1Milk is a unique universal. It is the only common food produced in humans, for humans. It is the only food every mother makes, the only food every person consumes, at the beginning of their life. Since breast-feeding is so central a part of infant-rearing it is unsurprising that a host of food historians (e.g. Hardyment 1983; Fildes 1986), medical researchers (e.g. Wickes 1953; Hamosh and Goldman 1986), anthropologists (e.g. Raphael 1973; Raphael and Davis 1985; Durham 1991) and feminists (e.g. Coward 1989; Carter 1995; Draper 1996) have studied its nutritional pros and cons, its cross-cultural variability, its patriarchal politics, the evolution of Western attitudes towards it, and the polemic about supplying powdered milk to mothers in the Third World. Yet in the midst of all this work about milk in the West and beyond, what is surprising is that one distinctive and historically significant topic appears to have been neglected: the establishment of a indissoluble 'milk-tie' between infants of different parents who suck at the same breast, and the important social, political, and economic consequences which follow from it. The aims of this paper are (1) to carry out the first (to my knowledge) systematic survey of the literature on the milk-tie, and (2) to try to derive some generalizations about this practice.
2The milk-tie is not a universal phenomenon: in some cultures women never give suck to others' infants, regarding the very idea as distasteful, unnatural or even dangerous to the life of the woman who attempted it (Fildes 1988: 265). Among the Baganda of East Africa, for instance, mothers refuse to allow surplus, expressed milk to be fed to other babies: those Baganda infants whose mothers are unable to feed them die (Jelliffe 1962: 22).
3But the milk-tie is a strikingly widespread phenomenon, practised by peoples from the Balkans to Bengal, from Marrakech to Mandalay. It is all the more disappointing then that the vast majority of data available on this custom consists of mere snippets of information collected by travellers, doctors and ethnographers. All that these here-gathered gobbets give us is some idea of the historical or contemporary extent of the practice.
4Starting with the westernmost examples first, Dunn writes of the campaign waged in the later eighteenth century by one Moroccan tribe against another. When the latter finally surrendered, the former did not subordinate them but
5United with them in a pact known in the region as tafargant (interdiction). The initiating ritual involved the exchange of milk from lactating mothers... Tafargant stipulated not only peaceful relations and mutual aid but also strict prohibition on marriage between the two tribes. This taboo implied symbolic brotherhood between them, but it may also have had the practical function of eliminating one major cause of tension (Dunn 1973: 97.)
6Maher (1984: 107) states that, in the Middle Atlas region of Morocco, a mother must have the permission of her husband before she nurses another's child. All the examples of milk-kin that Maher came across obtained between a mother and relatives of her husband. None obtained between a mother and her relatives. This suggests that each husband was careful to ensure that his wife's wet nursing only served to strengthen the links between his immediate family and his kin.
7Information on milk-kinship in the Near East is tantalizingly brief. Hammel (1968: 30) and Filipovic (1960) both mention that, in the Balkans, siblingship can be created between two children if they are suckled by the same mother, but neither author gives further details (but see Doja 1999). In traditional Turkey, a mother's child and her nurseling become 'milk-siblings' and thus cannot intermarry (Davis 1977: 237-38).
8There are more reports about milk-kinship in the Middle East, all of which state that while the practice is still maintained today, the number of children so nursed has declined greatly in recent decades. On the island of Socotra, 550 miles east of Aden, infants are nursed by women from outside the father's clan; if both parents are from different tribes, 'they are placed with the tribes or clans that stand closer to that of the mother' (Naumkin 1993: 282). In the southern Egyptian village studied by Ammar (1954: 102), people remembered who had wet nursed whom so that the local marriage taboos would not subsequently be broken. Wet nursing rarely occurred among close relatives, where future marriage of the children was anticipated. Ammar underlines the importance of the milk-tie by reporting the locally-repeated tale of Harun al-Rashid, an eighth-century caliph of Baghdad, who decided to kill one of his viziers and to punish the man's family because he believed they were plotting to depose him. Though the vizier's mother appealed to Harun, whom she had nursed as a child, he went ahead with his decision anyway. One chronicler considered this, 'An example of tyranny that overrides one's duties towards his breast-feeder who is like mother' (Ammar ibid.: 99).
