Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

Discussion of paper by Pollitt and Kariger


Dr. Colombo

An expert from the Institute on Brain Research in Frankfurt, who specialized in the growth and development of the brain in children, was sitting here a few months ago, and I asked him, "How long do these developments last?" I was expecting him to say for two or three years, but he said until puberty. Certainly, if there are clear differences, he said that they may show up at 12 months; but if we have to wait a long time, so many confounding factors intervene that the analysis may fail to give clear answers. As a layman, I can see that this would be a problem for the analysis. Now, I wish to give the floor to more respected colleagues. Thank you.

Dr. Lawrence

I am intrigued by the comment you made at the end that an area of enormous potential was assessing the role of breastfeeding in the formation of secure attachment in early childhood. Do you think of that as a maternal behaviour or an infant behaviour or both, and, if so, what is the mechanism that you postulate? This is particularly interesting in light of the question you asked me yesterday about the weakness of the evidence of maternal behaviour.

Dr. Pollitt

Infant attachment to the caretaker (biological or surrogate mother) is first observed at six to eight months, and it generally stems from the continuous social and emotional interactions between both members of the dyed. It is theoretically plausible that interactions leading to attachment are induced and sustained by behaviours or behavioural cues of both the infant (for example, smiling) and the caretaker (for example, scent of the nipple). Breastfeeding would provide a medium for the presence of particular behavioural interactions (for example, sucking and milk release) that would help the development of secure attachment between the infant and the caretaker.

Dr. Lawrence

Do you see this attachment in the breastfed infant as being singularly different from that of the bottlefed infant?

Dr. Pollitt

I do not know, but I suspect this is the case because bottle-feeding and breastfeeding represent different behavioural systems, including similar and distinct behaviours of both parties. For example, the organization of sucking behaviour occurs earlier in the breastfed infant than in the bottle-fed infant, and the scent of the mother's nipple is different from the scent of a nipple on a bottle. Further, whereas the composition of maternal milk changes during feeding, this does not happen in artificial feeding. These differences between bottle-feeding and breastfeeding might be independent of the social and emotional development of the child. Nevertheless, the nature of the behaviours supports the proposition that there is scientific and programmatic merit in testing the hypothesis in the field.

Dr. Garza

I'd like to explore your comment that even if we accept the validity of the present findings, with all of their limitations, the differences between breastfed and bottle-fed babies are rather modest. If we think of them as half a standard deviation, I agree with your assessment for those ranked around the mean. That is, I don't know whether half a standard deviation in either direction would have much functional significance. If, however, those differences are homogeneously distributed, they could theoretically have an enormous adverse impact on the number of functionally impaired people. Or the converse might be true: you would almost double the number of gifted people. Is there any reason to suspect that improvement would not be distributed homogeneously? Or is this such a theoretical construct that it has no merits?

Dr. Pollitt

The question refers to two separate but related issues. One is whether the alleged effects of breastfeeding are equally distributed along the entire distribution of scores, whatever cognitive measures are used. I do not have an answer, and I doubt whether at present there are relevant data to address this question. The second issue is the behavioural significance of the effects of breastfeeding on cognition. Even though the effects reported were relatively modest, they could still have significant developmental importance if they were found in different cognitive areas. Behaviourally, small benefits over a wide range of cognitive domains (for example, short- and long-term memory, attention, concept formation, and vocabulary) may be more advantageous than large benefits over a single domain (for example, short-term memory).

Dr. Garza

In relation to the comment made by Dr. Colombo, it may be useful for you to comment on the rather extraordinary finding from Guatemala that a limited intervention during gestation and the first 3 years of life was still of functional significance 15 years later, despite the fact that nothing had been done to sustain the intervention.

Dr. Pollitt

The study involved four rural communities. Pregnant and lactating women and children up to the age of seven years participated in this study from 1969 to 1976. Two of these communities received a high-energy (180 kcal x 180 ml) and high-protein (11.5 g x 180 ml) supplement called atole, and the remaining communities received a low-energy supplement (one-third of that in atole). A 1988 follow-up showed that those in the first group performed significantly better than those in the second group in a wide range of cognitive tests (for example, arithmetic, reading comprehension, and vocabulary). The data necessary to explain the mechanisms behind these effects are unavailable. It seems likely, however, that two different sets of mechanisms were involved. One is that atole made a difference in the development of particular areas in the brain or in the neurotransmitter systems that facilitated cognitive function. The other involves the benefits of atole on physical growth, neuromotor maturation, activity level, and interactions with the social and physical environment. These advantages pooled together could have increased the educational opportunities of the children and fostered cognitive development.

