necessarily post-hoc and observational. Certainly, space is inadequate to review all property theories, and it would be a misrepresentation, in any case, to claim to test a priori hypotheses with archived data. Rather, the goal of this program of research is to establish an empirical "factual" base upon which to criticize property theory and upon which future research might rest with confidence.

Method

Archived Data

Simmons' 1945 study used the same sample of societies examined in his 1937 study. In Table 1, on the left, are the names of the societies as identified by Simmons (1945), along with the amount of data missing for the 109 variables tabulated in 1937 and for the 112 measures (56 variables each measured separately for men and women) tabulated in 1945. The letter "p" indicates that data are missing for the measures of private ownership, which in the 1937 data are variables 28 and 29 measuring ownership of land and of objects, and which in the 1945 data are variables 127 and 128 measuring ownership by aged men and by aged women.

Table 1 Summary of the sample of cultures

Legend:

c = corrected data from Ethnology (1967-1971).
d = deleted because redundant sampling of culture cluster.
e = Ethnology (1962-1971) is sole source of data.
p = private property measures have missing data. 

On the right side in Table 1 are the names and identification codes for the same 71 societies according to Murdock's (1967) Ethnographic Atlas. In Murdock's categorization scheme, the culture cluster code numbers and names identify societies that are related by common derivation or by prolonged and intimate contact. Holocultural samples should avoid multiple representations of the same culture clusters. Within each culture cluster, the specific society is identified by name on the far right in Table 1.

Murdock's reports for these societies are mostly found in his bound 1967 Ethnographic Atlas, which was based on 21 installments published in the journal Ethnology between 1962 and 1967, covering a total of 1,264 cultures. Murdock had encouraged the anthropological community to criticize and correct his codings, and these corrigenda were published in Ethnology as they became available. The bound 1967 Atlas reported on the 862 societies with the most complete data on 44 selected topics and included all corrections to date. However, between 1967 and 1971, corrections and completions of missing values continued to be published in Ethnology (Murdock & Spoehr, 1967; 1968a; 1968b; 1968c, Brudner, Murdock, Schorr, Spoehr & Tuden, 1971). In Table 1, a letter "c" in front of Murdock's cluster numbers indicates that Murdock's 1967 Atlas data were corrected based on subsequent reports in Ethnology. A letter "e" indicates that the society had been omitted from the 1967 Atlas and that data therefore came entirely from tabulations in Ethnology.

Identifying Societies

For the most part, Simmons' names for the societies in his sample were easily matched to corresponding names in Murdock's Atlas. However, for seven of Simmons' societies (Bushmen, Tuareg, Berbers, Mongols, Munda, Pomo, and Awawak), Simmons' names corresponded to the names of culture clusters comprised of several societies, and it was necessary to use arbitrary criteria to match Simmons' societies with specific cultures within these clusters. Simmons' 1945 monograph, unlike his earlier 1937 report upon which Rudmin's 1992 study was based, included an "Index of Tribes", a "Bibliography" of ethnographic source materials, and footnotes linking tribes to bibliographic sources. Thus, comparing Simmons' 1937 and 1945 reports, the latter provided more information upon which to match Simmons' societies to those tabulated in Murdock's Ethnographic Atlas.

Re-examining Rudmin's (1992a) earlier matching of Simmons' 1937 data and Murdock's data, it seems that Simmons' Bushmen had been correctly identified to be the Naron, according to the title of Bleek's (1928) ethnography cited by Simmons. The Berbers had been correctly identified to be the Shluh Berbers of the Grand Atlas Mountains in Morocco, according to Simmons' and Murdock's mutual citation of Ubach and Rackow (1923) as an ethnographic source. Simmons' Munda also seem to have been correctly matched to Murdock's Hill Bhuiya, since both data bases refer to the ethnographic work of Sarat Chandra Roy (1912; 1935), whose name was miscited by Simmons as C.R. Sarat and by Murdock as S.C. Ray. Finally, Simmons' Pomo had been correctly identified to be the Eastern Pomo, according to Simmons' and Murdock's mutual citation of Loeb (1926) as an ethnographic source.