9Among the urban elite of Saudi Arabia, Altorki (1980: 240-41) argues, milk-kinship was created for two primary reasons:
Domestic Convenience. Since a woman was (and still to a great extent is) compelled to veil before any man other than a close kinsman, whom she could not marry, a man might ask his slave woman to nurse his daughters, so that they would not later have to veil to her son(s).
Forestalling Potential Marriages. Given the local preference for the intermarriage of cousins and the prevalence of extended households composed of the families of adult brothers, a jealous man could prevent an undesirable marriage by having his wife nurse the children of the envied brother.
10The members of the Iranian landowning elite with whom Khatib-Chahidi (1992: 119) spoke, thought both of these reasons highly bizarre. For the sake of domestic convenience, these Iranians did not have to establish milk-kinship with their servants. Instead they got around the problem by mean of temporary fictive marriage contracts (a traditional device) between the servant and one of the children of the household head. If they did have to call on the services of a wet nurse, their only concern was to avoid a choice of milk mother which could affect future marriage arrangements.
11According to Granqvist (1947: 252), who lived among Palestinian Bedouin in the 1920s, local Muslim infants could be nursed by Muslim, Christian or Jewish mothers. By the same token, a Muslim could nurse a Christian, and the consequences would be understood in local terms. In 1881 in Damascus, the milk of the British Consul's wife failed. As her son records:
12A girl was duly produced (by an astute sheikh of the 'Anizah group) and according to my mother's testimony I drank her milk for several weeks. This is in the eyes of the Badawin entitles me to a certain 'blood affinity' with the 'Anizah; for to drink a woman's milk in the desert is to become a child of the foster mother. This fact has been of assistance to me in my dealings with the Badawin' (Dickson 1952: 7).
13Going beyond the Middle East, Lyall (1882: 221) stated that among the Rajputs of India chiefs chose wet nurses for their children from a well-known pastoral tribe. The nurse's family held a recognized hereditary status of 'kinship by milk' and when the once-nursed man finally assumed the chieftancy, his milk-brothers often attained much influence and position at his court. Also, in the Muslim Thai village studied by Hanks (1963: 128), if a wet nurse was employed, her child had to be of the same sex as the child to be nursed. This was not only because the milk intended for a child of the opposite sex was believed to cause disease and even death in the nurseling, but because incest might occur otherwise when the milk-siblings became adults.
14Almost all the reports about milk-kinship cited so far are overly pithy and fail to site the phenomenon adequately in its various contexts. To gain a more exact idea of the ends to which milk-kinship was put, it is necessary for us to look at two further, more intricate examples, from the Caucasus and northern Pakistan.
15For a far-sighted father, marrying off his children strategically was one way to extend, in the right direction, his family's network of kin and relatives; using his children to create milk-ties with people they could not marry was another way. In pre-Revolutionary Russia, Christian villagers of northern Georgia and their Muslim counterparts over the Caucasian mountains used to sell their produce to one another by making hazardous treks through the high passes (Dragadze 1987). Since cross-creed marriages were prohibited, the best way for a pair of already close trading partners to strengthen their relationship was for one of them to nurse and bring up a child of the other. In order not to appear suspicious, the family of the 'adopted' child rarely made the trek over the mountains to visit. The child returned home before entering adolescence so that it could not be thought the 'adoptee' family were exploiting his, or her, labour. After all, the ultimate aim of this tie was not for the partners to make a balanced exchange but to ensure the survival of the partnership across generations, for the milk-siblings would speak the same language, know the same traditions, and help each other with transport, trade and hospitality.
16In Georgia today, milk-siblingship is seen as an important, intimate relationship. The parents exchange favours, gifts, and visits, and it is expected that the milk-siblings will be on close terms with each other throughout life. Neither they nor their children can intermarry.
