Too close to home: Dioxin contamination of breast milk andthe political agenda

KATHRYN HARRISONDepartment of Political Science, The University of British Columbia, C472-1866 Main Mall,Vancouver,BC, CanadaV6T 1Z1

Abstract. Both the Canadian and U.S. governments have determined that breast-fed infants areamong the populations most exposed to dioxins, receiving levels of exposure orders of magnitudeabove those considered acceptable. In light of the political controversy associated with dioxins andthe cultural signi¢cance of breast milk as a symbol of purity, one might have expected dioxincontamination of breast milk to achieve prominence on both the popular and governmentalagendas. Yet as this article demonstrates, this issue has received less media and governmentalattention than other environmental issues believed to present comparable or lower health risks.Consistent with recent literature on agenda denial strategies, there is some evidence that e¡orts byenvironmental groups to publicize levels of breast milk contamination have been rebu¡ed bygovernment o¤cials, physicians, and breastfeeding advocacy groups fearful that women will stopbreastfeeding. However, what is more striking is just how seldom environmentalists have attemptedto reframe this issue. The article argues that North American environmentalists have consciouslychosen not to press the dramatic issue of breast milk contamination out of concern that motherswould discontinue breastfeeding, as well as personal anxiety about an issue that fundamentallychallenges conceptions of our own bodies and our relationships with our children. Their self-restraint challenges the depiction by some authors of environmental groups as eager to capitalizeon any opportunity to provoke public concern and outrage to advance their agenda. The case studyalso suggests that the literature on agenda setting must look beyond active strategies of agendadenial by economically and politically powerful interests, to the role of shared cultural values inshaping ^ and restricting ^ the political agenda.

Introduction1

In 1990, the Canadian federal government released a report declaring dioxinsand the structurally-related furans to be toxic to human health and the environ-ment as de¢ned by the Canadian Environmental Protection Act (Governmentof Canada, 1990). Although not emphasized in the accompanying press release,a table in the report estimated that breast-fed infants receive 16.5 time the`tolerable daily intake' of dioxins set by the Canadian government. Four yearslater, the U.S. Environmental Protection Agency (EPA) released its draft `dioxinreassessment,' which o¡ered a comparable estimate of the daily dose of dioxinreceived by breast-fed infants. However, since EPA's virtually safe dose ofdioxin is much lower than Canada's, that estimate was 6000 times greater thanthe level of exposure considered acceptable by the U.S. agency (U.S. EPA, 1994a).

One might have expected these dramatic ¢ndings to receive widespreadmedia and public attention for a number of reasons. Dioxin is arguably the

35Policy Sciences 34: 35^62, 2001.ß 2001Kluwer Academic Publishers. Printed in the Netherlands.

ultimate `political chemical' (Salter, 1988), thanks to a combination of its long-standing reputation as the `most toxic chemical known to humankind' and thesustained attention to dioxin from environmental groups on both sides of theborder. Moreover, breast milk is laden with cultural signi¢cance as a symbolof purity and sustenance for the most vulnerable members of society. Theconnection between dioxin and breast milk thus might be expected to yield acombination of fear and outrage, and, at a minimum, interest.

Yet the potentially explosive government ¢ndings received little attentionfrom government, environmental groups, the media, or the public in eithercountry. Although industrial sources of dioxin and various sources of humanexposure, particularly contaminated ¢sh, have continued to receive public andgovernmental attention, the level of contamination of breast milk with dioxinshas gone relatively unnoticed in both Canada and the United States. Thepurpose of this paper is to explore the reasons for the absence of this issuefrom the political agenda in North America.

The paper begins with an overview of the literatures on risk perception andagenda setting, followed by consideration of the health risks of dioxins and thelack of prominence of dioxin in breast milk compared to other environmentalissues on the public agenda. Thereafter, several explanations are considered forthe absence of breast milk contamination from the political agenda. Althoughthere is some documentary evidence and support from interviews for agendadenial by actors who sought to keep the issue o¡ the political agenda,2 thepaper ¢nds the strongest support for an alternative explanation, that environ-mentalists simply chose not to press the issue of breast milk contamination inthe ¢rst place for a variety of reasons, including concern that breastfeedingmothers would discontinue breastfeeding and personal anxiety about an issuethat fundamentally challenges our conceptions of our own bodies and ofour relationships with our children. The case study of dioxins in breast milksuggests that the literature on agenda setting must look beyond active strategiesof agenda denial by economically and politically powerful interests, to therole of shared cultural values in shaping ^ and restricting ^ the politicalagenda.