However, some corrections of Rudmin's earlier (1992a) matching of Simmons' and Murdock's society listings are necessary. Simmons' Tuareg are not the Ahaggaren Tuareg as reported earlier, but the Asben Tuareg, according to Simmons' and Murdock's mutual citation of the ethnographic report of Lord Rodd (1926), who is also identified in bibliographic references as F.J. Rennell. Simmons' Mongols are not the Monguor located near the Tibetan border. According to Simmons' (1945) citations of Carruthers (1913), the data come from northwestern Mongolia and refer to the Altai Kalmucks, for whom Murdock (1967) has no corresponding data. Finally, Simmons' Arawak are not the interior Wapishana Arawak from southern Venezuela, but a coastal Arawak people. This is deduced from Menezes' (1977) report that one of Simmons' ethnographers, Roth (1908; 1916; 1922), studied Arawaks in Pomeroon near the coast. Thus, the best match for Simmons' Arawak would be the Locono from Murdock's Coastal Arawak cluster.

Comparing Rudmin's earlier (1992a) report and the present one, Simmons' data for the Tuareg and the Arawak are more correctly matched with Murdock's data in the present study, and Simmons' Mongols will join the Euahlayi in standing without any comparison group in Murdock's data base. These changes, of course, would have some effects on the results of Rudmin's 1992 analysis of Simmons' data. Of the correlations computed to compare Simmons' and Murdock's property measures for validity, the only two changes showed slightly stronger correlations when using the correctly matched data. Of the correlations computed to show the significant relationships between ownership and other societal practices, 24 became stronger, and only 8 moved in the null direction towards r=0.00. The largest change occurred with the correlation of Simmons' measure of "Use of Grain for Food" with Murdock's measure of land ownership, improving from r=.39 (n=48, p<.01) to a corrected value of r=.43 (n=47, p<.001). The only important consequence of these changes was the new inclusion of Simmons' variable 14, "Mining and Smelting of Metals", as a significant correlate of private ownership.

Final Sampling of Societies

As shown in Table 1 by the letter "d", three societies were deleted from Simmons' sample because they were redundant of sampled culture clusters: the Arunta, also known as the Aranda, were a duplicate sample of the Central Australia culture cluster, and the Point Barrow Eskimo and the Labrador Eskimo were duplicate samples of the Central and Eastern Eskimo culture cluster. As shown in Table 1, these deleted societies had more missing data than did the Dieri and the Polar Eskimo societies which were retained to represent their respective culture clusters.

Also shown in Table 1, five societies were effectively deleted because they had missing data for all of Simmons' measures of ownership. These were the Semang, the Euahlayi, the Chippewa, the Menomini, and the Yahgan. Thus, the effective sample for the present study consisted of 63 independent societies, only one of which, the Mongols, was without representation in Murdock's data base.

Property Measures

The index of private ownership developed for this study consisted of the disjunctive summation of Simmons' variables 127 and 128. These are measures of "Property Rights Including Slaves", coded separately for aged men and aged women. As with his other measures, Simmons used a four-point ordinal scale to summarize information from the ethnographic source materials:

0 = "absence or non-appearance of the trait, when definitely indicated in the sources";

1 = "incipient presence, slight elaboration, or cultural unimportance of the trait";

2 = "presence without dominance, moderate elaboration, or intermediate importance";

3 = "dominance, marked elaboration, or strong social importance of the trait".

The two gender-specified measures of private ownership, variables 127 and 128, showed slightly negative Kendall correlations with one another (r= -.07, n=30, p=.34). Hence, they could not be simply summed to create a composite index of private ownership. Rather, they were summed disjunctively. That is, the higher of the two values was taken as the measure of the degree to which a society allows at least some aged people to have private property. How property rights are distributed between aged men and aged women is less important for the present study than whether or not someone in the society has property rights. As shown in Table 2, the validity of this composite index of private property is demonstrated by its relatively strong positive correlations with indices of private ownership computed for other studies.