17The most radical, elaborate use of the milk tie for political purposes occurred in the tiny kingdom of Chitral, northern Pakistan (Biddulph 1880: 82-3; Jettmar n.d.) In this markedly hierarchical state of royalty, aristocrats, commoners and slaves, which survived as a functioning polity until the 1950s, the rulers used to give their children away at birth to be fostered by noblewomen. These women would share the nursing of the children with as many of their female dependents as possible. In this way the welfare of a royal child involved a large number of people. Foster-parents of a princess received land and other gifts on presenting her at court when she reached seven; on her marriage, one of her milk siblings would accompany her to her new home. Those who had reared a prince would not only receive similar gifts on his handing-over but stood to gain even more: if at an early age he were made the governor of a province, his milk brothers became his main advisors and ran his executive. A prince was usually made governor of the area where he had been fostered so that he would be fully acquainted with the locals and would consider himself, to a certain extent, one of them.
18Nobles, who headed tribes or clans, followed the royal example by farming out all their infants to families of lower status, who in return enjoyed particular privileges and were excused the payment of certain tithes. As a contemporary visitor to the region observed, 'The foster parents continually show great devotion and abnegation to this cuckoo in their nest, and their own children suffer. I have known cases where the foster fathers have spent all their subsistence on some useless brat of the aristocratic class' (Schomberg 1938: 225). Again like their superiors, nobles might have their babies passed around a whole village or local tribe. Thus an infantile aristocrat suckled by several dozen different women might grow up to have fifty milk-mothers, fifty-milk fathers and hundreds of milk-siblings, who would support and protect him if need be.
19It seems like a cunningly simple political system, a state ruled at all levels by massively extended families linked by blood, marriage and milk. Trouble is, the system had a fatal flaw: no regulated accession to the throne, whether by seniority or achievement, leading to very bloody conflicts between the parties of different pretenders. When a king died, aspiring princes rapidly mobilized all their kin. Pretenders reminded nobles of their milk ties and nobles did the same to their relatives of lower status. During these interregna, aristocrats acted as the bodyguard of their pretender; commoners and slaves formed the rank and file in the forces of their noble milk-brothers. Those groups who failed their prince, as did sometimes happen, were stigmatized inedibly. Generally, however, almost all groups gave unswerving loyalty and effective protection to their candidate for the crown.
20The stakes were high. If a pretender was successful, he gave high offices, land and other gifts to his aristocratic milk-brothers, who passed on some of this largesse to their own supportive milk kin. But if his bid for the throne failed and he was not killed in the process, he and all his noble backers were banished immediately. This enforced exile was unavoidable for it was feared, rightly, that otherwise a pretender's milk brothers would strive to assassinate the new incumbent. They fled to the mountainous areas on the margins of the monarchy, which the king's men could never properly bring under control and which acted traditionally as places of refuge for his defeated brothers and their milk kin. There the vanquished prince and his band made a living by obtaining help from the local peoples, or violently subjecting them to his own will and establishing his own mini-state. Those aristocrats who could not easily accompany them on their flight (old people and mothers with infants) lost their homes and land and were forced to live in misery. Commoners of the defeated party might be enslaved to victorious nobles or sold to traders.
21It is easy to see these conflicts from above, as primarily the bloody consequences of competition at court. But they may also be viewed from below, as battles between tribes for power, with the princes as valued assets or risky liabilities. Thus powerful tribes, headed by ambitious nobles, far from being the puppets of milk brother princes, could become the driving force of the system, persuading their royal kin, when the moment arrived, to make a bid for the throne.
22There were also other, less risky ways for tribes and clans to exploit the system to their own advantage. Once, when the Roshte tribe incurred the wrath of the king, he confiscated their house and lands, which he handed over to others. To save their own lives, the Roshte had to flee. Only one of them, a woman with an infant, decided to remain. A friend of hers had just borne the king his first son, of which he was very proud. The friend, clutching her baby, stole out of the palace and gave him to her to suckle. She immediately took the child to her former house, burst in, and proclaimed that she was now the milk mother of the prince. When its inhabitants realized what she was shouting, they had no option but to leave. The king, though enraged, could not object to the demands of his new kinswoman. He restored all rights to the Roshte (Parkes 2001).