Risk perception and political agenda-setting

Two di¡erent scholarly literatures can be brought to bear on this question,those concerning risk perception and political agenda setting. It has been longestablished in the psychological literature on risk perception that the lay public'srisk assessments are often very di¡erent from experts' estimates of statisticalrisk. Most people tend to underestimate familiar risks, such as the likelihood ofdying in a car accident, and overestimate risks that are unfamiliar, associatedwith delayed impacts, involuntary, or associated with `dread' (Slovic et al.,1979). Displaying all those characteristics, dioxin contamination of breastmilk would seem ripe for overreaction. This is reinforced by studies on gender

36

and risk perception, which ¢nd that, confronted with similar hazards, womentend to perceive greater risks than men (especially white men) (Flynn et al.,1994). Gustafson (1998) notes that not only do women tend to worry moreabout the same risks than men, but they also tend to worry about di¡erentrisks, paying greater attention to accidental and health risks. Davidson andFreudenberg (1996) ¢nd the strongest support for two explanations for thedi¡erences between men's and women's risk perceptions: That women aresocialized to nurture and maintain life and are thus more concerned abouthealth and safety; and that women have less trust in institutions associatedwith science and technology. Again, since dioxin in breast milk is both ascienti¢cally sophisticated issue, and a health concern intimately linked towomen's role as nurturers, one might expect a particularly strong reactionfrom women.

However, as Douglas and Wildavsky (1982) have noted, an important draw-back of the risk perception literature is its exclusive focus on the individual.While the lay public, and women in particular, may well have been primed toreact strongly to news of dioxin in breast milk, as discussed below they had fewopportunities to obtain that information from the popular media. The riskperception literature alone thus cannot tell us why this issue received so littleattention from environmental groups and the press. Douglas and Wildavskybegin from the perspective that risk is a collective, rather than individual,construct, and o¡er a sociological analysis for why some environmental risksbecome the object of widespread concern and others do not. They argue thatthe environmental movement, as a result of its `sectarian' organizational struc-ture and ideology, is inclined to selectively publicize risks that are involuntary,irreversible, and hidden, and for which blame can be attributed and conspiraciesalleged (ironically, quite similar to the conclusions drawn by the psychologicalliterature on individual risk perception of which the authors are so critical).`Purity becomes a dominant motif' according to Douglas and Wildavsky (1982:p. 124). Given the hidden and involuntary nature of dioxin contamination ofbreast milk, the irreversibility of cancer, which is the health impact most oftenassociated with dioxin, and the signi¢cance of breast milk as a symbol ofpurity, a cultural analysis of risk would seem to predict considerable e¡ort byenvironmentalists to provoke public alarm about dioxin in breast milk.

The political science literature on agenda setting also seeks to explain theprocesses by which we collectively select some issues for governmental attentionand ignore others. The foundational work on agenda setting by Cobb and Elder(1972) emphasized the characteristics of societal problems, including theirmagnitude and potential for disasters or `focusing events.' Birkland (1998)emphasized the importance of the nature and organization of policy commun-ities, arguing that well organized policy communities are better positioned totake advantage of focusing events to achieve policy change. Kingdon's (1984)`garbage can model' of agenda-setting also emphasized the role of policy entre-preneurs in bringing developments in the `policy stream' to the public's andthus the government's agenda. Applying Kingdon's model to agenda-setting

37

for environmental risks, Harrison and Hoberg (1991) noted that the discoveryof dioxins in pulp mill e¥uents only received public attention in the U.S. whenGreenpeace brought to light EPA's ¢nding (Van Strum and Merrell, 1987), andin Canada when Greenpeace released its own analyses of sediments from thereceiving environment of aVancouver Island pulp mill. In addition to focusingevents and interest groups, other authors have emphasized the importance ofinstitutions in constraining the political agenda, and of changes in institutionalvenue as a strategy for agenda change (Baumgartner and Jones, 1993).