Table 2 Validity and reliability of private property measures;
Kendal correlations of variables and indices

Legend:

M = Mean correlation between variables in an index.
n.a.= not applicable.

As was done in earlier studies, conservative, conjunctive criteria were used to determine whether or not a variable was a statistically significant correlate of private ownership. This required that other indices of ownership be available. From Simmons' 1937 data base, his measures of private property in land and in objects were summed to produce a second composite index of private ownership. As shown in Table 2, the two measures contributing to this index were positively correlated, attesting to the reliability of the index, and the correlations with other measures of private property were all positive, attesting to the validity of the index. A similar summation of Murdock's (1967) measures of inheritance of land and of movables produced a third index which also showed positive correlations, displayed in Table 2, attesting to reliability and validity.

Thus, 63 independent societies and three composite indices of private ownership were available for computing correlation coefficients with the remaining 217 of Simmons' variables: 107 from his 1937 study and 110 from his 1945 study. These are all listed in Table 3, using Simmons' category headings, numeric codes, and descriptive labels.

Results

For a cultural norm or social institution to be identified as a correlate of private ownership in this study, the following conjunctive significance criteria needed to be satisfied: a) correlations of the same sign with all three available indices of private property, and b) at least two of the three correlations large enough to reach statistical significance (p<.05). The binomial probability of three correlation coefficients having the same sign is p=.25. The probability of two of three correlation coefficients reaching statistical significance is p=.05 X .05 = .0025, times the three ways of choosing two of three, which yields p=.0075. To this must be added the remote probability of all three coefficients being statistically significant, which yields p=.0075 + .000125 = .007625. Because the sign of a correlation is independent of its magnitude, the statistical significance level using the two conjunctive criteria is p=.25 X .007625, which is p<.0019. With 217 variables to be examined for correlation with private ownership in this study, and with a significance criterion of p<.0019, it is unlikely that even one variable would be falsely identified as a correlate of private ownership due to random patterns in the data.

Because the data were ordinal, the Kendall correlation was preferred to the Pearson correlation. Because correlation coefficients for the three indices of ownership were expected to have the same sign, one-tailed probability estimates were used. The significant correlates of private ownership are noted in Table 3 by the mark # on the left. Of the 217 measures examined, 46 met the conjunctive significance criteria employed here. Included among these were 20 of the 22 correlates of private ownership identified in the earlier study (Rudmin, 1992a) of Simmons' 1937 data. The two exceptions were variable 65, "Marriage by Capture", and variable 93, "Exposure of the Dying." These each produced correlations of the appropriate sign, but each had only one of three correlations achieve significance.

It should also be noted that 30 of the 110 variables defined by Simmons for his 1945 study had fewer than 10 societies available for computing correlations with the composite measures of private ownership. Thus, many of Simmons' variables have not had the opportunity here to be properly tested as correlates of private ownership.

Table 3 Correlates of private ownership

Legend: Kendall Correlations # p<.0019; * p<.05

Replication

In the social sciences, every study has its faults and every researcher has his or her biases. Thus, confidence in a claim can only arise from replication, preferably by studies designed with different intentions in mind, with different scholars collecting the data from a variety of ethnographic sources, with different operational definitions for the variables, and with different samples of societies. The present program of research has been very much directed to seeking to establish confidence through replication.

For each of the five cross-cultural studies to date, the variables were defined and the data were collected for a number of different purposes, none of which were to examine property theory per se. Simmons' 1937 study was designed and directed to confirm by quantitative methods the conclusions reached by his mentors Sumner, Keller, and Davie in their 1927 non-quantitative cross-cultural study. Simmons' 1945 study had a very different purpose, to discover if older people in the world's societies had common problems and opportunities in relation to various social norms and institutions. The purpose of Swanson's 1960 study was to explore the functions of beliefs in high gods. Murdock's 1967 Atlas was intended to serve as a general ethnographic data base. Zelman's 1974 cross-cultural study was designed to discover gender roles and power relationships relative to reproduction.