Dr. Woolridge

I have two questions. The first regards the two slides on alcohol and marijuana use. If I understand them correctly, they both refer to breastfeeding mothers. Are there data from those studies to show what sort of effect you would get for non-breastfeeding mothers? One may assume that use of alcohol and marijuana may affect other aspects of the mother-child relationship. The effect could lead to lower development in the children of users. For my second question, I wanted some clarification on the issue of preterm births. I think I read in the reviews by Villar and Belizán that the proportion of low-birthweight babies who are pre-term is smaller in developing than in developed countries because there are so many intrauterine-growthretarded babies. Because the figures for pre-term births are expressed proportionately does not necessarily mean that the incidence of pre-term births is lower in less developed countries.

Dr. Pollitt

I agree with what was said regarding the prevalence of lowbirthweight babies in developed countries and have nothing to add. I also agree with the speaker that mothers who bottle-feed their infants and use alcohol and marijuana might hinder the development of their offspring through different channels.

Dr. Victora

When we are looking at contaminants in breastmilk, we must realize that the alternatives to breastmilk can have contaminants, too, although they may be of different types. In New Zealand recently, the soya bean milk formula was found to have oestrogens, and the babies' intake was equivalent to about 12 pills per day, which is quite significant. I was wondering what is the effect of low-lead exposure on IQ, because we are trying to get lead out of petrol and similar things because of its effects on children.

Dr. Pollitt

There is no consensus on the cut-off point that should be used to define elevated lead levels in the blood. Recently, the cut-off point in the United States was lowered to 10 µg/dl (0.48 µmol/L). Nevertheless, the World Health Organization still uses the former cut-off point. Studies on the effects of elevated lead levels on cognition in children are limited by problems of research design. The use of experimental protocols is precluded because of obvious ethical concerns. Further, strong criticisms have been made of the validity and reliability of the data collected. In my view, there is still no evidence to support the contention that cerebral function is at risk with blood lead levels less than 25 parts per million.

Dr. Garza

What level of lead exposure would be needed to effect a change of one-half of a standard deviation in IQ?

Dr. Pollitt

I do not know.

Dr. Hartmann

How is it possible to differentiate between the effects of marijuana or alcohol on the infant through the maternal milk and the direct effects during gestation when the mother was also using alcohol or marijuana?

Dr. Pollitt

The published studies do not discriminate between the effects on the offspring of the mothers' consumption of drugs (alcohol or marijuana) before and during lactation. The authors, however, caution the reader against drawing unwarranted conclusions.


Social and demographic aspects

The cultural context of breastfeeding and breastfeeding policy

Penny Van Esterik



Breastfeeding is not instinctive behaviour but is dependent on learning and is, therefore, influenced by social and cultural factors. Thus, the social sciences as well as the biological sciences should be engaged in explanatory research about breastfeeding To rebuild breastfeeding cultures to protect, support, and promote breastfeeding a biocultural model of breastfeeding and child care that takes a broader view of culture must be developed: a view that attends more to differences than similarities, that provides more detailed contingencies of context, that is more sensitive to the forces that constrain women's lives, and that can be more directly linked to policy-making. This article explores the interdisciplinary nature of breastfeeding research and suggests some areas where anthropological theory and method could be put to better use to ask new research and policy questions about breastfeeding



Breastfeeding is the epitome of a biocultural phenomenon in which the processes of biology and culture are inextricably linked [1]. It is not instinctive behaviour but is dependent on learning; it is, therefore, influenced by social and cultural factors. Because breastfeeding is at the intersection of biological and cultural processes, the social sciences as well as the biological sciences should be engaged in explanatory research about breastfeeding. This article explores the interdisciplinary nature of current breastfeeding research and examines cultural factors and cultural context as two different ways to approach the "cultural" in "biocultural." I suggest some areas where anthropological theory and method could be put to better use in breastfeeding promotion and conclude that breastfeeding policy can be influenced by social science research.


Interdisciplinary nature of breastfeeding research

The participants in this Workshop come from many different disciplines but share a commitment to research and policy to strengthen breastfeeding. To work profitably together, we must acknowledge disciplinary differences. These differences emerge out of hundreds of years of Western philosophy and logic that have kept the sciences, social sciences, and humanities apart-or, in the case of breastfeeding, brought them together on biomedical terms. Anthropology straddles these divisions and permits an easier integration of the biological and the cultural.