Much of the recent literature on agenda setting has adopted a constructivistapproach, focusing on how problems are de¢ned or `framed' by political actors.Thus, Stone (1988; 1997) focuses on strategic representation of problems bypolicy entrepreneurs, and in particular how they use causal stories to reframeaccepted societal `conditions' as political `problems.' Similarly, Schneider andIngram (1993) consider how stereotypes about target populations can increaseor decrease the likelihood of an issue a¡ecting those communities from reachingthe political agenda. Baumgartner and Jones (1993) o¡er a punctuated equili-brium model, in which a stable policy network and problem de¢nition canoccasionally be disrupted through popularization of an alternative problemde¢nition by actors seeking to focus public attention on the issue. Re£ectingon this literature, Rochefort and Cobb (1994) draw a link between the earlierpluralist literature on agenda setting and more recent constructivist literature.Ideas, language, and perceptions are the currency with which interest groupseager to place new issues on the agenda, and their opponents seeking toconstrain that agenda, engage in political struggle. Again, one might expect anissue such as dioxins in breast milk to be ripe for strategic representation andframing, and that an audience inclined to fear unfamiliar risks would bereceptive to the issue.

Of particular interest for this paper is an emerging thread in the constructi-vist literature that seeks to explain why some issues do not reach the agenda.Cobb and Ross (1997) analyse various strategies adopted by actors that opposereframing of an issue and its emergence on the political agenda, and when thosestrategies tend to succeed. They conclude that `agenda denial occurs not justbecause of the complicated mechanics of the policy process or the lack ofgovernmental resources, but often because a proposed action challenges existingworld-views and identities in unacceptable ways that opponents demonstratee¡ectively' (Cobb and Ross, 1997: p. 219). The implication for the breast milkcase is that one should look to the counter-activities, including strategic use ofsymbols and rhetoric, by those who stand to lose should widespread publicconcern about dioxins in breast milk be mobilized.

The foregoing suggests several alternative explanations for the absence ofdioxin contamination of breast milk from the political agenda. First, althoughthe risk of contaminated breast milk may well be characterized by unfamiliarity,delayed impacts, and dread, there simply may be other issues with similarcharacteristics that present greater health risks and thus have been moresuccessful in capturing the attention of interest groups, governments, and the

38

public. Second, the emergence of this issue on the political agenda may havebeen precluded by the absence of a focusing event. Third, the institutionalvenue may not have been conducive to successful reframing of the issue ineither country. Fourth, consistent with Cobb and Ross' analysis of agendadenial strategies, environmentalists' e¡orts to publicize the issue may have beenrebu¡ed by political opponents. Finally, environmentalists simply may havedeclined to press the issue in the ¢rst place, a form of self-censorship examinedin greater detail below. Before turning to these hypotheses, a brief review ofwhat is known about the risks of dioxins in breast milk is in order.

Dioxin and human health

The chemical usually referred to as `dioxin' is 2,3,7,8-dibenso-p-dioxin (2378TCDD) is the most toxic of a family of structurally similar chlorinated dioxinsand furans. Most analyses measure toxicity of all dioxins and furans collectivelyin terms of toxic equivalents or `TEQs' of 2378 TCDD. Dioxins and furans arenever intentionally manufactured, but are released as unintended byproductsof processes such as incineration of chlorinated compounds, manufacturing ofcertain chlorinated chemicals, and chlorination of naturally occurring com-pounds in wood during pulp and paper manufacturing. Dioxins have lowsolubility in water but high solubility in fat, and they are extremely stable,persisting in the environment in some cases for decades. As a result of thesetwo characteristics, dioxins tend to bioaccumulate as one moves up the foodchain. At the top of the food chain, humans receive most of their exposurethrough consumption of animal products ^ meat, ¢sh, chicken, eggs, and dairyproducts ^ and from human breast milk. Since dioxin is metabolized veryslowly by the human body, the accumulated `body burden' of dioxin stored infat tends to increase over time.3