The data for the five studies to date were collected by four different scholars. In the case of Simmons collecting data for two of the studies, there was almost a decade separation between his two projects, and the data took different forms, i.e., in his first study, his codes characterized societies, and in his second study, his codes characterized genders within societies. Thus, any biases in these five studies caused by the authors' intentions should be different for the different authors and studies.

For each of the five cross-cultural studies to date, a different operational definition of private ownership was used. These have already been defined, discussed, and displayed in Table 2, on the left. Simmons' 1937 index measure of ownership was created by summing his 1937 measures of property in land and in objects. Simmons' 1945 index measure of ownership equaled his measure of ownership by aged men or by aged women, which ever was greater. Zelman's index of ownership was created by averaging a measure of land ownership with a measure of land inheritance, and then summing that with a measure of ownership of animals. In some instances, the correlates of private ownership have also been measured by two or more different operational definitions. For example, Simmons (1937), Swanson (1960/1966) and Murdock (1967) each have a measure of how much a society gets its food from agriculture. In every instance, each of these three measures correlated positively with each of the five indices of private ownership. The demonstration of conceptual replication should be cause for further confidence.

Table 4 Summary of overlap in samples in five studies

Legend:

N = Number of societies in the sample.
n = Number of sampled societies with property measures.
( ) = Number of societies shared in common by two samples.
Grand N = Number of societies used in all studies to date.

To date, a total of five different samples of societies have been examined. Although the two studies by Simmons used the same sample, the study of Murdock's data employed two mutually exclusive samples. A summary comparison of the different samples of societies used in these studies appears in Table 4. In total, 510 societies have been examined. The numbers in brackets refer to societies common to any two intersecting studies. Murdock's and Zelman's data bases have most overlap, and Simmons' and Swanson's data bases have least overlap. Of the five samples of societies examined to date, a total of 416 societies have appeared uniquely in only one sample, 70 societies have appeared in two samples, 22 societies have appeared in three samples, and only two societies (the Nama and the Iban) have appeared in four samples. No societies have appeared in all five samples. Thus, if only the samples of societies are considered, these studies are not completely independent replications of one another. However, they are independent replications to the degree that the data were collected for different purposes, by different authors, using different operational definitions.

Table 5 Summary of robust correlates of private ownership

Legend:

Source of Variables: s=Simmons, w=Swanson, m=Murdock, z=Zelman

Correlations:
C=(positive correlation) and (p<.05). 
c=(positive correlation) but (p>.05).
? = (n<10) and (indeterminate).

Table 5 is a summary display of the robust correlates of private ownership identified in the five studies to date. Robust here means that the expected correlation is evident in every instance it which it is possible to compute a correlation, regardless of who collected the data, for whatever purposes, regardless of how private ownership was operationalized, and regardless of how small and unrepresentative the set of societies available for the computation. Thus, Table 5 may omit variables that might be true correlates of private ownership, but which are not robust.

In Table 5, the data bases for the five studies to date are displayed across the top, along with the number of societies with property measures in each study. For a variable to be listed in Table 5, its correlation with a measure of private ownership must have met strict significance criteria in at least one study. The replicated, robust correlates of private ownership are listed on the left. They are grouped into seven categories and ordered within each category from the most replicated to the least. In order to simplify the table by displaying only positive correlations, each negative correlate of private ownership had "NOT" inserted into its label in order to reverse its meaning and thus transform the negative correlation to a positive one. In other words, Table 5 indicates those social norms and institutions that tend to be present when private ownership is present and absent when private ownership is absent.

An upper case "C" indicates that the correlation was positive and was statistically significant at p<.05. A lower case "c" indicates that the correlation was positive but was not of sufficient magnitude to achieve statistical significance. In the case of Murdock's index of ownership correlating with one of his own measures, an upper case "C" was used only when the correlation coefficients were significant in both samples drawn from Murdock's Atlas, and a lower case "c" was used in the one instance of "Dual Class Stratification" in which the correlation was significant in one sample (r=.22, n=296, p<.001) of Murdock's data but not in the other (r=.13, n=145, p=.06). A question mark (?) indicates that fewer than ten societies were available for the computation and that the results were indeterminate, either because r=0.00 or because one of the measures was invariant. These symbols all refer to Kendall correlation values.

