Ending Footbinding and Infibulation: A Convention Account

Gerry Mackie

American Sociological Review, Vol. 61, No. 6. (Dec., 1996), pp. 999-1017.

Stable URL: http://links.jstor.org/sici?sici=0003-1224%28199612%2961%3A6%3C999%3AEFAIAC%3E2.0.CO%3B2-0

American Sociological Review is currently published by American Sociological Association.

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/about/terms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/journals/asa.html.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

The JSTOR Archive is a trusted digital repository providing for long-term preservation and access to leading academic journals and scholarly literature from around the world. The Archive is supported by libraries, scholarly societies, publishers, and foundations. It is an initiative of JSTOR, a not-for-profit organization with a mission to help the scholarly community take advantage of advances in technology. For more information regarding JSTOR, please contact [email protected].

http://www.jstor.org Sun Feb 3 03:03:03 2008 ENDING FOOTBINDING AND INFIBULATION: A CONVENTION ACCOUNT*

Gerry Mackie St. John's College, University of Oxford

Female genital mutilation in Africa persists despite modernization, public education, and legal prohibition. Female footbinding in China lasted for 1,000 years but ended in a single generation. 1 show that each practice is a self-enforcing convention, in Schelling's (1960) sense, maintained by inter- dependent expectations on the marriage market. Each practice originated under conditions of extreme resource polygyny as a means of enforcing the imperial male's exclusive sexual access to his female consorts. Extreme po- lygyny also caused a competitive upward flow of women and a downward flow of conjugal practices, accounting for diffusion of the practices. A Schelling coordination diagram explains how the three methods of the Chi- nese campaign to abolish footbinding succeeded in bringing it to a quick end. The pivotal innovation was to form associations of parents who pledged not to footbind their daughters nor let their sons marry footbound women. The "convention" hypothesis predicts that pronzotion of such pledge asso- ciations would help bring female genital mutilation to an end.

emale genital mutilation (clitoridectomy to 300 years, according to one educated Fand infibulation; hereafter FGM), a Sudanese (Lightfoot-Klein 1989: 135). Pain- painful and dangerous practice, is one of the ful and dangerous footbinding afflicted most major human rights and public health prob- Chinese women for a thousand years, and re- lems in the world today. It affects perhaps as form-minded Chinese women at one time many as 100 million women across some two "agreed that footbinding was of no use, but dozen countries in Africa. Opposition to could only be given up by degrees" (Little FGM is now part of American human rights 1899:152). Yet footbinding ended, for the policy (U.S. Department of State 1994:xvi), most part, in a single generation. I will show and the U.S. Agency for International Devel- that footbinding and infibulation are closely opment is assisting the African organizations equivalent practices, and thus that the suc- working to eradicate it (Mann 1994). cessful campaign to end footbinding in China Rather than diminishing with moderniza- has lessons for the efforts to end infibulation tion, the practice is spreading. Many observ- in Africa. ers, especially those indigenous, predict that Footbinding and infibulation correspond as ending infibulation will be slow-taking up follows: Both customs are nearly universal where practiced; they ..re persistent and are * Direct correspondence to Gerry Mackie, St. practiced even by those who oppose them. John's College, Oxford OX1 3JP, United King- Both control sexual access to females and dom ([email protected]). I especially ensure female chastity and fidelity. Both are thank James Fearon, Russell Hardin, David necessary for proper marriage and family Laitin, and also Gary Cox, Hill Gates, Michael honor. Both are believed to be sanctioned by Munger, John Padgett, Lynn Sanders, the ASR tradition. Both are said to be ethnic markers, Editor and reviewers, and others for advice and comments. This work was started while I was and distinct ethnic minorities may lack the supported by a Searle Fellowship at the Univer- practices. Both seem to have a past of conta- sity of Chicago. [Reviewers acknowledged by the gious diffusion. Both are exaggerated over author are Elise Boulding, Margo Wilson, and time and both increase with status. Both are Michael W. Macy. -ED.] supported and transmitted by women, are

American Sociological Review, 1996, Vol. 61 (December999-1017) 999 AMERICAN SOCIOLOGICAL REVIEW performed on girls about six to eight years paternity-confidence mechanisms under con- old, and are generally not initiation rites. ditions of imperial female slavery. The stron- Both are believed to promote health and fer- gest alternative theory-ascribing female tility. Both are defined as aesthetically pleas- mutilation to patriarchy-does not explain ing compared with the natural alternative. the nonuniversality of the practices under Both are said to properly exaggerate the universal patriarchy, maintenance beyond complementarity of the sexes, and both are originating conditions, or the end of claimed to make intercourse more pleasur- footbinding. I use a Schelling coordination able for the male. Important general differ- diagram to show that the end of footbinding ences between imperial China and Sudanic was a convention shift in direct response to Africa are elite concubinage in China versus the three methods of the abolition cam- commonplace polygyny in Africa, exogamy paigns. The major alternative theory of ces- versus endogamy, and agrarian and commer- sation-generalized modernization-does cial versus pastoral and horticultural produc- not seem to apply to footbinding or infibula- tion.' Important similarities between Impe- tion. Therefore the methods used to end rial China and Sudanic Africa are their histo- footbinding in China might well work to end ries of imperial female slavery and their rules FGM in Africa. of emancipation for the children of concu- bines. FOOTBINDING As far as I know. Ortner (1978) was the first on record with a theoiy relating the Beginning at about age six to eight, the fe- complex of female purity, family honor, se- male child's four smaller toes were bent un- clusion, chastity, and fidelity (with reference der the foot, the sole was forced to the heel, to footbinding and female genital mutilation) and then the foot was wrapped in a tight ban- to a past of highly stratified empires. Ex- dage day and night in order to mold a bowed treme resource polygyny and consequent and pointed four-inch-long appendage. Foot- hypergynous competition in an imperial past binding was extremely painful in the first 6 elicited the complex of honor and modesty, to 10 years of formative treatment. Compli- which persists beyond the originating condi- cations included ulceration, paralysis, gan- tion~.~Stacey (1983:40-43) applied Ortner's grene, and mortification of the lower limbs theory to the case of footbinding. Dickemann (Drucker 1981); perhaps 10 percent of girls (1979, 1981) presented a similar theory of did not survive the treatment (Fairbank paternity confidence in much greater detail, 1992). The saying was: A mother can't love but with a strongly sociobiological bent. I both her daughter and her daughter's feet at borrow and modify the paternity-confidence the same time (Blake 1994:682). Bound feet hypothesis and, with the help of simple game were malodorous, and treated women were theory, supplement it with an explicit theory crippled and largely housebound. The cus- of the endurance and demise of mutilating tom was defended even by women and was practices. I use Schelling's (1960, 1978) ac- transmitted by them. count of conventions as solutions to recur- Footbinding appeared in the Sung Dynasty rent coordination problems to explain the lo- (960-1279), a time of strong urbanization, cal universality and persistence of foot- expanding bureaucracy, commercialization binding and infibulation, which originated as of agriculture, monetization, and thriving trade. The respective imperial capitals at the I In this article, Sudanic refers to the region opposite ends of the Grand Canal were the across Africa, and Sudan refers to the country in largest and richest cities in the world Northeast Africa. (Fairbafik and Reischauer 1978:116-51). The For the periods under discussion, a man in status of women declined in the Sung: Con- China could have one wife and additional concu- cubinage expanded, upper-class dowries in- bines, in non-Islamic Sudanic Africa many wives and concubines, and under Islam four wives and creased, and a neo-Confucian ideology in- additional slave concubines: All are termed po- cluding tenets of female chastity, seclusion, lygyny for the purposes of this article. Hypergyny and subordination, emerged and came to is when women marry up-that is, a pattern of reign (Ebrey 1991). The first unequivocal females marrying into higher ranking families. record of footbinding is praise by an elev- ENDING FOOTBINDING AND INFIBULATION