The extent of risk to human health from dioxins is highly controversial. Theever-present di¤culties of demonstrating statistically signi¢cant e¡ects amongdiverse human populations is exacerbated in the case of dioxins by the fact theyare unintended contaminants. As a result, levels of contamination often werenot measured historically and exposure is typically confounded by concurrentexposure to other toxic substances. However, several large epidemiologicalstudies in the 1990s did provide evidence of carcinogenicity in humans, prompt-ing the International Agency for Research on Cancer in 1997 to declare 2378TCDD to be a known human carcinogen (IARC, 1997). Nonetheless, given thelimited number of epidemiological studies as well as researchers' capacity toexperimentally vary exposure among experimental animals, most risk assess-ments for dioxins continue to be based on studies of laboratory animals. TCDDhas been found to cause a variety of adverse health e¡ects in test species atextremely low doses.

Cancer risk has dominated regulatory decisions concerning dioxins to date,and thus will be the focus of the discussion here. However, it is noteworthy that

39

regulators are increasingly attentive to other end points, since evidence suggeststhat humans may be more sensitive to the immunological, reproductive, anddevelopmental e¡ects of exposure to dioxin. EPA's dioxin reassessment con-cludes that the margin of safety between doses that causes adverse e¡ects inlaboratory animals and current levels of exposure is less than a factor of 10 forthe average population for some of these non-cancer e¡ects, and that popula-tions receiving high-end exposures are approaching the levels at which adversee¡ects have been reported in laboratory animals (EPA, 1994a: pp. 9^82).

As previously analysed by the author (Harrison, 1991), the Canadian and U.S.governments have adopted very di¡erent approaches to cancer risk assessmentfor dioxin. The U.S. approach yields a more cautious `virtually safe dose' of0.01 picograms per kilogram body weight per day (pg/kg/day) (U.S. EPA, 1994a),1000 times less than Health Canada's `tolerable daily intake' of 10 pg/kg/day.4

The divergence between the Canadian and U.S. estimates of acceptable expo-sure might be expected to have signi¢cant policy implications, since estimatesof current exposure levels at 1^4 pg/kg/day fall squarely between the two.5

However, the fact that highly exposed populations, including those who con-sume large quantities of ¢sh, can be exposed to doses several times higher thanthe average has prompted similar regulatory action, for instance to controldischarges from pulp mills and restrict consumption of ¢sh, in both countries.

Dioxin was ¢rst detected in human milk samples from Germany in 1984,followed soon thereafter by con¢rmation of comparable levels of contamina-tion in other industrialized countries. During lactation, a portion of a mother'sbody fat, and with it dioxins accumulated therein since birth, is transferred toher breast milk, `e¡ectively putting babies one step higher on the food chain'(Great Lakes Science Advisory Board, 1991: p. 38). Although dioxin levelsbetween the mother and fetus tend to be in equilibrium before birth, after birththere is a net transfer of dioxins from mother to infant during breastfeeding.Indeed, one of the most e¡ective, if unintended, ways for a grown woman toreduce her body burden of dioxins is through lactation (Sullivan et al., 1991;Shecter et al., 1998).

The combination of the fact that breast milk is relatively contaminated withdioxin compared to other foods, that many infants initially subsist entirely onbreast milk, and that they consume a large quantity of food relative to theirbody weight, means that breastfeeding infants can receive a higher dose ofdioxin per unit body weight than any other population. Exposure estimates rangefrom EPA's at 60 pg TEQ/kg/day to Health Canada's at 165 pg TEQ/kg/day,depending on the averaging time.6 Even EPA's lower estimate is 35 times higherthan the typical background exposure for adults. EPA estimates that a breast-fedenfant receives 4 to 12% of its lifetime exposure to dioxin during the ¢rst yearof life.

