Why There Are So Few Women Warriors

Why There Are So Few Women Warriors

by David Adams

Behavior Science ResearchVolume 18, Number 3, 1983, Pages 196-212

SUMMARY

Cross-cultural methodology suggests that women are excluded from warfare not so much because of sex differences in aggressiveness or strength, but instead because of a contradiction arising from marital residency systems that arose, in turn, as a function of warfare. Under conditions of internal warfare war against neighboring communities sharing the same language many stateless cultures may have adopted patrilocal exogamous marital residency (the bride comes from a different community and comes to live with the family of the husband). Under these conditions the wife is faced with contradictory loyalties during warfare, because her husband may go to war against her brothers and father. It appears that women have been excluded historically from warfare in order to resolve this contradiction and protect the security of the warrior husbands. This explanation is supported by other findings that women do fight as warriors in certain cultures in which warfare or marital residency rules are structured in such a way that the contradiction does not arise.

1. Introduction

Why don't women take an active part in war? Is it because they lack a "so-called aggressive instinct" (Lorenz, 1966) present in men and responsible for war? These are important questions while it is imperative for us to abolish war within our lifetime, some people have argued that war cannot be abolished because it is caused by "aggressive instincts." In previous work on the brain mechanisms of aggressive behavior (Adams, 1979), I found no evidence that "aggressive instincts" could explain warfare. Therefore, I have turned to cross-cultural methodology to address the question. The results, presented in the following paper, support a completely different explanation that women do not go to war because there is an historical contradiction between the institutions of warfare and marriage .

Of course, all complex human behavior is at the same time both biological and cultural, both individual and social. On the one hand, the present paper emphasizes that historical contradiction between marriage and warfare is the immediate determining factor in the exclusion of women from war. On the other hand, there is no reason to deny that biological factors have also played a role in the historical process, as will be discussed later on.

Cross-cultural anthropological research by previous investigators, especially Carol and Melvin Ember (1971, 1974) have shown that there is a strong relationship between patterns of marital residency and patterns of warfare. In particular, it has been shown that there is an association between patrilocal marital residency (in which the bride goes to live with or near the husband's family) and the presence of some internal warfare (warfare against neighboring communities that share the same language), while matrilocal marital residency (in which the husband goes to live with or near the bride's family) is more likely to be found in cultures with exclusively external warfare (warfare against more distant communities that do not share the same language).

In cultures with patrilocal marital residency and internal warfare, there is a potential conflict of interest for women: war may be waged by their husbands on one side against their brothers and fathers on the other side. In reading various accounts of warfare in cultures without a state structure, I have been struck by how often one encounters this problem of split loyalties for women. For example, among the Mae Enga, a New Guinea people with frequent internal warfare and a patrilocal system of marital residency, Meggitt (1977 , p. 98) reports:

Mae women are wholly excluded from the meetings in which men decide whether or not to go to war. ...This restriction prevents women from learning the details of battle plans, and it is meant to do so. In this way a woman married in from another clan (and almost all are) is not faced with a difficult conflict or interests, if, as may well be the case, her natal group is the intended target .

Given the frequency of the preceding predicament (in the sample to be described later, it would be expected to occur in 7 I percent of the 58 cultures with frequent internal warfare), it seemed that it might be a major contributing factor to the exclusion of women from warfare. Therefore, I put it to a test by using classical cross-cultural survey techniques.

2. Cross-Cultural Methodology

Rather than drawing up an entirely new sample of cultures, the sample for this study was drawn from the sum of the samples used by nine previous investigators who have used cross-cultural techniques to examine the relation of warfare and marriage.1 All of the 204 cultures cited by those investigators were used providing they met the following criteria: (1) they had to be included in the Ethnographic Atlas (Murdock, 1967) and coded there for type of marriage and political complexity (all but 10 societies met this criterion) (2) they had to be stateless cultures, coded as 0 or 1 on column 33 of Murdock, since it was assumed that state systems might involve other complicating factors (44 cultures were excluded on these grounds) (3) when more than one source was available, the data had to be consistent in al1 of them or else the culture was excluded {the following six were excluded because of such contradictory information: Aymara, Chukchee, Cuna, Siwans, South Ute, White Knife) and {4) data had to be available on both the frequency and type of warfare (internal, external), as well as feuding (see below). Warfare was judged to have a low frequency if war and feuding were coded by previous authors as "rare," "infrequent," 'never ," or "absent," or if there was no mention of warfare despite considerable evidence about related aspects of the culture including economic structure or relations with neighboring cultures and communities. There were 29 societies excluded because of lack of adequate information about frequency or type of warfare. All told, there were 115 cultures that met all of the criteria and were included in the sample.

In the previous literature two different definitions have been used for warfare. A broad definition has been used by the Embers: "fighting that involves two or more territorial units so long as there is a group of fighters on at least one side. " A narrower definition has been used by Divale, Tefft, and Otterbein: for Divale and Tefft, warfare is "armed aggression between political communities or alliances of political communities" for Otterbein, it is 'an armed contest between two independent poIitical units, by means of organized military force, in the pursuit of a tribal or national policy." Since the narrower definitions insist upon the political community as the unit of war, they exclude feuding. In the present study the broader definition, including feuding, will be used because it would be expected to relate to marriage residency rules in the same way as other internal warfare. Therefore, in order to adapt the findings of Divale, Tefft, Otterbein and Horton to the present study, it was necessary to check for feuding in the cultures they listed and to reclassify four cultures in which there was frequent feuding and include them among cultures with internal warfare (Copper Eskimo, Monachi, Maria Gond and Papago).