enth-century poet of a dancer in "the palace by the American prohibitionist pledge to ab- style." A late fourteenth-century document stain from alcohol, the society introduced the quotes an early twelfth-century discussion of effective technique of pledging members not the recent origin of footbinding. Citing a to bind daughters nor let sons marry bound source otherwise unknown to China scholars, women. Western women organized the na- the same discussion also states that foot- tional Natural Foot Society, aimed at the binding was invented by a favorite dancer in non-Christian elite, in Shanghai in 1895. An the palace of Southern T'ang emperor and indigenous Anti-Footbinding Society estab- love poet Li Yu (937-978) and then spread lished headquarters in Shanghai in 1897 and by imitation until people were ashamed not eventually acquired 300,000 members. The to practice it. ~isualreferences become societies propagandized the disadvantages of more common from the late twelfth century footbinding in Chinese cultural terms, pro- on (Ebrey 1990; Levy 1966). moted pledge associations, and subtly con- The practice effloresced along three di- veyed international disapproval of the cus- mensions over several centuries. First, it tom. By 1908, leading Chinese public opin- spread from the imperial palace, to court ion was opposed to footbinding, and the circles, to the larger upper classes, and then leadership of the Natural Foot Society was to the middle and lower classes; eventually transferred to a committee of Chinese the higher the social status, the smaller the women. The Nationalist Revolution banned foot. Second, it became more exaggerated footbinding in 1912, and the decree suc- over time; a practice supposedly originating ceeded in many locales (Drucker 198 1). among dancers eventually made dance a for- Footbinding started to end in China be- gotten art. Third, it radiated from the impe- tween the Boxer Rebellion of 1900 and the rial capitals to the rest of the empire. Foot- Revolution of 191 1, certainly among the up- binding was clearly the normal practice by per strata of the larger cities. Although there the Ming Dynasty (1368-1644). As mea- was local variability in onset of cessation, sured in 1835, it prevailed in the whole em- available evidence is that whenever binding pire among the Chinese, affecting 50 to 80 did end, it ended rapidly. As measured by a percent of women depending on the locale, sociologist's data, for example, the popula- the disgraceful exceptions only among the tion of Tinghsien, a conservative rural area lowest classes, wherever woman's work was 125 miles south of Peking, went from 99 per- needed in the field or workshop (Levy cent bound in 1889 to 94 percent bound in 1966:23-106). The missionary Justus 1899 to zero bound in 1919 (Gamble 1943). Doolittle (1865: 201) reported that bound The Chinese offered various explanations feet were a sign of "gentility," that "many for footbinding. It was said to distinguish the poor families prefer to struggle along for a Chinese from the invading Mongols and precarious living, bringing up their daughters other barbarians and to enhance the differ- with small feet . . . in order to attain a more ence between men and women. It was be- competent support," and that some Chinese lieved to promote good health and fertility opposed footbinding as useless but felt (Blake 1994:686). For Chinese men, bound obliged to conform for the sake of proper feet were universally associated with higher- marriage. status love and sex, and so carried strong The Manchu conquerors opposed foot- connotations of both modesty and lascivious- binding, but their efforts to abolish it in 1665 ness. Bound feet became a sexual fetish; they failed entirely, despite intimidating penalties. were said to be conducive to better inter- An 1847 Manchu edict against footbinding course, but this claim was medically false also failed. Throughout- the nineteenth cen- (Van Gulik 1961 :219).The leading neo-Con- tury the practice was condemned by influen- fucian philosopher Chu Hsi imposed foot- tial liberal literati, but with no apparent ef- binding on a southern province in the twelfth fect (Levy 1966:68-74). Some of the Protes- century so as to enforce female chastity (an tant missionaries, particularly the women, apocryphal tale, according to Ebrey 1990). denounced the custom. The first antifoot- The historical record contains several ex- binding society was founded in 1874 by lo- plicit statements over the centuries that the cal missionaries for their converts. Inspired purpose of footbinding was to hobble women AMERICAN SOCIOLOGICAL REVIEW

and thereby promote their seclusion and fi- and education opportunities, improving trans- delity. Finally, the record is abundantly clear portation, and capitalist individualism were that the immediate explanation always given attracting women away from the weakened for footbinding in its heyday was to secure a patriarchal family and to the labor market. proper marriage (Levy 1966). Levy (1966) believes that footbinding INFIBULATION originated in aesthetic appreciation of the small foot and was maintained by male erotic FGM occurs in Egypt and in what was for- interest. Veblen (1934) regards it as an osten- merly called the Sudanic Belt in Africa (the tatious display of the practicing family's savanna lands between the desert to the north wealth. Freud (1961) concludes that foot- and the jungle to the south, or between the binding appeased male castration anxieties. Tropic of Cancer to the north and the Equa- Ebrey (1990) proposes that the practice arose tor to the south); from the Atlantic Coast of in the Sung as a way of ethnically differenti- western Africa to Egypt in the northeast and ating civilized Chinese from invading north- Kenya in the southeast. Infibulation, the ern barbarians and of maintaining gender dis- harshest practice, occurs contiguously in tinctions as refined Chinese males became Egyptian Nubia, the Sudan, Eritrea, Djibouti, more effeminate. Blake's "mindful-body" and Somalia, also known as Islamic North- theory interprets footbinding as a voluntary east Africa. ordeal by which mothers taught their daugh- The following classification of female ters to succeed in the male-controlled neo- genital operations is paraphrased from Confucian world. related to the intensifica- Kouba and Muasher (1985:96): tion of agricultural labor in the period of du- (1) Mild sunna: The pricking of the prepuce ration (all cited in Blake 1994). Gates (1989) of the clitoris with a sharp instrument points to the commodification of agriculture such as a pin, which leaves little or no and textile production in the Sung. some mis- damage. (Sunna means "tradition" in sionaries blame flawed national character Arabic.) (Greenhalgh 1977). Fairbank and Feuer- werker (1986:29) attribute footbinding's gen- (2) Clitoridectomy/excision: The removal of erality and maintenance to homogeneous cul- part or all of the clitoris as well as part ture and peasant subservience to elite norms. or all of the labia minora. The resulting Daly (1978: 130-33) links footbinding, scar tissue may be so extensive as to genital mutilation, and other misogynistic cover the vaginal opening. practices in a patriarchal, sado-ritual syn- (3) Infibulation/Pharaonic circumcision: drome that is obsessed with purity, is sanc- Clitoridectomy and the excision of the tioned bv tradition. and has "an inherent ten- labia minora as well as the inner walls dency to 'catch on' and spread, since [the ritu- of the labia majora. The raw edges of the als] appeal to imaginations conditioned by vulva are then sewn together with catgut the omnipresent ideology of male domina- or held against each other with thorns. tion," uses women as scapegoats and token The raw edges of the labia majora are torturers, ritualizes and "normativizes" sutured together or approximated so that atrocities, and is legitimated by male schol- the opposite sides will heal together and ars despite appearances of disapproval. form a wall over the vaginal opening. A Greenhalgh (1977) ties the custom to the pa- small sliver of wood is inserted into the triarchal Chinese family, arguing that foot- vagina to stop coalescence of the labia binding consolidates and perpetuates the kin- majora in front of the vaginal orifice and ship system. She takes note of upward mar- to.allow for the passage of urine and riage, of downwardly migrating custom, of menstrual flow. footbinding as the "essential criterion for any girl's marriageability," of the self-perpetuat- For the sake of brevity, I concentrate on in- ing character of the practice, and of the shock fibulation. de-livered by Western imperialism. The operation takes place from a few days Greenhalgh also offers an explicit theory of after birth to before the birth of the first child footbinding's end: Expanding employment (depending on local custom), but mostly ENDING FOOTBINDING AND INFIBULATIO seems to be performed on girls around age the Red Sea, wrote that one group excised eight, safely before puberty. It is usually done their women in the manner of the Egyptians, among women, in private, with little cer- and that another group "cut off in infancy emony; only rarely does it have the trappings with razors the whole portion that others cir- of an initiation rite. The girl is held down cumcise" (Agatharchides 1989: 11 l-12).3 amid singing and shouting, which drown out The geographic distribution of FGM sug- her screams. Except recently among the af- gests that it originated on the western coast fluent, the operation is inflicted without pain- of the Red Sea, where infibulation is most killer or antiseptic precaution. Then the girl intense, diminishing to clitoridectomy in lies with her legs tied together for several westward and southward radiation. Whatever weeks. Urination and (later) menstruation are the earliest origins of FGM, there is certainly difficult because of the pencil-point opening an association between infibulation and sla- left by the operation. After marriage, penetra- very. The Egyptians raided and traded the tion takes two weeks to two years to accom- Black south for slaves from dynastic to Byz- plish, or is facilitated by knife; it is traumati- antine times, and Sudanic slaves were ex- cally painful for the female, and some men ported through the Red Sea to the Persian feel guilt and revulsion at the cruelty in- Gulf before the rise of Islam (Beachey 1976: volved. Childbirth requires introcision and 2-4). The Islamic slave trade delivered many resewing of the genital area. Virtually every Sudanic concubines and maids to Egypt and ethnography and report states that FGM is Arabia. defended and transmitted by the women. The In 1609 Dos Santos reported that inland mothers who have this done to their daugh- from Mogadishu (Somalia) a group has "a ters love their children and want the best for custome to sew up their Females, specially them (Assaad 1980). their slaves being young to make them un- Health consequences are severe. Immedi- able for conception, which makes these ate: Pain, hemorrhage, shock, acute urinary Slaves sell dearer, both for their chastitie, retention, urinary infection, blood poisoning and for better confidence which their Mas- (septicemia), fever, tetanus, and death. Inter- ters put in them" (Fr. Joao dos Santos, Ethio- mediate: Delay in wound healing, pelvic in- pia Oriental, in Freeman-Grenville 1962). fection, dysmenorrhea, cysts and abscesses, Browne, reporting in 1799 on his African keloid scar, and painful intercourse. Late: travels, wrote that the Egyptians practice fe- Haematocolpos (vaginal closure and accu- male excision, and that infibulation to pre- mulation of menstrual fluid), infertility and vent pregnancy is general among female miscarriage, recurrent urinary tract infection, slaves, who come from the Black south. difficulty in urinating, calculus and stone for- Other early travelers to Egypt-Larrey in mation, hypersensitivity, and anal inconti- 1803 and Burckhardt in 18 19-confirm nence and fissure. Intercourse: Difficulty in Browne and claim that slave traders infibul- penetration, painful intercourse, and use of ated young female captives (Widstrand misplaced deinfibulation wound as false va- 1964:102).4 Curiously, infibulation is called gina. Delivery: Prolonged and obstructed la- "Sudanese circumcision" by the Egyptians, bor, hemorrhage leading to shock and death, perineal laceration, uterine inertia, and still- Today's Beja, pastoralists and recent slave- born or brain-damaged infants. Postnatal: holders inhabiting the same territory, are almost Urinary and rectal fistula causing odor and certainly their direct descendants. According to miscarriages, and prolapse of uterus and ad- Hicks (1993:215), some Beja populations excise; jacent organs. Sexual problems: Lack of or- others infibulate. gasm, anxiety, depression, and frustration Widstrand speculates that the travelers were (summarized from headings in Koso-Thomas misinformed, and that instead slave traders paid a 1987:25-28). higher price for women already infibulated in their home territory. This may be true, but The origins of FGM are obscure. Wid- Browne and Burckhardt each visited the Sudan. strand (1964: 116) traces several classical ref- Hicks (1993:22), relying on Cloudsley's (1983: erences to the second century B.C.E. geog- 112) misinterpretation of Widstrand's specula- rapher Agatharchides of Cnidus, who, report- tion, mistakenly denies that an association be- ing on tribes residing on the western coast of tween slavery and infibulation is established. AMERICAN SOCIOLOGICAL REVIEW but "Pharaonic circumcision" (i.e., Egyptian) ject women was 95 percent, the same as for by the Sudanese (Kouba and Muasher 1985). their mothersn5 Infibulation is not only nearly universal The most common explanation given by and persistent where practiced, but is ex- participants is that infibulation is required for panding its territory. It is spreading from marriage and honor. Infibulation prepares for Arabized northern Sudan further into indig- marriage, is a prerequisite for marriage, enously populated areas of southern and makes for better marriage prospects, makes western Sudan; as Arabized traders enter or possible the security available through mar- as indigenes urbanize, the less advantaged riage, and so on. It fosters virginity, first, be- adopt infibulation to make their daughters cause the physical barrier prevents rape, and more marriageable to the high-status outsid- second, because the physical barrier and the ers. For example, infibulation was unknown attenuation of sexual desire protect the sup- in Nyala in west Sudan 50 years ago, but it posedly oversexed and promiscuous woman now saturates the area and is universal in from temptation. It is proof of virginity and Nyertete merely 20 years after introduction. secures fidelity by reduction of female desire Where practiced in the Sudan, FGM affects and by reinfibulation upon the prolonged ab- 99 percent of women; 83 percent are infi- sence of the husband. Infibulation is closely bulated. Moreover, infibulation has been fur- associated with the modesty code. ther exaggerated in the Sudan, beginning One common explanation, sometimes ne- among educated urban dwellers and spread- glected due to the inquisitive bias of intel- ing to the uneducated in the villages, with the lectuals, is simply that "such is the custom new practice of reinfibulation to pinhole size or tradition here." A variation on the custom after each birth; this is quite damaging to the explanation is that "this is a practice that dis- woman's physical integrity over time (reinfi- tinguishes us from neighboring groups." A bulation is reported among 2 percent of related explanation of FGM that is conven- women over age 64 and among 54 percent tional among some European-educated ages 25 to 34; among 70 percent of urban Africanists is that the practice functions to women and 28 percent of rural women) promote the solidarity of the group. That ex- (Dareer 1982:57; Lightfoot-Klein 1989:48- planation originated with Kenyatta, the func- 49, 31. 98-99). tionalist Malinowski's pupil, who used Brit- In Somalia, 99 percent of women are mu- ish attempts to prohibit clitoridectomy tilated (76 percent of those are infibulated), among his Kikuyu as a theme of ethnic uni- and the infibulation rate nears 95 percent fication and nationalist agitation. Religious among the noble northern clans (Grassivaro- command is another common explanation: Gallo and Abdisamed 1985). Infibulation For many of the practitioners, tradition and was not reported among the southern Sab at religion mean one and the same thing (Boddy the beginning of the century, but has com- 1991). menced among them. Reportedly they want FGM is found only in or adjacent to Is- to gain higher social status by emulating the lamic groups (some neighboring Christians noble northerners (Grassivaro-Gallo and practice it to avoid damnation). This is curi- Viviani 1992). Grassivaro-Gallo and ous because FGM, beyond the mild sunna Abdisamed (1985) conducted a large but supposedly akin to male circumcision, is not nonrandom survey of women in the capital. found in most Islamic countries nor is it re- According to their analysis of a subsample quired by Islam. Mutilation is not practiced (parental origin and place of operation in in Mecca or Medina, and Saudis reportedly the same region), infibulation is attenuating find the custom pagan. The Koran is silent to clitoridectomy. However, disaggregating on FGM, but several hadith (sayings attrib- these data into high-infibulation regional origin and low-infibulation regional origin shows that the apparent attenuation is due Of the capital residents originating from high- infibulation regions, 49 percent expect that their simply to the inclusion of the southern Sab daughters will not be infibulated (but will be mu- in the sample. Other than among respon- tilated); from low-infibulation areas, 65 percent. dents originating from the capital or the I doubt, however, that Somali hopes will be real- atypical south, the infibulation rate of sub- ized short of a convention shift. ENDING FOOTBINDING AND INFIBULATION