Interdisciplinary work is not easy, as the different epistemological assumptions of each discipline need to be acknowledged. The term "biocultural" stresses the relation between biology and culture and, hence, is important for our discussions of breastfeeding and society. Most research on breastfeeding however, examines either the cultural or the biological aspects. The few people who put the two together often write and speak as if the biology part provides unquestioned facts, and the culture part provides the less important, ephemeral context.

Medical anthropology has been found to be "a powerful tool for reassessing what is taken as natural and normal in connection with the human body" [2]. Science has determined generalities or laws about breastfeeding many of which have been reviewed at this Workshop. But generations of mothers who breastfeed their children have also formed some generalities about the process. Our task should be to bring these two kinds of generalities or laws into a specifiable relation with each other, not to dismiss one, or to reduce one to the other. Millard [3] notes that the "guidelines" that Mexican women have about infant feeding are principles; that is, they are statements of natural tendencies or laws, conceptualizations of ultimate causes, and guides to practice. These guidelines also have commonalities with scientific practice. Application of cultural principles involves logic and situational decision-making on the part of the mother. These principles are used in conjunction with other sets of principles, such as the humoral system, to decide how individual infants should be fed.

Some concepts hinder and others enhance communication across disciplinary boundaries. In the past, "culture" and "cultural factors" have been used in ways that have not always increased understanding of infant-feeding practices or encouraged interdisciplinary collaboration between anthropology and the health sciences. As a term such as culture crosses disciplinary boundaries, its meaning and significance often change.

Defining culture and stating that cultural factors are important for understanding infant-feeding decisions does not alter the fact that the concept is very difficult to use. Just as biologists have difficulty agreeing on a definition of life, theologians on a definition of God, and doctors on a definition of health, anthropologists seldom agree on a definition of culture. We are unlikely to be able to find narrowly defined or easily measured cultural variables or indices to capture the variations within or between different infantfeeding patterns. Continuing to search for these elusive cultural factors may result in culture being blamed for generating attitudes and beliefs that are seen by health planners as obstacles that must be overcome by education.


Cultural factors in biomedical models

What sociocultural factors have been singled out for integration into biomedical and epidemiologic models of infant feeding? Common factors include religion, marital relationship, family composition, residence pattern, and values. Qualitative data have been used successfully to support findings from epidemiologic and biomedical research, but qualitative data can also be used to raise new questions and examine relations that are not amenable to reductionist models. Questions about values, attachment, nurturance, and sexuality-all intimately connected with breastfeeding and society-require interpretation from social science paradigms.

There is in the literature on breastfeeding clear agreement that maternal attitudes towards breastfeeding and bottle-feeding are culturally conditioned and, indeed, influence infant-feeding decisions. The most widely accepted method in health and nutrition research is the construction of knowledge, attitude, and belief questions that can be administered by a single context questionnaire, where respondents are asked whether, or to what extent, they agree or disagree with a statement about infant feeding. Scores on these questions may then be related to a mother's actual infantfeeding decisions.

There are numerous problems with this approach. Infant feeding in general and breastfeeding in particular are very personal and emotionally charged subjects. For this reason, it is difficult to obtain reliable, valid information on mothers' attitudes, beliefs, and knowledge through survey methods. Because the interviewers often have had no previous contact with the respondents, it is difficult for them to evaluate how the interview setting affects mothers' responses.

Interviewers' biases may encourage mothers to predict what the interviewer wants to hear and answer accordingly. This problem is complicated by respondents' desire to be polite to strangers or their hesitancy to speak frankly to authority figures perceived to possess power. In addition, standardized knowledge and attitude statements about infant feeding often reflect verbal cliches or key images developed through health-education or breastfeeding promotion campaigns. The problem of phrasing culturally appropriate questions or statements without biasing the response is particularly difficult.

Serious conceptual problems also arise in relating knowledge, attitude, and belief to actual practice. The links between knowledge and belief, belief and attitude, and attitude and practice are both poorly understood and ambiguous. Not only can we not demonstrate a causal relationship between attitude and behaviour, but we often cannot even predict the direction or order of the relationship. Behaviour may well change before attitude changes, not after. For example, breastfeeding-promotion campaigns, based on linear reasoning linking knowledge to belief, attitude, and practice, often aim to increase correct knowledge about the importance of breastfeeding. It is therefore not surprising that breastfeeding-promotion campaigns based on improving knowledge, attitudes, and beliefs about the benefits of breastfeeding have not, by themselves, proved to be an effective means for changing infant-feeding patterns. The success of these methods depends on whether women's knowledge of breastfeeding was acquired through formal classes, observation, books, or women's stories. Learning the practical management of breastfeeding is quite distinct from trying to teach someone else to do it right.