We know relatively little about the impacts of these levels of dioxin exposureon breastfeeding infants. However, several ongoing epidemiological studieso¡er the potential to advance our understanding of the developmental impactsof dioxins and other persistent toxic substances. The earliest of these studies

40

found a developmental de¢cit among the children of women who consumed¢sh contaminated with PCBs and other toxic substances from the Great Lakes(Jacobson et al., 1984; Jacobson and Jacobson, 1996). A subsequent study ofbreast-fed children in North Carolina found lower psychomotor skills duringthe ¢rst two years among children whose mothers had higher levels of variouscontaminants in their breast milk (Rogan and Gladen, 1991), but those e¡ectswere no longer observable by age 5 (Rogan and Gladen, 1993). A Dutch studycomparing breast-fed and formula-fed infants similarly found negative e¡ectsof PCBs and dioxins in breast milk only at 7 months of age, though the negativee¡ects were outweighed by the bene¢ts of breastfeeding. The researchers did,however, ¢nd a negative correlation between prenatal exposure to toxins and avariety of developmental e¡ects in infants (Huisman et al., 1995; Koopman-Esseboom et al., 1996; Patindin et al., 1998, 1999a).7 This led them to concludethat, even though exposure from breast milk may be less problematic thanprenatal exposure, to the extent that it represents a signi¢cant contribution tothe body burden of the next generation of women of reproductive age, it couldstill have an important delayed impact (Patindin et al., 1999b). A recent ¢ndingthat breast-fed infants whose mothers' milk contains higher levels of dioxins aremore likely to have tooth defects is consistent with hypothesized developmentalimpacts (Ralo¡, 1999; Alaluusua et al., 1999). Using risk estimates from animalstudies, Hoover estimated that dioxins contribute three quarters of the lifetimecancer risk associated with some two dozen toxic substances in breast milk(Hoover, 1999).

It is noteworthy that despite pronounced di¡erences between U.S. andCanadian regulators' cancer risk assessments, the exposure received by breast-fed infants greatly exceeds both the EPA and Health Canada estimates ofacceptable daily intake of dioxin (at 16.5 times the Canadian tolerable dailyintake and 6000 times the U.S. virtually safe dose). The question remains,however, how to interpret a relatively high exposure for a short period of time,when animal tests and the risk estimates derived from them are typically basedon lifetime averaged exposures. Consistent with its more risk tolerant approach,the Canadian government's assessment (1990: p. 44) notes that although theexposure received by breast-fed infants is `relatively high,' `breast feeding onlyoccurs for a short part of the life span .. . and lower exposures throughout theremainder of the life span reduce lifetime exposure below [the 10 pg/kg/day]guideline.'8 In contrast, EPA expressed concern that reliance on lifetime riskestimates may actually underestimate the risk to breast-feeding infants, sincethey are exposed during a time of rapid cell division and growth (EPA, 1994a:pp. 9^19). EPA thus identi¢es breast-fed infants among the `highly exposedpopulations' of particular concern.9

41

Breast milk contamination and the popular agenda

It is, of course, impossible to prove something's non-existence, thus presentinga fundamental obstacle to studying why some issues fail to reach the politicalagenda. However, it is not asserted here that dioxin in breast milk was on noone 's public policy agenda. The issue was undoubtedly a concern for somemembers of the public and some actors within government. Rather, the focushere is the inattention to this issue relative to other environmental health risksthat have received more widespread media, public, and governmental attention.

Cobb and Elder (1972) distinguish between the popular agenda and thegovernmental agenda. The primary focus of this paper is the former, de¢nedby Cobb and Elder as the set of `issues commonly perceived by members of thepolitical community as meriting public attention and as involving matters withinthe legitimate jurisdiction of government authority.' In the absence of detailedpublic opinion polls comparing attitudes toward di¡erent environmental healthrisks, newspaper coverage will be used as a surrogate for public attention.Although extensive media coverage of an issue does not ensure that it will beperceived by the public as warranting a public policy response, given the techni-cally complex nature of the issues that will be compared it is highly unlikelythat an issue could reach the public agenda in the absence of extensive mediacoverage.

How does media coverage of dioxins in breast milk compare to that ofcomparable environmental health risks? Table 1 compares the number of storieson various environmental and health issues in the New York Times and theCanadian Newspaper Index between 1985 and May 1999. The absolute ¢guresare not comparable between Canada and the U.S., since the Canadian indexincludes multiple newspapers.10 Rather, it is the relative attention to di¡erentissues within each index that is of interest here. The issues presented wereselected because, like dioxins in breast milk, each involves either risks fromfood contamination or risks to children, or both.