uted to Mohammed) recommend attenuating women; a precaution against rape . . . ; a nec- the practice for the woman's sake, praise it essary preliminary to adult status and mar- as noble but not commanded, or advise that riage" (Hicks 1993: 181). female converts refrain from mutilation be- Infibulation seems to have attenuated to cause even if pleasing to the husband it is clitoridectomy in Egyptian Nubia (Kennedy painful to the wife. In Egypt, the Christian 1970). Kennedy explains this change with a Copts follow the same practice of clitoridec- modernization hypothesis: As Nubians lost tomy as their Islamic neighbors; in the Sudan their land to the flooding of the Aswan dam, the tiny Coptic minority follows the same took up wage labor, and came into contact practice of infibulation as their Islamic with urban values, along with increased edu- neighbors (Cloudsley 1983). Catholic mis- cation and communication, then dependence sionaries among the Egyptian Copts in the on the patrilineage decreased, the nuclear seventeenth century initially forbade clitoral family thrived, modern egalitarian ideas were excision as an allegedly Jewish practice, but adopted, and so on. Hayes (1975:630-31) male converts to Catholicism refused to finds that the manifest function of infibula- marry intact women and so would return to tion "is to regulate female sexuality in order the Coptic Church. Then a special investiga- to safeguard virginity, thus protecting the tion by the Vatican decreed that the oversized honor of a woman's and her husband's clitoris alleged among Egyptian women jus- patrilineages." Latent functions include con- tified excision as conducive to marriage tributing to the village economy, maintaining (Meinardus 1967). the status and role of midwives, and limiting The remaining explanations seem to be population. Building on Kennedy, Hayes pre- post hoc. ~nfibulationis sometimes justified dicts that industrialization and urbanization as clean and aesthetic. It is said to prevent weaken the corporate nature of the lineage malodorous discharge; however, the prolon- and that FGM then becomes less severe. gation of urination and menstruation and Some explanations fail the comparative other complications make that a false belief. test. Boddy (1982) relates Sudanese infibu- The judgment that natural genitals are ugly lation to a complex of symbols for pure, is like the old Chinese judgment that natural clean, and smooth (she is correct. I believe. feet are ugly. Infibulation is also said to pro- that infibulation emphasizes female fertility mote health and fertility, and to make child- by de-emphasizing female sexuality). bearing easier; obviously these are false be- Grassivaro-Gallo and Viviani (1992) claim liefs. Infibulation is said to enhance male in- that Somali infibulation is an evolutionary tercourse, but reports from Nubia and the adaptation to make female shepherds and Sudan indicate that men aware of the differ- goatherds odorless and safe from prey. The ence prefer intact or excised women to symbolic associations of Boddy's riverine infibulated women (Boddy 1982; Lightfoot- agriculturists, however, do not explain in- Klein 1989:97). The story of enlarged geni- fibulation on the pastoral plains of Somalia; tals accepted by Catholic worthies with re- nor does Grassivaro-Gallo's hypothesis apply spect to the Egyptians, and by Arabs with re- to infibulation along the Sudanese Nile. Nei- spect to excising Africans, is medically false. ther theory accounts for the maintenance of Freud (as related by Bonaparte 1953:191- infibulation among urban dwellers and edu- 92) hypothesized that FGM assists the tran- cated elites. sition from clitoral orgasm to vaginal or- gasm, in other words, that FGM is a realiza- tion of his theoretical clitoridectomy. The THE CONVENTION HYPOTHESIS most ambitious study to date, comparative- Conventions ethnographic in approach, agrees with the in- digenous interpretation that infibulation is This is an account of conventions as solutions related to guaranteeing virginity and chastity, to coordination problems, as developed by and with academic interpretations relating Schelling (1960) and Lewis (1969). (This infibulation to "the requirement of virginity game-theoretic account of coordination prob- and chastity in societies where the honor of lems assumes strategic rationality: "Rational- the family is contingent on the honor of its ity" means choosing what one wants more 1006 AMERICAN SOCIOLOGICAL REVIEW