Underlying much of the literature on the determinants of infant-feeding practices is a nagging, unanswered question. Can we, given our current research methods, distinguish between mothers who want to breastfeed yet cannot, and mothers who do not want to breastfeed and do not? Women belonging to both of these categories may fail to initiate breastfeeding or may breastfeed for only a short period of time. We assume that women in the the former group have the knowledge and attitudes to support breastfeeding but are unable to do so. They are constrained by something, such as inappropriate medical practices, excessive promotion of breastmilk substitutes, or lack of social support. We assume that women in the latter group do not want to breastfeed or will breastfeed for only a short time.

The difference between these two categories of women lies not only in their demographic characteristics but also in their heads-in the ideas, beliefs, and assumptions about infant feeding that make up the cognitive and affective dimensions of human behaviour. Studies of the determinants of infant-feeding practices often include these factors under the variables of maternal knowledge and attitudes or cultural factors. Studies that do not emphasize these factors acknowledge their possible importance.

This approach to knowledge and attitudes, based on social-psychological methods and theories, does not reflect the way anthropologists approach human behaviour. Nor does it do justice to the complexity and flexibility of human decision-making about infant feeding. For example, in the complex calculus of infant-feeding decisions, a woman may weigh such factors as preference for dresses zippered up the back, avoiding sexual intercourse, or thwarting her mother-in-law. Even an apparently straightforward interpretation of behaviour may generate erroneous assumptions about cultural attitudes. For example, the conclusion that the use of Western-style blouses indicates a desire for modesty and a negative attitude towards exposure of the breast may be unjustified. In areas of the world where it is believed that the evil eye can poison breastmilk, blouses may, in fact, be worn to protect the breasts and breastmilk from danger.

Moreover, people have the capacity to tolerate an amazing degree of inconsistency between what they say they "believe," what they do believe, and what they do. The capacity of mothers to tolerate inconsistency, to hold contradictory beliefs and attitudes about infant feeding, and to interpret options reinforces the need for means of data collection and analysis beyond traditional knowledge and attitude assessment. Whereas cultural factors are determinants in chains of cause and effect, cultural context is a much less circumscribed concept that permits other kinds of information, such as the political context or the gender context, to guide explanations. This broad background information is referred to here as the "cultural context" of infant-feeding decisions. This context may be thought of as an additional level of analysis that integrates social, cultural, and biological factors. Such an additional level of analysis is necessary for interpreting attitudinal data. Survey responses alone seldom provide the data necessary for such contextual interpretation. For example, the knowledge that women agree that breastfeeding makes breasts sag, without the accompanying knowledge of how sagging breasts relate to women's body image and selfimage, may easily result in misinterpretation of this information or distortion of its meaning. Body image, beliefs about breast size, and ideas of beauty are embedded in culturespecific systems of gender ideology.


Breastfeeding context

Breastfeeding as a process is culturally constructed: that is, in spite of its physiological base, the process itself, its meaning, and the way it is integrated into cultural systems varies globally. Until recently, we have had minimal information on the sociocultural context of breastfeeding from detailed, long-term ethnographic analysis. This evidence is now available, and it demonstrates that women's lives and child-care practices are changing rapidly.

One conceptual model emerging from these ethnographic data approaches cultural context as the interaction between style and structure. Style refers to the manner of expression characteristic of an individual, a period of time, and a place. The concept of infant-feeding style communicates fundamental cultural assumptions underlying infant-feeding decisions. It refers to the manner of feeding infants in particular communities and includes both the way to feed an infant and the values, attitudes, and beliefs associated with that behaviour. Infant-feeding style includes the style of interaction between mothers and infants, eating style (how does the infant-feeding pattern fit with the household meal pattern?), breastfeeding style (how is breastfeeding accomplished?), and feeding "in style," reflecting the fact that infant-infant feeding choices are part of dynamic, changing trends and fashions.

To understand differences in breastfeeding style, it may be useful to distinguish between breastfeeding as a process and breastmilk as a product. Process or product interpretations may be emphasized in different contexts [4].

Both personal and shared styles of infant feeding interact with organizational and institutional structures, such as health-care institutions and marketing systems. These structures are important in influencing mothers' infantfeeding choices. The interaction between style and structure should allow us to predict how infant-feeding choices might be affected by different policy options.