The determination of whether or not women take part as warriors was made on a subsample. The subsample consisted of those cultures that: (1) had frequent or occasional warfare, (2) were surveyed in the Human Relations Area Files, and (3) had sufficient material in the Area Files so that it could be determined if women fought as warriors. Of the 83 cultures with frequent or occasional warfare, 71 were found in the Human Relations Area Files and all but four could be coded for female participation in warfare (the exceptions were the Lesu, Maria Gond, Copper Eskimo, and Yavapai for which there were insufficient data). All of the material in the Human Relations Area File was read by the author and searched for descriptions of women taking part in warfare within the following sections of the files: 628, inter-community relations 669, revolution 70-708, armed forces and 72-729, war (Murdock et al., 1967).

3. Societies with Women Warriors

In nine of the 67 cultures in the subsample it was found that women take part in war, at least occasionally, as active warriors (Tables 1 and 2). All nine were among the 33 cultures characterized by exclusive external warfare and/or community endogamy (i.e., marriage partners come from the same community). In contrast, there was not a single case of a culture in which women are allowed to fight as warriors among the other 34 cultures characterized by internal warfare and some community exogamy (i.e., at least some marriage partners come from different communities). These data, significant at a .01 probability level by Fisher's Exact Test, support the hypothesis that women are excluded from warfare at least in part because of the problem of split loyalties derived from the contradiction between marriage and warfare systems.

Women warriors have been described in five of the sampled cultures with exclusive external warfare. Among the Crow, "there are memories of a woman who went to war. ..indeed one of my woman informants claimed, to have struck a coup" (Lowie, 1935, p. 215). Among the Navaho "a woman, if she wished, might join a war party. There were never more than two women in a party. They fought just as did the men " (Hill, 1936, p. 3). Among the Gros Ventre, "women not infrequently went with the war parties" (Kroeber, 1908, p. 192). Among the Fox women could go to war with their husbands and it has been reported that "even some women have become warrior women" (Michelson, 1937, p. 11). Among the Delaware, although "the women seldom go into battle, they have however a right to fight, which entitles them to a place in the ranks" (Kinietz, 1946, p. 132).

Women warriors have also been described in four cultures with exclusive local or community endogamy. Among the Comanche women sometimes would "snipe with bows and arrows from fringes of the fray" (Wallace and Hoebel, 1952, p. 253). Among the Majuro of the Marshall Islands, "women take part in war, not only when they have to defend the home ground from the enemy, but also in attacks, and although in the minority, they form part of the squadron. ..(throwing) stones with their bare hands" (Erdland, 1914, p. 93). Among the Maori, women occasionally took part in the fighting and also accompanied raiding parties (Best, 1924, p. 231). Finally, among the Orokaiva "the women were always ready to urge on the fighting men and even to mingle in the fray as W.E. Armit found. ..when they finally were beaten off that two of the women lay dead with spears in their hands" (Williams, 1930, p. 164).

4. The Exclusion of Women from War

In the various cultures in which women are excluded from participation in active fighting, the male monopoly on warfare generally extends far beyond the battle field. Women are generally excluded from all meetings and councils and discussions of warfare and they are usually forbidden to make, own or use the weapons of warfare such as spears or bows-and-arrows. Furthermore, in most cases they are also forbidden to use such weapons in hunting or fishing as well. In one culture the middle fingers on the right hand which would be necessary in order to pull a bow string are systematically cut off in young girls (Heider, 1970, p. 238). In this way the question of split loyalty may be avoided. Since women know nothing of war, there is no question about which side they might take. The question of split loyalties during warfare is not an abstract issue, but a practical matter of security for the warriors and the entire community. In most warfare, the most critical and deadly phase consists of surprise raids, often carried out before daybreak under the conditions of absolute secrecy. If even one woman were to know and betray the plans, making possible an ambush, the result would most likely be disastrous.

Even in those cultures in which there is no question of split loyalties, the proportion of women warriors is very small, much less than might be expected by chance. How can this be explained? To attempt an answer, we must go beyond a simple analysis of the contradiction between marriage and warfare and develop a general picture of the prehistory of war based upon a further analysis of its relationships with marriage systems.

5. Type of Warfare Determines Marital Residency

The causal relationship between marriage and warfare appear, to be a two way relationship not only may the type of marital residency help determine whether or not there are women warriors, but also the type of warfare may help to determine the type of marital residency. This may be seen from an examination of the full sample of cultures shown in Table 1. Cultures with internal warfare have, in most cases patrilocal marital residency and exogamy. Cultures with exclusively external warfare have, in most cases, matrilocal marital residency and, in most cases, there is no restriction that marriage must be exogamous. Finally, cultures with low frequencies of warfare tend to have no strict rules at all with regard to marriage, for example, they require neither patrilocality nor matrilocality and neither exogamy nor endogamy. These differences are significant by Chi Square tests, in which the cultures with internal war or exclusive external war, respectively, are compared to cultures with low rates of war, which suggests that the presence and type of warfare is a causal factor. Whereas 44 of 58 cultures with internal war are patrilocal, there is no such tendency in cultures with low rates of war (15 of 32) Chi Square = 6.4, probability less than .01 (Table 3). Whereas 14 of 25 cultures with exclusively external war are matrilocal, only 5 of 32 cultures with low frequency of war are matrilocal Chi Square = 8.7, probability less than .01 (Table 4). Finally, exogamy is associated with patrilocality. Of 67 patrilocal cultures, 30 are exogamous whereas in the appropriate control group of 24 bilocal and neolocal cultures, only two are exogamous, Chi square = 8.8, probability less than .01 (Table 5).