The Ecological-Evolutionary Typology of Human Societies and the Evolution of Social Inequality

6.2Polygyny

Because they are not inequality-conscious, traditional descriptions of marriage systems do not provide direct measures of inequality in the distribution of wives. The distinction between "occasional" and "common" polygyny is useless in a competition perspective, because it confounds situations that differ radically in the degree of inequality. The best one can do is to use existing information to classify societies as strictly monogamous (so that a man cannot legally have more than one wife) and those that are not, and thus in which women are potentially distributed in a highly skewed fashion. (I have counted the 4 polyandrous societies in the EA as monogamous.) Table 6.1 shows the distribution of monogamy/polygyny by type of society.
------ Table 6.1 ------
The marginals of Table 6.1 remind us of the fact that monogamy defined in the strict sense adopted here is practiced by only a minority of human societies (153 out of 857 in the EA, or 17.9%). Comparing the incidence of monogamy (absence of polygyny) across societies it appears that the AG type has by far the largest incidence (41.5%). Along the main sequence of types of society the high figure for AG represents a real jump relative to the immediately preceding type AH, where the incidence of monogamy is only 6.5%. Since the data set if rather large (the 41.5% figure for AG is supported by 135 cases) we can be reasonably sure that this instance of "AG reversal" is not an artifact. Thus, once again, it appears that for this dimension AG is associated with less inequality, and the nadir of inequality corresponds to AH, contrary to the monotonic pattern of increasing inequality from HG to AG originally postulated by Lenski.

6.3Maximum Harem Size

A better approximation to a measure of inequality in the distribution of women than the monogamy-polygyny dichotomy of Table 6.2 is an actual measure of maximum harem size (Alexander et al. 1979). Betzig (1986) has provided such a measure for a subset of cases in SS. (Her decision to exclude societies that could not be considered fully autonomous politically reduced observations from 186 to 77, for this variable.) Maximum harem size is defined as the number of conjugal relationships enjoyed by the man at the head of the social hierarchy (or, where there is no hierarchy, the most polygynous man). The codes categorize the number of wives into intervals as 1 = 3 or fewer, 2 = 4 to 10, 3 = 11-100, 4 = more than 100 (note that 8 cases fall into the last category).
------ Table 6.2 about here ------
Table 6.2 reveals clear differences among types of societies with respect to maximum harem size. Societies of the HG and SH types are concentrated in categories 1 and 2 at the low end of harem size. AH societies tend to lie at the other extreme with a plurality (45.5%) in category 4 of over 100 wives. By contrast AG societies exhibit a sharp decline in the degree of inequality, with a mode (61.5%) in category 1 (3 wives or fewer). FI and HE exhibit patterns not unlike that for HG, SH, or AG societies, but these percentages are based on very few cases. In any case these results suggest that the equality nadir is situated in AH rather than AG societies.

This clear instance of AG reversal in the trend of monotonically increasing inequality is unfortunately based on only 77 observations. While the chi-square is significant at p = .038, the large number of sparse cells renders this significance estimate suspect.²

7Sociocultural Evolution Revisited: Was There an AG Reversal?

7.1Evidence for the AG Reversal

------ Table 7.1 about here ------

Table 7.1 summarizes the empirical results. The pattern of the relationship between a dimension of inequality and type of society is classified as either Monotonic (the dimension increases monotonically over the sequence HG, SH, AH, AG) or AG Reversal (the dimension increases over the sequence HG, SH, AH, but then decreases from AH to AG). The table shows that there are 5 clear instances of AG reversals (excepting Games of Strategy, for which the decline AH -> AG is very small). One reversal corresponds to a measure of social rigidity (inheritance of headman's office), two concern clear measures of (absence) of freedom (presence of slavery and of despotism), and two correspond to the measures of sexual and reproductive inequality (polygyny allowed and maximum harem size). None of the measures of social complexity exhibits the AG reversal pattern.