(A) Single Equilibrium (B) Coordination: Indifferent Equilibria

C1 C2 C 1 C2

(C) Coordination: (D) Coordination: Ranked Equilibria Conflicting Equilibria

Figure 1. Game Matrices Note: Shaded boxes indicate equilibrium choices over what one wants less, whether self-re- coordinate on Rl Cl and carry home a payoff garding or other-regarding, given beliefs and of 2 each; the same is true if they coordinate constraints; a "strategic" situation is one on R2C2. If coordination fails-say they characterized by the interdependence of de- choose RlC2, or R2Cl-then each gets noth- cisions; see Elster 1986.) Schelling said that ing. The usual illustration of this is whether the coordination game lies behind the stabil- to drive on the left side of the road or the ity of institutions and traditions, yet can also right. It doesn't matter which side I drive on, explain rapid change. How can the same so long as everyone else does the same. mechanism explain both stability and Consider now the game in Figure IA. change? Look at the sample coordination Here, only RlCl is an equilibrium choice. game in Figure IB. (Generally for the games Figure 1A does not represent a coordination in Figure 1, there are two players; each player problem; for that, there must be at least two has two strategies; the lower left payoff in proper coordination equilibria, according to any box is that of player Row-Chooser, and Lewis (l969). Figure IB does represent a co- the upper right payoff in any box is that of ordination problem. So does Figure IC. In player Column-Chooser. Assume that the Figure IC, RlCl is better for each player players can't talk to each other and that they than R2C2, and R2C2 is better for each play pure strategies, not a probabilistic mix- player than the miscoordination at RlC2 or ture of strategies.) In Figure lB, Row- R2C1. If people are stuck at inferior equilib- Chooser chooses Row 1 (Rl) and Column- rium R2C2, they may lack a concerted way Chooser chooses Column 1 (CI), then they to move to superior equilibrium RICI. In ENDING FOOTBINDING AND INFIBULATION 1007 this paper, I concentrate on that type of prob- would be better, they could never spontane- lem. Figure 1D also represents a coordina- ously, by some invisible hand, get to the bet- tion problem, but now with a bargaining as- ter coordination equilibrium. Convention is pect. Here, Column-Chooser does best at self-enforcing: Any one person driving to the RlC1, while Row-Chooser does best at right to demonstrate its advantages would R2C2. and each likes either of these coordi- end up dead. For example, in left-driving Pa- nation equilibria better than miscoordination kistan a local religious party decreed that the at R1C2 or R2C1. This game has all the in- pious must drive on the right. The decree was gredients of power and tragedy (Knight 1992 rescinded in two weeks after a number of se- is the definitive treatment). Any game with rious accidents (Bedi 1994). two or more proper coordination equilibria represents a coordination problem. Singling out a coordination equilibrium is Paternity Confidence a matter of concordant mutual expectations. How do the practices of footbinding and in- If there are two of us and we can talk, then fibulation fit the convention model just de- we can each promise to choose either right scribed? For the explanation to succeed, one or left, and the promise is self-enforcing. If assumption must be true: that humans there are hundreds of us, however, express strongly desire to successfully raise their bio- agreement is difficult. Many conventions logical ~hildren.~The various forms of mar- suggesting a single choice of equilibrium in riage and family serve this end. Females are a coordination problem are not expressly certain of maternity, but males are not cer- agreed to; rather, they are tacit. Schelling tain of paternity. In the standard premodern (1960) urges that there is no logical solution case, the female requires assurances of re- to the tacit coordination problem; rather, so- source support for bearing and rearing chil- lutions are suggested by their psychological dren, and the male re~uiresassurances that salience. The salient choice is not uniquely the children are his offspring. Because of the good, just noticeably unique. It all depends desire for children, each party prefers mar- on what the players believe about each other. riage to nonmarriage. Marriage is a deeply In novel play of Game lC, absent credible interdependent choice and is a coordination communication, superior RlCl stands out as eauilibrium. Families advertise their male the salient choice for most people. But in a offspring as capable of providing both gen- recurring game, precedent is strongly salient. erous and sustained support, and their female If we played the same Game 1C yesterday at offspring as both fertile and faithful.' RlC1, then RlCl is the salient choice today. Under conditions of resource equality, hu- If the choice made yesterday was the inferior mans compete in conveying the many signs R2C2, R2C2 is weakly salient today; and if of trustworthiness to possible marriage part- R2C2 was the choice in our last 10 games, it ners. Under conditions of resource inequal- is strongly salient in our next. Coordination ity, conventions of wealth and honor emerge by precedent is convention. as signs of higher rank. When the inequality At 5:00 A.M., Sunday, September 22, 1967, of resource control reaches a certain extreme, Sweden switched from driving on the left polygyny and hypergyny appear because a side of the road to driving on the right female is then more likely to raise children (Hardin 1988; Ullman-Margalit 1977). Swe- successfully as the second wife of a high- den, or at least its authorities, saw driving on ranking man than as the first wife of a low- the left as more like the game in 1C than the ranking man. The richest families will also game in 1B. The rest of continental Europe drove on the right, so as international traffic increased, visitors to Sweden caused acci- 1 simplify and purge of sociobiological as- dents by driving in the wrong lane as did sumptions Dickemann's (1979, 1981) brilliant ar- gument. nonchalant Swedes abroad. Thus Swedes ' Consider, for example, the Ming Dynasty would be better off driving on the right, mov- (1368-1644) saying quoted by Levy (1966): "If ing from R2C2 to RlCl in something like you care for a son, you don't go easy on his stud- Game 1C. Even if the millions of Swedes ies; if you care for a daughter, you don't go easy were each convinced that driving on the right on her footbinding" (p. 49). AMERICAN SOCIOLOGICAL REVIEW then prefer male children to female children such scrupulous guarding-the "two-Marys" because each polygynous son will generate or "madonna-whore" complex. The false be- more grandchildren than the equivalent lief in enlarged genitals as the rationale for daughter. The higher the male's rank, the FGM is based on similar reasoning. greater the resource support he offers, the Mutilation is a conventional prerequisite greater number of consorts he attracts, the of marriage, originating as a sign of pater- greater his costs of fidelity control, and thus nity confidence in a game of coordination the greater the competition among female with conflict like that shown in Figure ID. families to guarantee paternity confidence. Mutilation is a superior equilibrium for the Therefore, families will advertise the honor originating male emperor in terms of maxi- of their lines, the purity of their females, and mizing the number of his own children; it is their members' commitment to the values of an inferior equilibrium for the females under chastity and fidelity, the so-called modesty his control. A convention signifying paternity code. confidence was imposed at the social apex, A highly polygynous apex induces an up- and then hypertrophied and diffused in posi- ward flow of women (hypergyny) and a tional competition. Mutilation is maintained downward flow of conjugal practices. If hu- as the inferior equilibrium in a game of co- mans want to have their own children, then ordination with ranked equilibria like the one an emperor will take costly measures to en- in Figure 1C. It is an inferior convention for sure that the several thousand women he sup- the ordinary male and for his one female ports are sexually reserved for him, while the mate: The monogamous male has trivial fi- interests of his wives and concubines will be delity-assurance costs, and both male and fe- to seek clandestine insemination from men male could better afford children if the more available than he. It is then in the female's fertility, productivity, and longevity emperor's interest to inflict costly methods were not damaged by mutilation. For the pur- of fidelity control. The next lower stratum, poses of reform it is important to realize that competing to provide wives and concubines the ordinary male, when fully informed, may to the apex, will imitate and exaggerate the be motivated to support convention shift. fidelity-control practice so as to gain eco- Men and women prefer marriage to non- nomic, social, and reproductive access to the marriage. Both men and women would be palace. The vacuum of women in the first better off marrying without the mutilating lower stratum will be filled by women mov- practice (RlCl), but they are trapped by the ing up from the second lower stratum, who inferior convention (R2C2). However the in turn will adopt the fidelity-,control conven- custom originated, as soon as women be- tion, and so on, all the way down. lieved that men would not marry an Local conventions of modesty emerge; unmutilated woman, and men believed that footbinding in one place, infibulation in an- an unmutilated woman would not be a faith- other. Under extreme polygyny and hyper- ful partner in marriage, and so forth, expec- gyny, the conventions are most intense at the tations were mutually concordant and a self- highest rank but domino down to the lowest enforcing convention was locked in. A rank that can afford the practice. Female woman would not choose nonmarriage and modesty in these circumstances is a posi- not to have her own children; a man would tional good (valued not for attaining a stan- not choose an unfaithful partner and not to dard, but for its rank; not "excellent," but have his own children. "best") and thus is driven to maximum af- fordable values on the conventions: "One Imperial Female Slavery wrong word about his sister and he'll kill you"; "The errant daughter shall die"; "The FGM is pre-Islamic but was exaggerated by smaller the foot, the better the family"; "The its intersection with the Islamic modesty code smaller the infibulation opening, the better of family honor, female purity, virginity, the girl's reputation." Naive observation of chastity, fidelity, and seclusion. I propose that runaway modesty practices gives rise to the imperial polygyny in Arabia and Egypt, and folk belief, in honor cultures, that women indirectly in Istanbul, induced an eastward must be excessively lustful to necessitate flow of female slaves through the mainly po- ENDING FOOTBINDING AND INFIBULATION 1009