In both countries, there has been greater media attention to dioxins in ¢shthan in breast milk. The former has also received greater attention from govern-ment regulators, who have responded to detection of dioxins in ¢sh and shell¢shwith ¢sheries closures and advisories in both countries. The issue of dioxinleaching from paper milk cartons into cow's milk received comparable atten-tion to dioxin in breast milk in the New York Times, though less in Canadiannewspapers. However, it is noteworthy that, unlike breast milk contamination,contaminated cow's milk elicited a policy response in both countries.11

Coverage of dioxins in foods is minor in comparison to two other potentialfood risks ^ `mad cow' disease and bovine growth hormone ^ though the latterreceived proportionately greater attention in Canada. Children's exposure topesticides in foods, Alar in apples, and phthalates in infant toys also allreceived greater coverage than dioxins in breast milk, especially in the U.S.where each of these issues have been the focus of prominent campaigns byenvironmentalists.

42

The Canadian newspaper coverage of dioxins in breast milk reported inTable 1largely re£ects two wire service stories, each of which was covered in multiplenewspapers. The ¢rst was a report in April 1988 that dioxins had been detectedin samples of breast milk from Quebec (the ¢rst sample to be analysed in anational study), which was picked up by Quebec newspapers and the Globe andMail.12 Then, in October 1990, references to high levels of infant exposure todioxins from breast milk in the federal government's dioxin assessment werepublicized in a widely reprinted story by the Canadian Press wire service.13

A leaked draft of the U.S. EPA dioxin reassessment also received front pagecoverage in the New York Times in May 1994, though that story gave lessprominence to breast milk as a source of exposure.14 In other stories, the NewYork Times focused on dioxins in cow's milk, bleached co¡ee ¢lters, andcontaminated ¢sh, and did not even mention breast milk.15

Interestingly, with the exception of a couple of follow-up stories in whichexperts and politicians sought to allay public concern, each of these storiesquickly faded from public view.16 In both Canada and the U.S., bureaucrats andpoliticians stressed that the bene¢ts of breastfeeding outweigh any risks. How-ever, Canadian bureaucrats tended to provide stronger reassurance of absolutelysafety, as when one Health and Welfare Canada o¤cials stated, `It has norelevance to human safety when you get down to these extremely low levels.'17

There has been somewhat greater media coverage of PCB contamination ofbreast milk. However, the focus has been on elevated levels of PCBs in the breastmilk of Inuit women, which has been the subject of an ongoing study by Quebec

Table 1. Canadian and U.S. newspaper coverage of selected environmental health risks.

Search NewYorkTimes Canadian News Index

Dioxin and breast milkb 18 13Dioxin and cow's milkc 18 4Dioxin and ¢shd 139 45Alare 161 29Mad cow diseasef 41000 225Phthalates and toysg 275 25Pesticides in children's foodh 425 41Bovine growth hormonei 110 161

a Searches were completed 13^15 May 1999. The NewYorkTimes was searched using Lexus-Nexus.The Canadian News Index was searched using Canadian Business and Current A¡airsb Search strategy: dioxin$ and milk and (breast$ or mother$).c dioxin$ and milk and (cow$ or carton$).d dioxin$ and (¢sh$ or shell¢sh or ¢sher$).e Alar.f (mad and cow$) or (Creudtzfeldt).g (phthalates$ or toxin$ or toxic$ or vinyl) and (infant$ or baby$ or children$) and (teething ortoy$).h child$ and pesticide$ and (food$ or diet).i (bovine growth hormone$) or bgh or bst.

43

researchers (Milly and Leiss, 1997), not the general population. As well, thequestion of dioxin in breast milk appears to have received greater public attentionrecently in Japan, suggesting an opportunity for future comparative research.18

In summary, there is no question that the chemical `dioxin' has receivedample attention from the media, the public, and governments in both Canadaand the United States. However, in light of the relatively high exposures asso-ciated with dioxin contamination of breast milk, and the cultural sensitivity toinfant exposure and breast milk as a symbol of purity, it is striking how littleattention there has been to this particular issue. The absence of signi¢cant presscoverage, and the relatively low pro¢le of environmentalists in that coverage,suggests that environmentalists have not attempted to publicize this particularissue, whether as a concern in and of itself or as a way to mobilize publicconcern for persistent toxic substances more generally.