23So far, I have given a list of instances of milk-siblingship, founded on the sucking at the same breast by genealogically unrelated infants. Yet it is important to note that, in areas where milk-kinship is already established, the symbolism of the practice may be exploited performatively in order to create structurally similar relations between adults. In these cases, the milk-kinship based on infantile breast-feeding is used as a model for the creation of a life-long tie. Granqvist (1947: 114) states that a Palestinian woman who wished to adopt a stranger boy or man, could do so by publicly putting her nipple into his mouth, saying, 'Thou art my son in God's Book, thou hast sucked from my breast'. According to Granqvist (1931: 65), 'One reason for adoption is often the fact that a woman has to be alone with a strange man for some time as on a journey, and to protect her reputation she adopts him'.
24The existence of milk-relations may also be used by women to achieve some degree of gender equality. Patricia Daugherty, who worked among the Yorul community in the Taurus mountains above the south Turkish town of Alanya, has reported that its women may well take advantage of the milk-tie in order to gain some measure of control over the otherwise male-dominated system of arranged marriages (Daugherty pers comm.)
25Furthermore, the creation of a milk-tie may be exploited in order to shore up a reputation that is on the point of being ruined. In Georgia, if a husband thought his wife unfaithful, he called the suspected paramour to his house, bared his wife's right breast, put salt on it, and asked the man to kiss it. The suspect had no option: if he kissed it, he would be milk-tied to the woman for life and so could not, under threat of punitive retribution, have sex with her; if he did not kiss it, he incriminated himself and faced punitive retribution. Once the deed was done, the husband would address the couple: 'Man, behold your mother. Woman, behold your son'. He could now rest assured; his wife and new milk son-in-law could meet openly without fear of raising any suspicion, for incest was out of the question.
26A similar use of the milk-tie was employed among the Afghans of the Hindu Kush, as noted by a British colonial officer:
27In cases where conclusive proof (of adultery) is wanting, and which are brought for settlement before the ruler, guarantee is taken for the future by the accused placing his lips to the woman's breast. She thenceforth is regarded as his foster-mother, and no other relations but those of mother and son can exist between them. So sacred is the tie thus established esteemed, that it has never been known to be broken (Biddulph 1880: 77).
28One British colonial officer, working in the same general area, who had agreed to participate in a ceremony of elective kinship with a local youth, found that the ritual entailed more than he had expected:
29A goat was procured, quickly killed and its kidneys were removed. These were cooked at a fire and cut into morsels by an officiating (local) who fed us both... on the point of a knife. At short intervals we had to turn our heads to one another and go through the (sickly) motion of kissing with our lips a foot or so apart. But - the surprise was in reserve... My coat and shirt were suddenly torn open and some butter was placed on my left breast, to which the youth applied his lips with the greatest energy and earnestness! I jumped up as if shot - but the thing was over! (Robertson 1896: 30-31, cited in Parkes n.d.: 2).
30What the official had not anticipated was that the ceremony would turn into a rite of submission, whereby the subordinate places oneself in a state of symbolic dependency, like that of a child to a mother.
31* * *
32Our list done, we can attempt to generalize about the nature of the milk-tie.
33The first generalization we can make is that the extent of the tie is greatly variable between cultures, and even at times within cultures. Some authors speak of milk-kinship as though it were exclusively a relation between a pair of milk-siblings (e.g. Simoons 1976: 316; Farb and Armelagos 1980: 96). But these representations of the practice are far too restrictive, for milk-kinship, even in the most restrictive cases, always involves more people than just the pair of siblings. In the Moroccan case studied by Maher, the milk-kin are a close circle of the husband's relatives. In contrast, in the eighteenth-century Moroccan example provided by Dunn, two whole tribes entered into a single milk-tie. In Georgia, a pair of families are united by each tie while in Chitral a consciously created network of such ties could unite whole regions of the former kingdom.
34In Islam, even though the Koran and the hadith (pronouncements of the Prophet) are explicit and fulsome about whom milk-kin may not marry, Muslim jurisprudentialists have debated, in an at times tortuous manner, the definition of what exactly constituted a milk-tie (Altorki 1980). To them, the tie was not knotted by a single feeding but determined by a complex formula. Learned scholars crossed pens on such matters as exactly how many feeds, and of what volume, were necessary to knot the milk tie. Estimates ranged from one to ten feedings to a few drops on separate occasions. Some jurists contended that an infant's involuntary pause while sucking marked the end of one feed and the beginning of another; others would only recognize this break if the mother had deliberately interrupted her suckling. Some argued it was sufficient for a child to drink five times from a pot containing a woman's milk; their opponents counter-argued a child had only to drink from a pot containing a woman's milk collected on five separate occasions. Further issues in this arcane debate included the nature of testimony required for proof of the relationship, and whether cheese made from a woman's milk was an acceptable substitute.