lygynous Sudanic Belt into infibulating slave concubines, courtesans. and housemaids as centers in Sudan and a westward flow of Is- well as government-owned brothels for low- lamization and FGM. Arabized pastoralists ranking officers and soldiers (Van Gulik raided northeastern Africa for slaves, and, 1961 :212-35). The many women of the em- because Islam forbade the enslavement of peror and of other officials no doubt needed Moslems, ventured further as closer sources guarding against escape, trysts, rape, and converted. The Sudanic slaves were shipped bastardy. Initial adoption of footbinding by down Nile Valley routes or through the Red the imperial seraglio may have had a second- Sea to Egyptian or Arabian markets. The dis- ary aesthetic aspect, but the fidelity-control tribution of infibulation in Sudan in the nine- aspect must have been primary. Notice as teenth century, by far the peak period of slav- well that footbinding- made it harder for bar- ing (almost one-half million were exported barian raiders to steal the palace women be- by way of the Red Sea, [Clarence-Smith cause they would have had to be carried 1989:5]), follows the caravan routes (Hicks rather than driven. 1993:21-23). Slave raiding and slave concu- binage continue in today's Sudan (U.S. De- Belief Traps partment of State 1994:285). The further ra- diation of clitoridectomy follows the channel The women who practice infibulation are of raiding and trading west to the Atlantic and caught in a belief trap. The Bambara of Mali southeast to Kenya. A practice associated believe that the clitoris will kill a man if it with shameful female slavery came to stand comes in contact with the penis during inter- for honor. Since defenders of infibulation re- course. In Nigeria, some groups believe that gard the uninfibulated as no better than con- a baby will die if its head touches the clitoris temptible slaves, demonstrating its origin in during delivery (Lightfoot-Klein 1989:38-9). the slave trade might contribute to its eradi- I call these self-enforcing- beliefs: a belief cation. that cannot be revised because the believed Hereditary stratification decreased in costs of testing the belief are too high. Koso- "egalitarian" Sung China (Ebrey 199 l), but Thomas (1987) interviewed 50 women in Si- wealth stratification increased with strong erra Leone who had known sexual experi- urbanization (Ma 197 1: 144). Commercial- ence before clitoridectomy. All reported de- ization probably decreased the relative price creased sexual satisfaction after the opera- of female labor in agriculture, outdoor work, tion, but they were unaware of the causal re- and increased its price in indoor work, such lationship until informed by the interviewer. as commodity textile production, household ~ronically,some of these women had become services, and male entertainment. In the promiscuous in their search for lost satisfac- China of female infanticide, poor and tion. Lightfoot-Klein's (1989:22, 59) initial middle-range families in the southern Sung interviews with Sudanese women elicited the capital preferred female children because of response, for example, that urination was their bright prospects as concubines and "normal." She then switched to more de- maids for rich families (Ma 1971:143). Per- scriptive questions such as "How long does haps footbinding arose as a slave traders' re- it take you to urinate?" The answer then was straint on girl children; its obvious purpose "Normal-about 15 minutes." was to keep them from running away. Then, The painful surgery, prolonged urination nubile courtesans adapted dance perfor- and menstruation, traumatic penetration, and mance to the restraint. The eleventh-century unbearable childbirth accompanying infibu- poet wrote, "Anointed with fragrance, she lation are all accepted as normal. Because it takes lotus steps;/ Though often sad, she is inflicted on all girls before puberty they steps with swift lightness" (Levy 1966:47). have no basis of comparison; the connection Why is she sad? Because, she remembers her between cause and effect is remote; and the impoverished freedom in her prosperous im- cost of testing pertinent beliefs is prohibitive prisonment? The imperial capitals contained to any one individual. Once in place, conven- not only the thousands of concubines, danc- tions regulating access to reproduction are ers, and servants of the emperor, but also an deeply entrenched, in part because dissent- upper class serviced by additional legions of ers fail to have descendants. Adult-to-child AMERICAN SOCIOLOGICAL REVIEW transmission augments persistence. One con- character explanation (still alive in some of sequence of a culturally insulated group's be- the denunciations of FGM) neglects refer- coming stuck in an inferior convention is that ences to honor and modesty, practice by op- as the period of origination passes, knowl- posing families, and especially the end of edge of the superior coordination equilib- footbinding. Otherwise, alternative theories rium disappears, as if the northwest quadrant are not so much conflicting as incomplete. of the matrix in Figure 1C were to fade away. The aesthetic interest, erotic interest, and os- Exposure to an alternative convention is a tentation theories each explain only a few of necessary but not sufficient condition for the observations, and especially fail at ac- convention shift. Such exposure may have counting for local universality, persistence, contributed to the abandonment of Chinese and practice by opposing families. The eth- footbinding. nic differentiation theory is consistent with the observations except for references to con- trolling sexual access to females, practice by Evaluation opposing families, and the end of foot- By making three assumptions-that people binding. Fairbank and Feuerwerker's (1986) are strategically rational, that people desire appeal to cultural homogeneity and compli- to successfully raise their own children, and ance with elite norms explains contagion, lo- in each case that there was an originating cal universality, and stubborn persistence as condition of imperial female slavery-I have well as references to tradition, ethnicity and been able to show why female mutilation is status; but the convention hypothesis ex- a self-enforcing convention, nearly universal plains their homogeneity and norm compli- where practiced, persistent, and practiced ance, and explains sudden change. even by those who oppose it; why it is nec- Blake (1994) carefully recounts the ob- essary for a proper marriage; and why it is served facts on footbinding, and his compli- such an obviously compelling tradition to cated theory that (among other things) it practitioners. That the convention is univer- functioned to obscure the labor power of sal within overlapping marriage markets ex- women accounts for the major observations. plains why distinct ethnic minorities may not Blake detects a puzzling conflict between the conform and why practitioners later come to strict Confucian norm against mutilation of regard it as an ethnic marker. The fact that the body and the widespread norm of muti- overlapping marriage markets are initially lating women's feet. Those norms are con- hypergynous, and the consequent positional sistent under the convention account, how- competition, explain contagious diffusion, ever, because each was justified in terms of exaggeration over time, and status grading. promoting fertility. He surveys the relation The men who perpetrate female enslave- between the duration of the patrilineal fam- ments and mutilations are not absolved un- ily and hobbling, on the one hand, and the der this hypothesis; but, as Ortner (1978:32) changing demand for female labor, on the discerns, the resulting hypergynous competi- other, but characterizes that relation equivo- tion explains why women as well as men ac- cally. With one exception, the convention hy- tively perpetuate the honor and modesty pothesis is the only theory that shows how code, and resolves a major puzzle. Moreover, mutilation could be practiced by those who the convention hypothesis explains how in- oppose it and that explains and predicts centives to conform linger after originating change. The exception is the theory that the conditions have faded. The belief that muti- patrilineage rises and falls with economic lation is excellent for health and fertility is conditions. The patrilineage theory, however, self-enforcing, a trap. Similarly, the tales of does not explain the absence of mutilation aesthetic appreciation, sexual pleasure, and under similar economic and family condi- gender complementarity are associationist tions elsewhere. Functionalism explains ev- folk explanations. erything but change, which may be why this Freudian theory offers two disparate mech- theory is found linked with the patrilineage anisms that explain nothing about the details theory; functional stability disrupted by or correspondences of infibulation and foot- modernization covers the observations but is binding. The missionaries' flawed-national- vague on origins and on variations. ENDING FOOTBINDING AND INFIBULATION 1011

0 N Number of Others Choosing R(Natura1 Feet)

Figure 2. Schelling Coordination Diagram for Chinese Footbinding Note: L = utility of choice resulting in an inferior equilibrium (footbinding); R = utility of choice result- ing in a superior equilibrium (natural feet): L'= utility of footbinding after hearing of its disadvantages; R' = utility of natural feet after hearing of its advantages; k = minimum number of people required to be better off together choosing natural feet; k' = new smaller k resulting from a successful propaganda campaign; t = Point at which marriage market tips to either footbinding or to natural feet; t'= new tipping point resulting from a successful propaganda campaign; 0 = arbitrary zero value of inferior equilibrium (footbinfing); 0'= arbitrary zero value of footbinding after hearing of its disadvantage.