Rebuilding breastfeeding cultures

A meeting of WHO and UNICEF policy makers in 1990 resulted in the adoption by 30 governments of a global initiative, the Innocenti Declaration, on the Protection, Promotion, and Support of Breastfeeding [5]. This Declaration stated that for optimal breastfeeding,

all women should be enabled to practice exclusive breastfeeding, and all infants should be fed exclusively on breastmilk from birth to four to six months of age. Thereafter, children should continue to be breastfed, while receiving appropriate and adequate complementary foods, for up to two years of age or beyond.

In order to bring this about,

efforts should be made to increase women's confidence in their ability to breastfeed. Such empowerment involves the removal of constraints and influences that manipulate perceptions and behaviour towards breastfeeding, often by subtle and indirect means. This requires sensitivity, continued vigilance, and a responsive and comprehensive communications strategy involving all media and addressed to all levels of society. Furthermore, obstacles to breastfeeding within the health system, the workplace, and the community must be eliminated.

This carefully worded statement is a challenge to change many priorities of the modern world to rebuild breastfeeding cultures. The language stresses the empowerment of women to breastfeed, rather than their duty to breastfeed, a change that should bring more advocates for women's health to support breastfeeding policies.

To rebuild breastfeeding cultures that will protect, support, and promote breastfeeding, we must work from models of breastfeeding and child care that are more sensitive to gender and the forces that constrain women's lives. What is entailed in recreating breastfeeding cultures? In some parts of the world, changing assumptions about body image is a priority; ensuring women are adequately fed is of higher priority elsewhere. A great deal of political will to make changes benefiting women and children is crucial in all state and international institutions.

At some point, policy makers need numbers to evaluate and finance policies and programmes. But that does not mean that ethnographic observations and other cultural information are merely illustrative anecdotes. The linkage between cultural data and policy can be directly relevant to establishing priorities. Because ethnographic description is both holistic and richly contextualized, it is easy for policy makers to picture the real-life conditions of the families that their policies will affect. Some products of ethnographic fieldwork, such as community sketches and mothers' life histories, are available much faster than survey data, which require substantial processing.

Solutions to infant-feeding problems must come from the cultural context underlying infant-feeding decisions. The same "culture" that some policy makers view as an obstacle to development must ultimately provide solutions assembled from available options, ideas, and strategies already in the cultural repertoire. In this case, the task of the policy maker is to choose options and implementation strategies that are most compatible with the infant feeding style in different countries.

Knowledge of style without consideration of structural constraints or supports would, however, be unproductive. Structural constraints, such as the powerful influence of the health-care system, the marketing practices of transnational manufacturers of infant foods, and structural supports, such as vendors selling porridges or traditional midwives, must also be included in policy decisions to suggest new directions to improve infant-feeding practices.

Breastfeeding-promotion activities, mothersupport projects, and consumer-advocacy campaigns are often viewed in isolation from each other and from other programmes. Nevertheless, breastfeeding has been linked to related childsurvival campaigns, such as immunization, family planning, growth monitoring, and oral rehydration therapy. To encourage the changes envisioned by the Innocenti Declaration, breastfeeding advocates must seek new allies and closely examine the concerns such potential allies have with regard to past promotion of breastfeeding. A number of potential allies could be called on to offer support for breastfeeding initiatives.

Links have already been made between breastfeeding and child-spacing programmes. The World Bank's "A case for promoting breastfeeding in projects to limit fertility" [6] identifies three arguments that have been used against the promotion of breastfeeding for its contraceptive effect:

The World Bank report challenges these arguments and demonstrates the advantages of breastfeeding for child spacing. It is, however, important to note that African women's groups in particular have warned against linking breastfeeding to family planning too closely and have argued for breastfeeding to be promoted for its own sake.

Fewer commonalities have been recognized with environmental groups. Lactation and breastfeeding had no place in the report of the World Commission on the Environment and Development entitled Our Common Future (unpublished report, Geneva, 1987), although the report's treatment of issues such as sustainable development, the international economy, population, human resources, food security, energy, and industry all could be related to infant-feeding choices. The environmental movement and advocates for sustainable development are potential allies for breastfeeding could be shown how breastfeeding furthers their interests and encouraged to link breastfeeding into their planning. But breastfeeding promotion that treats women as producers of natural resources is bound to fail.