To read Table 5, consider the first entry, "Agriculture." Simmons ("s"), Swanson ("w"), and Murdock ("m") had each defined and coded data for a measure of the degree to which a society engaged in agriculture. A composite measure of ownership of land and objects from Simmons' 1937 study had significant (p<.05), positive Kendall correlations with all three measures of agriculture. Similarly, with Simmons' 1945 data, a composite measure of ownership by aged men or aged women had a significant, positive correlation with all three measures of agriculture. Swanson's 1960 measure of "Ownership of Economically Significant Property" had a positive correlation with all three measures of agriculture, but did not achieve statistical significance with Simmons' measure of agriculture (r=.15, n=8, p=.33). Murdock's composite measure of the inheritance of land and of movables had significant, positive correlations with Simmons' and Swanson's measures of agriculture. Furthermore, Murdock's measure of ownership had significant, positive correlations with his own measure of agriculture for both of the samples drawn from his data (r=.46, n=147, p<.001 and r=.44, n=312, p<.001). Finally, Zelman's composite measure of the ownership and inheritance of land and the ownership of domestic animals had a positive correlation with all three measures of agriculture, but not of sufficient magnitude to achieve significance with Simmons' measures of agriculture (r=.30, n=19, p=.08) nor with Swanson's measure (r=.29, n=12, p=.14). In those instances where statistical significance was not demonstrated, there were simply too few societies available to make statistical significance a reasonable expectation.

Even disregarding the magnitudes of the correlation coefficients and their levels of statistical significance, summary statistical probabilities can be computed from just the signs of the correlations. If the 16 computed correlations between private ownership and agriculture are presumed to be independent of one another, then the binomial probability of achieving 16 positive correlations due to random patterns in the data is p=.0000153. Thus, it is nearly certain that agricultural food sources tend to coincide with private ownership.

Independence means that each of the correlations is free to be either positive or negative, that the sign of one correlation is not constrained by the other correlations. Independence seems a correct presumption since the property measures have been uniquely defined and the data have been uniquely collected, or in the case of Murdock's data, the two samples are exclusive, having no societies in common. However, the presumption of independence might be challenged because the samples for the other studies do have some societies in common. Two societies, the Nama and the Iban, appear in four of the five samples. To examine the effects of non-independent samples on conceptual replication, the Nama and the Iban can be assigned to the Swanson study, which has the smallest sample, and can be deleted from the other studies. A recomputation of the correlations showed that in no instance did the deletion of the Nama and the Iban cause the signs of the correlations reported in Table 5 to change. Thus, the overlap of samples seems not to be responsible for the conceptual replications.

Discussion

The social institutions and behavioral norms listed in Table 5 might be conceived to be "objective observations" of the societal characteristics that tend to co-exist with private property. Each correlation is an "observation", something "seen" by the "observer" who defined the measures, collected the data, and organized them for the computation of correlations. These observations are "objective" in the sense that several observers, whether they know it or not, have agreed that they see the same thing. This agreement exists even though each observer has unique interests and none of them was even "looking" to see property relations.

Thus, Table 5 presents some "facts" about private ownership, to the extent that social science can observe "facts" when casting its eyes around the globe and across numerous cultures. Although the observations have been post hoc, opportunistic, and atheoretical, theories about property should be compatible with these facts where they pertain. Empirical tests of theories need not always be a priori.