Daly's (1978) inductively constructed pa- that China was losing face in the world and triarchal sado-ritual syndrome accounts for was subject to international ridicule. Second, most observations, but not for variations in their education campaign explained the ad- the presence and type of mutilation under a vantages of natural feet and the disadvantages universal patriarchy. The sudden end of of bound feet. Third, they formed natural-foot footbinding before the end of male domina- societies, whose members pledged not to bind tion she assigns to the similarly malicious their daughters' feet nor to let their sons category of patriarchal "reversal." Dicke- marry women with bound feet. These three mann's (1981) paternity confidence theory is tactics are appropriate for escaping an infe- comprehensive but does not, as does the con- rior convention. I will illustrate with a vention hypothesis, explain endurance be- Schelling diagram (see Figure 2). yond originating conditions or the rapid de- The Schelling coordination diagram is a mise of such practices. The strength of the way of visualizing a many-person version of convention hypothesis is in explaining the the two-person coordination game in Figure end of footbinding, thereby disclosing tactics 1C. Suppose a reference population of n+l for the eradication of FGM. homogeneous individuals in a marriage mar- ket. The vertical axis represents higher or lower value to the typical chooser. The hori- ESCAPING INFERIOR CONVENTIONS zontal axis runs from zero other people on The work of the antifootbinding reformers the left to n other people on the right. In this had three aspects (Drucker 1981; Levy 1966: particular application, one curve, L, repre- 74-88). First, they carried out a modern edu- sents the choice resulting in an inferior equi- cation campaign, which explained that the librium (footbinding); it starts at an arbitrary rest of the world did not bind women's feet- zero value on the left and rises to the right. L AMERICAN SOCIOLOGICAL REVIEW