The strongest supporters of breastfeeding should be women's groups. Breastfeeding is seldom considered an important issue to be discussed in national or regional women's organizations in developed or developing countries. For example, national women's organizations were not immediately supportive of Brazil's breastfeeding initiatives because they saw it as a means of isolating women in their homes. Similarly, women's groups in India protested against legislation to control the marketing of infant formula, calling the bill draconian, fearing it would damage women's careers and force them back to the kitchen.

According to a review of American maternity policies by Kamerman [7]:

Although the women's movement clearly has played an important role in improving the status of women, maternity and other family benefits have not been at the forefront of its agenda.... Heretofore, the women's movement has viewed family politics as either too traditional a focus for unenlightened women, or as too open to criticism as conservative.

In Scandinavian countries, efforts to support breastfeeding have been more strongly supported by the women's movement:

The new feminist movement in Norway has encouraged women's self-confidence and pride- both prerequisites for breastfeeding. Unlike similar movements, the Norwegian feminist movement has never seen a contradiction between women's liberation and breastfeeding [8].

The antagonism to breastfeeding that exists in many parts of the European and North American women's movements results in difficulties in establishing breastfeeding as a priority in international women's meetings. This antagonism or apathy will be particularly destructive to efforts to implement the Innocenti Declaration unless it is countered by clear recognition of breastfeeding as a woman's right. Thus, there is a need to reconsider breastfeeding in relation to human rights. Programmes that promote the child's right to breastmilk rather than the mother's right to breastfeed without consideration of maternity entitlements are likely to be damaging in the long run.

The issue of breastfeeding has been invisible in the women's movement, either because it is taken for granted in some parts of the world or because it is totally out of women's consciousness in other parts of the world. It is the responsibility of breastfeeding advocates to integrate their issues with other concerns of women. This process was begun at the Beijing Conference on Women in September 1995, where a number of measures protecting breastfeeding were included in the final action plan. Women's organizations need to be informed about how breastfeeding entitlements can complement their other policy objectives and then be mobilized to support breastfeeding policies. If feminist theorists and activists supported breastfeeding, they might well be able to reformulate a more global, holistic approach to gender, care, mothering, and empowerment.



Indigenous cultural knowledge about child feeding is critically important for successful breastfeeding research and programmes, but there are ways this knowledge can be better used. I suggest that anthropologists can be most useful to breastfeeding research if they resist the pressure to reify culture, or to provide checklists of cultural factors affecting infant feeding. No list would be adequate to the task, and the temptation to reduce culture to one or two factors within a biomedical framework may draw attention away from more important questions about power, class, or gender. Whenever we use the concept of culture, we should ensure that it is firmly situated in understandings of material conditions of life and fields of power. Anthropologists should develop innovative ways to communicate about other patterns of infant feeding without resorting to homogenizing generalizations and dangerous reifications that can be used to create or perpetuate stereotypes: "All Bangladeshi women reject colostrum," "All Mexican women wean abruptly," and so forth. We should strive to make visible the cultural systems in which we operate, examining assumptions about our own health-care systems and the tenets of biomedicine and not treating them as natural or as givers. Finally, we should resist the commoditization of culture by insisting on the necessary theoretical framing of the concept and locating it more directly within various social science disciplines. In this way, we can make room for metaphor as well as measurement in breastfeeding research, programmes, and policy.



1. Stuart-Macadam P. Dettwyler K, eds. Breastfeeding: biocultural perspectives. New York: Aldine De Grnyter, 1995.

2. Lindenbaum S. Lock M, eds. Knowledge, power and practice: the anthropology of medicine. Berkeley, Calif, USA: University of California Press, 1993.

3. Millard A, Graham MA. Weaning decisions by rural Mexican women. Working Paper on Women in International Development no. 73. East Lansing, Mich, USA: Michigan State University, 1984:1.

4. Van Esterik P. Commentary: an anthropological perspective on infant feeding in Oceania. In: Marshall LB, ed. Infant care and feeding in the South Pacific. New York: Gordon and Breach, 1985:357.

5. WHO/UNICEF meeting on Breastfeeding in the 1990s; a global initiative. Innocenti Declaration on the Protection, Promotion and Support of Breastfeeding, Florence, Italy, 1 August 1990. New York: UNICEF, 1990.

6. Berg A, Brems S. A case for promoting Breastfeeding in projects to limit fertility. Washington, DC: World Bank, 1989.

7. Kamerman S. Maternity policies and working women. New York: Columbia University Press, 1983.

8. Baumslag N. Breastfeeding: the passport to life. New York: UNICEF, 1989.

Contents - Previous - Next