For example, John Locke presented his labor theory of property in his Second Treatise on Government in 1690. He argued that it is self-evident that humans naturally own their own bodies, and, therefore, they own their bodies' labor and whatever gets mixed with that labor. Locke imagined that the derivation of ownership from labor was most clearly evident in the hunting and gathering of food by North American native peoples:

"The fruit or venison which nourishes the wild Indian, who knows no enclosure and is still a tenant in common, must be his, and so his, i.e., a part of him. . . He that is nourished by the acorns he picked up under an oak, or the apples he gathered from the trees in the wood, has certainly appropriated them to himself. Nobody can deny but the nourishment is his. . . Thus this law of reason makes the deer that Indian's who has killed it; it is allowed to be his goods who has bestowed his labor upon it, though before it was the common right of every one." (Locke, 1690/1952,pp. 17-19.)

The cross-cultural observations reported in Table 5, however, do not support Locke. Where people do in fact gather acorns and apples, where they do hunt venison, in explicitly those conditions, private ownership tends not to be the norm. To the contrary, hunting and gathering peoples tend not to have private ownership of land or of goods. Typically, they secure their sustenance, not by private rights and not by means of exclusive access to resources, but by rights of sharing and by means of intra-communal and inter-communal access to resources (Gould, 1982; Hayden, 1992a; 1992b; 1993).

Locke's labor theory is one of the most influential and widely discussed theories of property. For example, the 1997 Social Sciences Citation Index (Anon., 1998) shows that his Second Treatise on Government was cited almost 100 times that year. Nevertheless, Locke is wrong. He himself was an ardent empiricist, and it seems likely that he would have acquiesced to the "facts" had he known them. That his theory seemed reasonable at the time, and to subsequent generations of readers, probably relates to the need to justify unequal distributions of property or to the need for whole populations to internalize motivations for work, both of which were issues of the Protestant movements of Locke's day (Aaron, 1972; Schama, 1987; Tawney, 1926; Weber, 1930).

Locke's labor theory might best be salvaged by shifting the focus away from physical labor per se to the expectations that motivate labor. As argued by Bentham and Litwinski (Rudmin, 1990), possession is essentially cognitive and only incidentally physical. Humans naturally have prospective imaginations and acquire and secure objects, information, and interpersonal relationships for their expected future utility. Possessions can be acquired in many ways, including labor, discovery, inheritance, trading, gift, and theft, but in Locke's mental experiment of trying to describe first acts of ownership by an isolated person in primitive conditions, labor seems the only way. The "facts" show, however, that hunting and gathering are not the ideal conditions for predicting the appearance of private property, possibly because the windfall features of these activities, and the relatively short shelf-life of the food acquired, make them not amenable to prospective imagination and norms of possession. In fact, time studies of hunting-gathering peoples show that they labor very little and live a leisurely, though materially simple, lifestyle (Lee, 1968).

In comparison to hunting and gathering, grain agriculture requires permanent settlement near maintained and protected agricultural land, requires seed grain, draft animals, and work implements to be secured for future use, requires intensive, organized labor, and requires that all of this be motivated by reasonable expectations of a harvest that can be stored for future consumption. Agriculture, in comparison to hunting and gathering, yields surpluses that can support large populations, including relatively non-productive groups such as rulers, warriors, priests, and the aged. Private ownership would be predictable in an agricultural society.

This is but one example of how a foundation of facts about private ownership might be used to eliminate incorrect theories. In this case, it was fortuitous that Locke had cast his theory into the context of hunting and gathering, and that cross-cultural data on property norms in hunting and gathering societies are abundantly available. A lesser known theory of property is George Mead's (1982) explanation that one way for the organic community to adjust to strangers and unknown personalities is to adopt abstract, rule-governed norms, such as private ownership. However, one of the robust, positive correlates shown in Table 5 is "NOT Outgroup Intimacy". Thus, Mead is exactly wrong: private property tends to be absent where there is intimate contact with outgroup personalities. Again, it is fortuitous that data were available to test his claim.

The present program of research will continue to examine existing, archived, cross-cultural data bases that have useful measures. However, after those resources have been exhausted, new studies will be commissioned to define variables and to collect new data explicitly to test property theory.

References

Aaron, R.I. (1972). "Locke", in Encyclopaedia Britannica, Vol. 14, pp. 188-193, William Benton, Chicago.