slopes down to the left because the greater Next, notice where L intersects R, labeled the proportion of people who footbind, the t. This is the tipping point. Any number of less is footbinding a positional advantage in choosers to the left of this point drives the securing marriage. A second curve, R, repre- marriage market (with the exception of any sents the choice resulting in a superior equi- organized k) to the equilibrium of all foot- librium (natural feet); it starts below L on the binding. Any number of choosers to the right left, intersects L, and ends above L on the of this point drives individuals to the equi- right. R slopes up to the right because the librium of all natural feet. If we get just to greater the number of potential marriage part- the right of the intersection t, we will tip all ners, the better the individual's marriage the way to the superior equilibrium on the match.8 The point n/3 along the horizontal right vertical axis. axis would represent the value to individual There is some minimum number k of us n+l of choosing L or R when one-third of the who benefit from getting together on the others choose R and two-thirds of them practice of natural feet, even at the cost of choose L. There are two equilibria: the infe- fewer potential marriage partners to choose rior, where everyone chooses L (where L from. Suppose that it's fairly easy to get the meets the left vertical axis), and the superior, requisite k together, but much harder to orga- where everyone chooses R (where R meets nize the t people we need to tip the equilib- the right vertical axis). If we are stuck.at L, rium (the situation is such that k is small and the footbinding equilibrium, how can we get t is large). What else can we do? We should to R, the natural feet equilibrium? Schelling recruit more members to our antifootbinding (1978, chap. 7) calls the process "getting over association because each new member above the hump" and "beyond the intersection." k benefits every individual member of the as- The reformers' campaign stressed that sociation. Also, we should conduct an educa- China was alone in the world in binding fe- tion and propaganda campaign on the disad- male feet. This provides an escape from the vantages of footbinding and on the advan- belief trap, a realization that the superior tages of natural feet. To the extent our cam- convention R (right) is available. Next, if ev- paign is successful, people come to value erybody is choosing L (left), footbinding, footbinding less and natural feet more, even then notice in the diagram that there is some though they are still stuck at the inferior equi- number k of us who would be better off librium: The L curve shifts down and right to choosing natural feet. The critical number k L', the R curve shifts up and left to R'. One is reached where the rising R curve attains result of our information campaign is a new the elevation of the left extremity of the L k', smaller than the old k, so fewer people are curve (an arbitrary zero in the diagram). needed for an effective antifootbinding asso- Thus, if k or more of us can organize into an ciation. Another result is a new t', smaller antifootbinding association, pledging not to than the old t, so fewer people are needed to footbind daughters nor let sons marry tip the marriage market into the superior equi- footbound women, we are better off; beyond librium. The way I have constructed the dia- this threshold, the advantages of natural feet gram, t'is less thank; thus, we are beyond the outweigh the disadvantages of a smaller intersection and so achieve the superior equi- mate-selection pool. The initial k formations librium. To reiterate: First, k organizes; sec- were enforced by public pledge within ond, k propagandizes; third, if the propaganda church fellowships, and the proven technique is successful the tipping point t'becomes less then was borrowed and spread by nonchurch than the k already organized; thus, fourth, the reformers. Reneging on a public pledge is entire marriage market tips to the superior damaging to family honor, and any tempta- equilibrium. tion to renege disappears upon reaching k: I have given a temporal interpretation to Getting to k is getting over the hump. Figure 2. The diagram can also be interpreted to represent heterogeneity. R' and L' could ROSSand Sotomayor (1990) demonstrate that, represent the preferences of those who have for any individual, a larger marriage pool is never come to believe that footbinding is worse than worse and may (often) be better than a smaller the average person thinks it is, while R and L marriage pool. represent the beliefs of the average person. A ENDING FOOTBINDING AND INFIBULATION small cadre of fanatics, who think that foot- the R curve up and left so that now the left binding is extremely evil, can organize at the beginning of the R' curve is above the left smaller k', and their successful proselyti- beginning of the L'curve. After effective edu- zation can then accrete k'and erode t. Also, cation the population spontaneously chooses marriage markets tend to be local but over- all R. This is an optimistic analysis, however. lapping. Suppose that some local marriage It is like an outsider coming along and say- markets complete their tipping to natural feet. ing, "Don't you people know that sitting is Their overlapping membership in other un- better than standing?" To which the insiders tipped local marriage markets may then com- can reply, "Maybe, maybe not, but we're all pose sufficient k and t to tip some of those standers around here, and sitting wouldn't other markets, and so on. The global collec- work." To the extent that people depend on tion of overlapping local marriage markets others for the evaluation of information, a sort will follow a logic of change similar to that of suboptimal belief equilibrium might per- of a local marriage market. sist. Practically, reforming associations might An analogy may ease comprehension of still be recommended because of their fre- the model. For originating circumstances, quency and demonstration effects (whereby a imagine a seated audience where the tallest noticeable number of people make the change people have grabbed the front row. The view without suffering harm). of the tallest (extreme polygynists) in the The natural-foot movement was identified front row is obscured by being too close to with liberal modernizers and women's rights the elevated stage, so they stand (footbind). advocates and proceeded in the years of Thus the second, third, and all the rest of change culminating in the Revolution of the rows must stand to regain their views of 191 1. It would be reasonable to consider the the stage. The front row is better off, but ev- reform as part of a wave of modernization. eryone else is worse off because their view Urban economic development encourages is no better than before, but now they incur migration from the countryside and provides the cost of standing. (Behind the first row alternative opportunities of support for wo- the advantage of tallness is accentuated by men (and young men). Each effect weakens standing, but the tall still would rather sit if the traditional family, but strengthens the in- enough other people would do so). For dependence and bargaining power of women, maintaining circumstances represented in and so But marriage modernization is Figure 2, imagine that over time the tallest standardly supposed to have occurred after drift away from the audience and that the 1949, while even dissidents from the stan- ease of sitting (natural feet) is forgotten. dard view trace it no further back than the Standing is now entrenched as the conven- 1930s (Zang 1993). Zang's 1982-1983 urban tion. Visitors tell people that elsewhere au- survey data demonstrate modernization phe- diences sit. People begin to think that sitting nomena: 54 percent parental arrangement of might be better, but only if enough other mate choice for those married in 1900-1938, people sit; any one person sitting alone gets 31 percent in 1946-1949, 20 percent in no view of the stage (reproductive death). If 1950-1953, and 1 percent in 1977-1982; and a column (k) can be organized to sit, its median bride age increased, proportion of members suffer a poor view of the stage but nuclear families increased slightly, and pro- are compensated by the ease of sitting. The portion of stem families and joint families members of k then have two incentives, to decreased slightly from 1900-1938 to 1977- recruit the contiguous columns and to in- 1982. form everyone that sitting is better and that Contrast the relative lateness, gradualness, standing is worse than people thought. and incompleteness of those urban family As I have drawn it, both the reforming as- modernization trends with the earliness, sud- sociation and the education campaign were denness, and completeness of the end of necessary for the reform. In an easier situa- footbinding in urban China (1895-1912). I tion, education alone might suffice. New in- believe that exposure to a specific superior formation shows that L is extremely disad- vantageous and that R is extremely advanta- %n Chinese marriage modernization gener- geous; the L curve shifts down and right and ally, see Croll (1981, chap.7). AMERICAN SOCIOLOGICAL REVIEW alien convention and the specific innovation perous Egyptians around them, just as immi- of the 1895 organization and education cam- grant Egyptian Copts adopted infibulation in paign, rather than any generalized moderniz- urban Sudan (Cloudsley 1983). Second, as ing trend, were responsible for the reform.I0 we have seen, in the Sudan modernization is Doolittle (1865) reported that early modern- accompanied by an increase in both the de- ization was intensifying the practice: "It is mographic expanse and the physical intensity believed that, were it not for the poverty of of infibulation; similarly, in Somalia infibu- the people, all the females would in a gen- lation has increased with development. eration or two have small feet" (p. 201). Why The success of the tactics of the antifoot- would footbinding come to an end just as it binding societies supports the hypothesis that became more affordable? footbinding was a self-enforcing convention. Kennedy (1970) and Hayes (1975) predict What evidence is there that infibulation is as that infibulation will come to an end with in- well? Educated Sudanese men and women creasing modernization. I have two objec- interviewed by Lightfoot-Klein (1989:99) tions tdthe modernization hypothesis. First, blamed the other sex for perpetuating reinfi- Kennedy's infibulating Nubians who began bulation. Similarly, educated respondents in to circulate in Egyptian society did not aban- Somalia told Grassivaro-Gallo and Abdis- don FGM but merely adopted the clitoridec- amed (1985:317) that their spouses were re- tomy of the more numerous and more pros- sponsible for inflicting infibulation on their daughters. Their data also show that the type lo Shortly before this article went to press, and frequency of FGM are predicted by re- Gates (forthcoming) provided me with some of gion of present residence, not by the birth- her work in progress: Original survey data and a theory on the differential distribution and disap- place of the mother or father; this finding pearance of footbinding in China. Some of her suggests that local marriage-market calcula- observations support parts of my argument, for tions prevail over simple conformist trans- example, that footbinding was necessary to a mission. In the northern Sudan village on the proper marriage and that wherever footbinding Nile studied by Boddy (1982:685), the local ended, it ended rapidly. Gates's respondents al- marriage convention had shifted from infibu- most never report having received any direct lation to excision around 1969. This was dis- antifootbinding propaganda, however, and thus closed by the existence of a cohort of unmar- she believes it is unlikely that the reforming as- ried infibulated women whose younger, ex- sociations had much to do with cessation. Gates argues, therefore, that the convention hypothesis cised sisters and cousins had married. is unsupported, except perhaps among elite or ur- Abdalla's (1982) survey in 1980 of 70 So- ban populations. Gates shows that the local pro- mali female and 40 male university students portion of women bound was higher where the revealed that 60 percent of the women and work available to females could be done with 58 percent of the men believed that FGM bound feet (e.g., household textile production). should be abolished, although 66 percent of Footbinding declined, according to Gates, as the women and 50 percent of men planned to cheaper products of modern industry displaced mutilate their daughters. Thus a majority the products of female household labor. (acting collectively) would abolish the prac- My view is that Gates's data and parts of her tice, while a majority (acting individually) theory complement rather than contradict the con- vention hypothesis. There is evidence in her work would inflict it on their daughters. This is a for indirect influence of reforming associations sure sign of being trapped in an inferior con- and elite practices on the rural masses. The con- vention. As Abdalla (1982:94-95) puts it, vention hypothesis recognizes the trade-off be- "No one dares to be the first to abandon it." tween higher marital status and family need for heavy labor that Gates confirms was explicitly discussed in families. Gates takes note of the pe- CONCLUSIONS AND culiarity of the permanent disappearance of RECOMMENDATIONS footbinding as compared to uneven changes in Colonial efforts to outlaw FGM in the Sudan other gender customs. Finally, the fact that FGM is less of a hindrance to female labor than was and Kenya sparked political reaction. More footbinding may have implications for reform recent international and national efforts to policy. Full discussion of these intriguing issues discourage or prohibit it have had no appar- must await another occasion. ent effect. Only two instances of recession ENDING FOOTBINDING AND INFIBULATION are known, according to Lightfoot-Klein of FGM, and such judgment should be con- (1989:SO-51); both are doubtful on closer veyed tactfully to practicing populations. examination (see Brooks 1995; Megafu Some African nationalists and cultural rela- 1983). Lightfoot-Klein reports that infibula- tivists claim that FGM is a matter to be tion is being introduced in parts of Uganda worked out by African women." This is so, as an imagined return to African traditions. but in the Chinese case, indigenous reform- Legal prohibition has not worked. The con- ers were more helped by the exposure to an sensus approach to eradication is to publicize alternative convention and by international the objectively bad consequences of female condemnation of footbinding as wrong than mutilation (initiatives are summarized in they were hindered by the obtuse xenopho- Hosken 1993:390418). bia of some of the Western reformers. The Infibulation is an absolutely necessary pre- followers of mutilation are good people who condition to marriage in the Sudan, accord- love their children; any campaign that insinu- ing to Gruenbaum (1982), who continues ates otherwise is doomed to provoke defen- that policy recommendations "have seldom sive reaction. Because these parents love recognized the significance of linkage be- their children, they will be motivated to tween the operations and the social goal of change, once they learn of the bad health maintaining the reputations and marriage- consequences of FGM, and once a way to ability of daughters in a strongly patriarchal change is found. FGM will end sooner or society" (p. 5). Testimony of an Islamic later; better that it ends sooner than later.I2 scholar that infibulation is not religiously Outside opposition to mutilation, whether commanded, or promoting greater awareness from various feminist groups or now by of medical hazards, "would not be sufficient American human rights policy, will have an for a mother to risk her daughter's marriage- eventual effect, and thus is worthwhile. ability" (p. 7) because of the importance of Third, associations of parents who pledge marriage and children. Gruenbaum recom- not to infibulate their daughters nor let their mends increased employment and educa- sons marry infibulated women may, as in tional opportunities to promote the indepen- China, decisively lead to change. Tribal fa- dence of women. cial scarring is common in the Sudan, but is The convention hypothesis predicts that if in disrepute among westernized youths in a convention ends, it will end quickly. The the capital. Why then is infibulation not in sudden end of footbinding in response to abo- disrepute? One of Lightfoot-Klein's (1989: lition campaigns supports the convention hy- 49, 116) respondents explained that if you pothesis and predicts that equivalent African give up scarring you lose your beauty, but to campaigns could help end FGM. The conven- give up infibulation is to become shameful. tion hypothesis recommends three types of Losing your beauty is sad, but losing your action. First, let people know the physiologi- honor is unthinkable. Scarring is on the cal facts, explaining to women and men the wane, I think, because it is part of an "aver- advantages of natural genitals and the disad- age opinion game" (Van Huyck, Battalio, vantages of mutilation. An education cam- and Beil 1991): Faces are visible, and inter- paign is a banally obvious measure. What is national images enlarge the reference group. not so obvious is that an ultimately success- Infibulation is invisible, however, and is ful education campaign may not have any also a matter of honor, a positional good behavioral effects until the moment of con- unamenable to compromise. Honor depends vention shift. Knowing that we are at an infe- rior coordination equilibrium is necessary but not sufficient for change. We may each come See, for example, Dawit and Mekuria to believe that sitting is better than standing, (1993): "A media campalgn in the West will not but only if enough others make the change at stop genital mutilation" (p. A27). '? One self-fulfilling reform tactic might be to the same time. Reformers and their funders warn parents to wait because future convention should therefore measure progress by attitu- shifts may make their irreversibly mutilated dinal rather than behavioral change. daughters unmarriageable. The failure of such a Second, international public opinion tactic would deepen the convention trap, how- should deplore the bad health consequences ever. 1016 AMERICAN SOCIOLOGICAL REVIEW on local marriage markets, not on interna- the Nineteenth Century. London. England: tional images. Frank Cass. Pledge associations might be even more Cloudsley, Anne. 1983. Womerz of Orndurmarz: important in Africa than in China because the Life, Love and the Cult of Virginity. New York: St. Martin's. public invisibility of FGM makes it harder Croll, Elisabeth. 1981. The Politics of Marriage for people to judge any change in the expec- in Contemporary Chirza. Cambridge, England: tations of others; they are less sure of where Cambridge University Press. they are on the Schelling diagram (Figure 2) Daly, Mary. 1978. Gyn/Ecology. Boston, MA: I drew to describe the end of footbinding. Beacon Press. The technique of pledge associations is not Dareer, Asma El. 1982. Womarz, Why do You reported in the FGM literature, and explain- Weep?: Circumcisiorz and Its Consequerzces. ing its potential is intended as the central London, England: Zed Press. contribution of this article. Dawit, Seble and Salem Mekuria. 1993. "The West Just Doesn't Get It." The New York Gerry Mackie is Jurzior Research Fellow in Poli- Times, December 7, p. A27. tics at St. John's College, University of 0.xford. Dickemann, Mildred. 1979. "The Ecology of He has recently published jourrzal articles on Mating Systems in Hypergynous Dowry Soci- workplace democracy arzd orz international mi- eties." Social Science Information 18: 163-95. gration. He is completing his Ph.D. in political . 1981. "Paternal Confidence and Dowry science from the University of Chicago. His rlis- Competition: A Biocultural Analysis of sertation criticizes social choice arzd corzsensual Purdah." Pp. 417-38 in Natural Selection and models of democratic deliberatiorz, arzd applies Social Behavior, edited by R.D. Alexander and the parallel constraint satisfaction model to po- D. W. Tinkle. New York: Chiron Press. litical debate. He plans research on the micro- Doolittle, Justus. 1865. Social Life ofthe Chinese. foundations of political cultures arzd orz empiri- New York: Harper and Brothers. cal approaches to social ethics. Drucker, Alison R. 1981. "The Influence of West- ern Women on the Anti-Footbinding Move- ment 1840-191 1." Historical Reflections 8: REFERENCES 179-99. Abdalla, Raqiya Haji Dualeh. 1982. Sisters in Af- Ebrey, Patricia. 1990. "Women, Marriage, and fliction: Circumcision arzd lnfib~dationof the Family in Chinese History." Pp. 197-223 Women in Africa. London, England: Zed Press. in Heritage of Chirza, edited by P.S. Ropp. Ber- Agatharchides. 1989. Orz the Erythraearz Sea. keley, CA: University of California Press. Translated by S.M. Burstein. London, England: 1991. "Shifts in Marriage Finance from The Hakluyt Society. the Sixth to the Thirteenth Century." Pp. 97- Assaad, Marie Bassili. 1980. "Female Circumci- 132 in Marriage arzd Inequality in Chirzese So- sion in Egypt: Social Implications, Current Re- ciety, edited by R.S. Watson and P.B. Ebrey. search, and Prospects for Change." Studies in Berkeley, CA: University of California Press. Family Planning 11:3-16. Elster, Jon. 1986. "Introduction.'' Pp. 1-33 in Ra- Beachey, R. W. 1976. The Slave Trade of East- tional Choice, edited by J. Elster. New York: ern Africa. New York: Barnes and Noble. New York University Press. Bedi, Rahul. 1994. "Pakistan: Religious Right Fairbank, John. 1992. Chirza: A New History. Puts Bhutto in a Spot." Interpress Service, May Cambridge, MA: Harvard University Press. 27. Fairbank, John and Albert Feuerwerker, eds. Blake, C. Fred. 1994. "Footbinding in Neo-Con- 1986. The Cambridge History of China, vol. fucian China and the Appropriation of Female 13. Cambridge, England: Cambridge Univer- Labor." Sigrzs 19:676-712. sity Press. Boddy, Janice. 1982. "Womb as Oasis: Pharaonic Fairbank, John and Edwin 0. Reischauer. 1978. Circumcision in Rural Northern Sudan." Chirza. Boston, MA: Houghton-Mifflin. Americarz Ethnologist 9:682-98. Freeman-Grenville, G.S.P., ed. 1962. The East . 1991. "Body Politics: Continuing the Africnrz Coast: Select Documerztsfrom the First Anti-Circumcision Crusade." Medical Arzthro- to the Earlier Nineteerzth Century. Oxford, En- pology Quarterly n.s.5:15-17. gland: Clarendon Press. Bonaparte, Marie. 1953. Female Sexuality. New Freud, Sigmund. 1961. "Fetishism." Pp. 149-57 York: International Universities Press. in The Stanriarri Edition of the Complete Psy- Brooks, Geraldine. 1995. Nine Parts of Desire. chological Works of S. Freud, vol. 21, edited New York: Anchor Books. and translated by J. Strachey. London, En- Clarence-Smith, William Gervase, ed. 1989. TIE gland: Hogarth. Ecorzomics ofthe lrzdiarz Ocean Slave Trade in Gamble, Sidney D. 1943. "The Disappearance of http://www.jstor.org