Explaining the absence of breast milk contamination from the public agenda

Severity of the problem

The remainder of this article considers alternative hypotheses for the absenceof breast milk contamination from the public agenda. First might the lowerlevel of media and governmental attention to the issue of dioxin in breast milkrelative to other environmental health risks simply be a re£ection of the factthat those other risks are more serious? This section compares risk assessmentof dioxin in breast milk with four other environmental issues that have receivedgreater media and/or governmental attention. In each case, the estimated risksfrom dioxins in breast milk are comparable or greater, suggesting that magni-tude of risk alone does not explain the absence of breast milk contaminationfrom the political agenda.

As discussed above, dioxins in ¢sh have received both greater media atten-tion and governmental attention, the latter re£ected in ¢sheries closures andadvisories in both countries. The U.S. EPA estimates that the average personconsumes 0.1 pg TEQ/kg/day of dioxin from ¢sh, though subsistence ¢shers andethnic populations that eat more ¢sh can consume as much as 2 pg/kg/day.19 Incontrast, EPA estimates that a breast-fed infant receives a daily dose of 60 pgTEQ/kg/day. It is, however, true that one consumes ¢sh for a higher proportionof one's life than breast milk. Taking that into account, estimates of the lifetimecancer risks are comparable: 6.6% of the background cancer risk from dioxin isfrom ¢sh compared to 4^12% from breast milk.20 The greater public andgovernmental attention to dioxin in ¢sh and shell¢sh thus cannot be explainedby the severity of risks, since the risks are of similar magnitude.

The revelation that dioxin had been seeping from bleached cardboard cartonsinto cow's milk (it was not present at the same levels in milk stored in plastic orglass containers), received comparable media attention to dioxins in breastmilk in the U.S., though less in Canada. However, it did achieve greater

44

prominance on the governmental agenda ^ U.S. regulators and their counter-parts in at least one Canadian province (Quebec) responded promptly, eventhough the dairy and paper industries had committed to eliminating the prob-lem voluntarily by switching to dioxin-free cartons. Moreover, even limitedmedia coverage of dioxin contamination of cow's milk provoked a publicreaction, when sales of milk in cartons declined in Canada in October 1988and in the U.S. in September 1989.21 Yet, levels of 2,3,7,8 TCDD in cow's milkwere detected at 0.02 to 0.07 ppt, compared to a typical concentration in breastmilk of 17 ppt, over 100 times higher.22 Again, the fact that one consumes cow'smilk for a longer proportion of one's life should be taken into account. How-ever, even then FDA estimated a lifetime risk from consumption of milk fromchlorine-bleached cartons of between 5 and 11 in a million, compared with acontribution to lifetime cancer risk from breast milk estimated by EPA to bebetween 4 and 100 in a million, again suggesting comparable or greater risksfrom the latter.23

More recently, the discovery in June 1999 that animal feed given to cows andchickens in Belgium was contaminated with dioxins received internationalpress coverage, and prompted consumer alarm and bans of Belgian beef,chicken, and dairy products by governments around the world. The Belgianfood scare is widely considered to have been a signi¢cant factor in the defeat ofthe governing party in national elections in Belgium that same month. Yet, theWorld Health Organization estimates the daily intake of dioxins associatedwith a typical diet of contaminated Belgian food to be 8 to 10 pg/kg (WHO,1999), roughly 10 times less than the same organization estimates is consumedby breast-fed infants around the world every day.

Finally, the case of Alar, a growth regulator that was used to promote a ¢rmtexture and unblemished appearance in apples, is similar in many respects tobreast milk contamination. Like dioxin in breat milk, Alar, or more accuratelythe chemical UDMH which was formed when Alar-treated products wereprocessed, was considered a particular risk to children, since they consumemuch larger quantities of apples, especially in the form of apple juice, per unitof body weight. Alar was the subject of an extensive media campaign by U.S.environmental groups in 1989, re£ected in greater media coverage in Table 1,and was ultimately banned in both Canada and the U.S. The U.S. EPA estimatedan upper limit lifetime risk of 4.5 ¾ 10º5 for Alar (Harrison and Hoberg, 1994:p. 69), roughly at the midpoint of EPA's range for dioxin in breast milk asreported above.