Anon. (1998). Social Sciences Citation Index: 1997, Annual, Part 2, Institute for Scientific Information, Philadelphia.

Avila, c (1983). Ownership: Early Christian Teaching, Sheed and Ward, London.

Bleek, D.F. (1928). The Naron: A Bushman Tribe of the Central Kalahari, Cambridge University Press, Cambridge.

Brudner, L.A., Murdock, G.P., Schoor, T., Spoehr, A. and Tuden, A. (1971). "Ethnographic atlas", Ethnology 10, 122-127.

Carruthers, A.D. (1913). Unknown Mongolia: A Record of Travel and Exploration in Northwest Mongolia and Dzungaaria, Hutchinson, London.

Engels, F. (1920). "The origin of the family, private property, and the state", in The Marx-Engels Reader, 2nd ed., (R.c Tucker, Ed.), pp. 734-759, W.W. Norton, New York. (Originally published in 1884.)

Gould, R.A. (1982). "To have and have not: The ecology of sharing among hunter-gatherers", in Resource Managers: North American and Australian Hunter-Gatherers (N.M. Williams and E.S. Hunn, Eds.), pp. 69-91, Westview Press, Boulder, CO.

Hayden, B. (1992a). "Introduction: Ecology and culture", in A Complex Culture of the British Columbia Plateau (B. Hayden, Ed.), pp. 1-46, University of British Columbia Press, Vancouver.

Hayden, B. (1992b). "Conclusions: Ecology and complex hunter/gatherers", in A Complex Culture of the British Columbia Plateau (B. Hayden, Ed.), pp. 525-563, University of British Columbia Press, Vancouver.

Hayden, B. (1993). Archeology: The Science of Once and Future Things, W.H. Freeman, New York.

Hobbes, T. (1839). "De Corpore Politico, or the Elements of Law", in The English Works of Thomas Hobbes, vol. 4, (W. Molesworth, Ed.), pp. 77-228, John Bohm, London. (Originally published in 1655.)

Hobhouse, L. T., Wheeler, G. and Ginsberg, M. (1915). The Material Culture and Social Institutions of the Simpler Peoples: An Essay in Correlation, Chapman and Hall, London.

Jaffa, H.V. (1963). "Aristotle", in History of Political Philosophy (L. Strauss and J. Cropsey, Eds.), pp. 64-123, Rand McNally, Chicago.

Lee, R.B. (1968). "What hunters do for a living, or, how to make out on scarce resources", in Man the Hunter (R.B. Lee and I. DeVore, Eds.), pp. 30-48, Aldine, Chicago.

Locke, J. (1952). The Second Treatise on Government, Bobbs- Merrill, New York. (Originally published in 1690.)

Loeb, E.M. (1926). "Pomo Folkways", University of California Publications in American Archaeology and Ethnology 24, 149- 409.

Marx, K. (1972). The Ethnological Notebooks of Karl Marx, (L. Krader, Ed. and Trans.), Van Gorgum, Assen.

McClelland, D.c (1961). The Achieving Society, Van Nostrand, New York.

Mead, G.H. (1982). The Individual and the Social Self: The Unpublished Work of George Herbert Mead (D.L. Miller, Ed.), University of Chicago Press, Chicago.

Menezes, M.N. (1977). British Policies Toward the Amerindians in British Guiana, 1803-1823, Clarendon, Oxford.

Morgan, L. (1878). Ancient Society, Holt, New York.

Murdock, G.P. (1967). Ethnographic Atlas, University of Pittsburg Press, Pittsburg.

Murdock, G.P. (1981). Atlas of World Cultures, University of Pittsburg Press, Pittsburg.

Murdock, G.P. and Spohr, A. (1967). "Ethnographic atlas XXXIII", Ethnology 6, 481-487.

Murdock, G.P. and Spohr, A. (1968a). "Ethnographic atlas XXXIV", Ethnology 7, 106-111.

Murdock, G.P. and Spohr, A. (1968b). "Ethnographic atlas XXXV", Ethnology 7, 218-224.