35These debates might, at first glance, seem to be a culturally particular example of the legal pedantry which jurisprudentialists of any culture might indulge in. But these debates were not exercises in Islamic hair-splitting. Exactly how long, and when, an infant sucked at a woman's breasts were relevant legal questions, as the belated discovery of a distant milk relationship between husband and wife led to mandatory divorce. Further, as Khatib-Chahidi argues (1992: 123), practising Muslims need to know to whom they are related and in what ways if they are to perform their devotional duties effectively, because, for some of them, merely looking at a potential marriage partner 'while saying their prayers would mean their devotions were nullified and had to be repeated'. Thus some of the rulings relating to milk-kinship may
36Reflect the behavioural implications of the forbidden degrees of kinship for marriage: they represent not so much an indication of whom a person may or may not marry, but those of the opposite sex with whom one may or may not act in a relaxed manner (ibid.)
37The second (and key) generalization we must make is about what the milk-tie does. Khatib-Chahidi (ibid.: 124) characterizes it as 'a means to get friends or allies'. This is far too restrictive. As a well-grounded generalization, all we can state is that the milk-tie is a means of establishing a lasting connection between two groups of people. It is a tie that in many ways is meant to be as significant, and as lasting, as those based on blood. Like human blood, human milk is an essential life-enabling substance. And by using milk (whether a nutritionally significant amount or not) to create deliberately a link of kinship, people try to naturalize the cultural. In other words, they exploit the parallels between blood and milk in order to make this fictive form of kinship appear as natural as that based on genealogy. In this context it is revealing that Waldren, who has done over two decades of research in the Majorcan village of Deia, reports that ‘an old woman recently died leaving her house to a young man she had nursed, rather than to her children or grandchildren’ (Waldren pers comm.)
38Perhaps what is most interesting about the milk-tie is the variety of purposes to which that connection can be put, and the associations we can make between certain purposes and certain forms of social organization. It is evident that, in societies where rules of intermarriage consolidate the coherence of certain sub-groups (and so isolate them in the process), the establishment of milk-ties is a means of creating lasting links with groups of people whom one is not allowed to marry into. This is clearly the case for caste-divided societies such as the Chitrali; for groups who need to maintain co-operative relations with their neighbours, such as Albanians, the Rajputs, or the tribes of eighteenth-century Morocco; and for religiously-defined groups who wish to uphold links with members of other faiths, such as the Christian villagers of northern Georgia or the Palestinian Arabs studied by Granqvist (Doja 1999, Parkes 2003). (The prompt provision by a sheikh of a local wet nurse for the British Consul's son may be interpreted in the same vein.)
39The establishment of a milk-tie may be used, not to bring people together in a mutually beneficial manner, but to increase the social distance of one group from others to the ultimate benefit of only one party. Thus, in the example of the jealous Saudi brother, he was able to exploit the consequences of creating a milk-tie in order to forestall marriage between his children and those of his envied sibling.
40The examples of the suspicious husbands and, in eighteenth-century Morocco, of the imposition of peace on a subordinate tribe illustrate a further end to which the milk-tie can be put: making others one's milk-kin as a way to control their behaviour.
41* * *
42As far as the evidence will allow us to state, the practice of milk-kinship is in decline, thanks to the increasingly widespread distribution of formula milk, to the general shift from extended rural families to nuclear urban ones, and to the gradual decline of kinship as the central model of co-operative interaction. This present decline is no reason for us to continue to ignore the practice. For milk-kinship is not just a historical oddity (though knowledge of its existence and types is important for the historical record); it is not just an ethnographic curiosity (though information about its nature and functioning is important for our understanding of the varieties of kinship). Perhaps above all, knowledge of the milk-tie and its diversities helps to illuminate the social uses to which food may be put, and the ways that the uniqueness of one food may be put to unique ends.