LINKED CITATIONS - Page 1 of 3 -

You have printed the following article: Ending Footbinding and Infibulation: A Convention Account Gerry Mackie American Sociological Review, Vol. 61, No. 6. (Dec., 1996), pp. 999-1017. Stable URL: http://links.jstor.org/sici?sici=0003-1224%28199612%2961%3A6%3C999%3AEFAIAC%3E2.0.CO%3B2-0

This article references the following linked citations. If you are trying to access articles from an off-campus location, you may be required to first logon via your library web site to access JSTOR. Please visit your library's website or contact a librarian to learn about options for remote access to JSTOR.

References

Female Circumcision in Egypt: Social Implications, Current Research, and Prospects for Change Marie Bassili Assaad Studies in Family Planning, Vol. 11, No. 1. (Jan., 1980), pp. 3-16. Stable URL: http://links.jstor.org/sici?sici=0039-3665%28198001%2911%3A1%3C3%3AFCIESI%3E2.0.CO%3B2-H

Foot-Binding in Neo-Confucian China and the Appropriation of Female Labor C. Fred Blake Signs, Vol. 19, No. 3. (Spring, 1994), pp. 676-712. Stable URL: http://links.jstor.org/sici?sici=0097-9740%28199421%2919%3A3%3C676%3AFINCAT%3E2.0.CO%3B2-X

Womb as Oasis: The Symbolic Context of Pharaonic Circumcision in Rural Northern Sudan Janice Boddy American Ethnologist, Vol. 9, No. 4, Symbolism and Cognition II. (Nov., 1982), pp. 682-698. Stable URL: http://links.jstor.org/sici?sici=0094-0496%28198211%299%3A4%3C682%3AWAOTSC%3E2.0.CO%3B2-7

Body Politics: Continuing the Anticircumcision Crusade Janice Boddy Medical Anthropology Quarterly, New Series, Vol. 5, No. 1. (Mar., 1991), pp. 15-17. Stable URL: http://links.jstor.org/sici?sici=0745-5194%28199103%292%3A5%3A1%3C15%3ABPCTAC%3E2.0.CO%3B2-Y http://www.jstor.org

LINKED CITATIONS - Page 2 of 3 -

The Disappearance of Foot-Binding in Tinghsien Sidney D. Gamble The American Journal of Sociology, Vol. 49, No. 2. (Sep., 1943), pp. 181-183. Stable URL: http://links.jstor.org/sici?sici=0002-9602%28194309%2949%3A2%3C181%3ATDOFIT%3E2.0.CO%3B2-%23

The Commoditization of Chinese Women Hill Gates Signs, Vol. 14, No. 4, Common Grounds and Crossroads: Race, Ethnicity, and Class in Women's Lives. (Summer, 1989), pp. 799-832. Stable URL: http://links.jstor.org/sici?sici=0097-9740%28198922%2914%3A4%3C799%3ATCOCW%3E2.0.CO%3B2-O

Female Genital Mutilation, Fertility Control, Women's Roles, and the Patrilineage in Modern Sudan: A Functional Analysis Rose Oldfield Hayes American Ethnologist, Vol. 2, No. 4, Sex Roles in Cross-Cultural Perspective. (Nov., 1975), pp. 617-633. Stable URL: http://links.jstor.org/sici?sici=0094-0496%28197511%292%3A4%3C617%3AFGMFCW%3E2.0.CO%3B2-T

Circumcision and Excision in Egyptian Nubia John G. Kennedy Man, New Series, Vol. 5, No. 2. (Jun., 1970), pp. 175-191. Stable URL: http://links.jstor.org/sici?sici=0025-1496%28197006%292%3A5%3A2%3C175%3ACAEIEN%3E2.0.CO%3B2-B

Female Circumcision in Africa: An Overview Leonard J. Kouba; Judith Muasher African Studies Review, Vol. 28, No. 1. (Mar., 1985), pp. 95-110. Stable URL: http://links.jstor.org/sici?sici=0002-0206%28198503%2928%3A1%3C95%3AFCIAAO%3E2.0.CO%3B2-B

The Virgin and the State Sherry B. Ortner Feminist Studies, Vol. 4, No. 3. (Oct., 1978), pp. 19-35. Stable URL: http://links.jstor.org/sici?sici=0046-3663%28197810%294%3A3%3C19%3ATVATS%3E2.0.CO%3B2-5 http://www.jstor.org

LINKED CITATIONS - Page 3 of 3 -

Strategic Uncertainty, Equilibrium Selection, and Coordination Failure in Average Opinion Games John B. Van Huyck; Raymond C. Battalio; Richard O. Beil The Quarterly Journal of Economics, Vol. 106, No. 3. (Aug., 1991), pp. 885-910. Stable URL: http://links.jstor.org/sici?sici=0033-5533%28199108%29106%3A3%3C885%3ASUESAC%3E2.0.CO%3B2-0