Murdock, G.P. and Spohr, A. (1968c). "Ethnographic atlas XXXVI", Ethnology 7, 327-329.

Rennell, F.J., alias Rodd, F.R. (1926). People of the Veil, Clowes, London.

Roth, W.E. (1908). "An inquiry into the animism and folk-lore of the Guiana Indians", Annual Reports of the Bureau of American Ethnology 30, 103-386.

Roth, W.E. (1916). "An introductory study of the arts, crafts, and customs of the Guiana Indians", Annual Reports of the Bureau of American Ethnology 38, 23-745.

Roth, W.E. (Ed. & Trans.) (1922-1923). Richard Schomburgk's Travels in British Guiana, Daily Chronicle, Georgetown.

Rudmin, F.W. (1988). "Dominance, social control, and ownership: A history and a cross-cultural study of motivations for private property", Behavior Science Research 22, 130-160. [Note: The first two galley pages were reversed during printing. The paper should begin on paragraph 5 on page 132. Paragraphs 1 through 4 should follow paragraph 8 on page 133.]

Rudmin, F.W. (1990). "The economic psychology of Leon Litwinski (1887-1969): A program of cognitive research on possession and property". Journal of Economic Psychology 11, 307-339.

Rudmin, F.W. (1992a). "Cross-cultural correlates of ownership of private property", Social Science Research 21, 57-83.

Rudmin, F.W. (1992b). "Cross-cultural correlates of ownership of private property: A look from another data base", Anthroplogica 34, 71-88.

Rudmin, F.W. (1993). "Ownership of economic property: McClelland's 1961 cross-cultural data", Psychological Reports 73, 794.

Rudmin, F.W. (1995). "Cross-cultural correlates of the ownership of private property: Two samples of Murdock's data", Journal of Socio-Economics 24, 345-373.

Rudmin, F.W. (1996). "Cross-cultural correlates of the ownership of private property: Zelman's gender data revisited", Cross-Cultural Research 30, 115-153.

Rudmin, F.W., Belk, R.W. and Furby, L. (1987). Social science Bibliography on Property, Ownership, and Possession: 1580 Citations from Psychology, Anthropology, Sociology and Related Disciplines, Vance, Monticello, IL.

Roy, S.c (1912). The Mundas, Kuntaline Press, Calcutta.

Roy, S.C. (1935). The Hill Bhuiyas of Orissa, Man in India Office, Ranchi.

Schama, S. (1987). The Embarrassment of Riches, Collins, London.

Schlatter, R. (1951). Private Property: The History of an Idea, Rutgers University Press, New Brunswick, NJ.

Simmons, L.W. (1937). "Statistical correlations in the science of society", in Studies in the Science of Society (G.P. Murdock, Ed.), pp. 493-517, Yale University Press, New Haven.

Simmons, L.W. (1945). The Role of the Aged in Primitive Society, Yale University Press, New Haven.

Sumner, W.G., Keller, A.G. and Davie, M.R. (1927). The Science of Society, Yale University Press, New Haven.

Swanson, G.E. (1966). The Birth of the Gods: The Origin of Primitive Beliefs, 2nd printing, University of Michigan Press, Ann Arbor. (Originally published in 1960.)

Tawney, R.H. (1926). Religion and the Rise of Capitalism: A Historical Study, Harcourt Brace, New York.

Textor, R.B. (1967). A Cross-Cultural Summary, Human Relations Area Files Press, New Haven, Ct.

Ubach, E. and Rackow, E. (1923). Sitte und Recht in Nordafrika, Enke, Stuttgart.

Veblen, T. (1899). Theory of the Leisure Class, Macmillan, New York.

Weber, M. (1930). The Protestant Ethic and the Spirit of Capitalism, G. Allen and Unwin, London.

Zelman, E.A. (1974). Women's Rights and Women's Rites: A Cross-Cultural Study of Womanpower and Reproductive Ritual, doctoral dissertation, University of Michigan, Ann Arbor.