Beyond the results presented in this study, one circumstance supporting the existence of the AG reversal for some dimensions of inequality may be the compatibility of this pattern with the traditional notion of "progress" with the emergence of "civilization" (which connotes the emergence of cities and is therefore identified with the AG type of society). It is compatible with the idea that, however harsh the punishment meted out in early agrarian codes of justice such as Hammurabi's, for example, these codes represented progress in freedom relative to earlier stages of sociocultural development.

The AG reversal pattern is also compatible with recent research showing that the fate of ordinary people in AG societies may not have been as bleak as previously supposed by Childe or Lenski (find ref. on Egyptian research, with letters of workers, etc; also comment in book on ancient near east about relative social fluidity in Egypt.)

7.2Implications of the AG Reversal

What would the existence of the AG reversal pattern for at least some important dimensions of social inequality mean for the understanding of broad trends of sociocultural evolution? There are empirical as well as theoretical implications.

7.2.1The Empirical Challenge

Among other dimensions, the AG reversal appears clearly in two dimensions related to sexual and reproductive inequality. Betzig's (1986) measure of maximum harem size is available only for a subset of the SS. An immediate empirical task would seem to be to obtain this, and perhaps other measures of sexual and reproductive inequality for a larger set of human societies, ideally for the entire EA set. Thus, an immediate project would be to code, for at least the whole SS and perhaps for the entire EA:

• 
  Maximum harem size (Betzig 1986); the advantage of this measure is that it is perhaps easiest to find in ethnographic sources; a drawback is that, like the range as a measure of dispersion, it may not be a robust statistic (Alexander et al. 1979)
  
• 
  Sufficient information on the distribution of wives among men to calculate a summary measure of inequality, such as the Gini coefficient
  
• 
  Sufficient information on the distribution of reproductive success (number of offspring) for men and women separately to estimate the effective rate of polygyny (calculated as the ratio of the variance in reproductive success for men to the variance in reproductive success for females; see Daly and Wilson 1983).
  

7.2.2The Theoretical Challenge

Confirmation of the AG reversal pattern would pose a major theoretical challenge while opening the door to new theoretical insights.

AG reversals challenge traditional explanations of trends in social inequality and human freedom. Lenski's historical scheme as proposed in Power and Privilege and Human Societies describes freedom for the average member of society as declining monotonically along the "main sequence" of types HG, SH, AH, AG, to reach a nadir in AG societies. Freedom for ordinary people does not begin to improve until past the onset of industrialization. The theoretical interpretation of this historical pattern is fairly straightforward. The monotonic decline in freedom is associated with the parallel increase in the size of the surplus made possible by technological advances amplifying the productivity of human labor. A greater surplus can be taken away from the producers to feed a more complex (and unequal) social structure. The reversal in the freedom trend during industrialization can then be attributed by mechanisms inherent to the "logic of industrialization". If the break in monotonicity occurred earlier during the agrarian era (or at the transition between AH and AG eras), industrialization can no longer account for the improvement in average human fate. Thus theoretical explanations for the upturn in freedom need to be at least adapted, and perhaps overhauled entirely. The new theoretical question becomes: What mechanisms operating in the agrarian system might motivate or force the elite to restrain their despotic tendencies?

The finding that the AG reversal takes place in the context of a decline in harem size, as well as the extension of strict monogamy, certainly suggests that there may be a connection between the upturn in human freedom and the watershed in sociocultural evolution represented by the rise of socially imposed monogamy (or socially imposed limitations on polygyny, as common today in the Moslem world). Thus the debate that has recently re-emerged in sociology about the origins of socially imposed monogamy (or limited polygyny) may be extremely relevant to the issue of the AG reversal (Kanazawa and Still 1999; see also MacDonald 2001; Sanderson 2001; Kanazawa 2001a, b). Theoretical questions to be posed in this context are: What were the mechanisms of the shift to socially imposed monogamy or limited polygyny? Are these mechanisms associated with the AG subsistence system, and if so how?