The point of these four examples is not to provoke alarm about dioxin inbreast milk. EPA's risk assessment methodology is intentionally conservative,yielding what the agency itself describes as upper-bound risk estimates. Actualrisks may well be much lower in each of the above cases. However, the fact thatrisks estimated using the same methodology24 to be comparable or even lowerthan those associated with dioxin contamination of breast milk have receivedgreater public and/or governmental attention indicates that the magnitude ofrisk alone cannot explain the lack of attention to breast milk contamination.

45

The breast milk case may be unusual in one respect, however. There is strongevidence that breast feeding o¡ers physiological bene¢ts to both mother andinfant, the latter including protection against infection, immunological bene-¢ts, and protection against allergies (Canadian Paediatric Society and HealthCanada, 1998). The relevant comparison thus may be net risk (or bene¢t),rather than environmental risk alone. Indeed, despite the presence of contam-inants, physicians, government spokespersons, breast-feeding advocates, andeven environmentalists invariably stress that the bene¢ts of breastfeeding out-weigh the risks.25 Although in the examples discussed above, apples, milk, and¢sh also have nutritional value, equally nutritious substitutes for those productsare, at least in theory, available, which is not the case for breast milk.26 Asdiscussed below, the perceived bene¢ts of breastfeeding relative to the risks wereclearly a consideration for all players in the relevant policy network. However,for this to explain the lack of attention to dioxin in breast milk, one would haveto assume that the only policy option available, or the inevitable consequenceof even raising the issue, would be to discourage breastfeeding.Yet is conceivablethat one could express concern about the extent of contamination and emphasizethe need for reduction of dioxin sources, even while encouraging women tocontinue breastfeeding. Indeed, one might expect those most aware of thebene¢ts of breastfeeding to be among those most outraged, and most inclinedto seek action to prevent further contamination of breast milk.

Absence of focusing events

From Cobb and Elder (1972) to Birkland (1998), students of agenda settinghave emphasized the role of focusing events in bringing new issues to thepolitical agenda. `Sudden, dramatic and often harmful, focusing events givepro-change groups signi¢cant advantages' in overcoming opposition fromthose that seek to keep an issue o¡ the political agenda (Birkland, 1998: p. 56).It is beyond question that if there had been an epidemic of babies becomingnoticeably sick from drinking contaminated breast milk, this issue would haveassumed dramatic proportions. However, while the absense of a focusing eventwould appear to be a contributing factor to the low pro¢le of this issue, itcannot o¡er a complete explanation since focusing events are neither necessarynor su¤cient for agenda change.

Not only do many issues reach the political agenda without a focusing event,but many focusing events are orchestrated by policy entrepreneurs (Kingdon,1984; Harrison and Hoberg, 1991). Environmental groups have held pressconferences to release their own studies (as Greenpeace did in releasing analysesof dioxins in sediments near pulp mills), to publicize expert reports that theythemselves have commissioned (as U.S. enviromentalists did in the case of alar),and to bring little-noticed government or academic studies to the media's andthus the public's attention (as Greenpeace did in publicizing Health Canada'sdetection of dioxins in cow's milk).27 `Discoveries' thus can become focusing

46

events with a little help from policy entrepreneurs. Yet, as discussed below,environmentalists have made little e¡ort to generate such attention to dioxinsin breast milk. It is also noteworthy that even dioxin-related focusing eventsapparently have not been su¤cient to focus attention on this issue. Althoughthe discovery of widespread dioxin contamination of the Belgian food supplywas a high pro¢le issue not only in Belgium but throughout Europe in 1999,there appears to have been little attention to the question of breast milkcontamination as a result.

Institutional factors

The low visibility of breast milk contamination relatively to other environ-mental health issues is thus not readily explained by risk assessments alone orthe absence of focusing events. This section considers the in£uence of politicalinstitutions on agenda setting. Elsewhere, I have o¡ered an institutional explan-ation for the divergent Canadian and U.S. risk assessments for dioxin notedabove (Harrison, 1991). The more open regulatory process in the U.S. hasresulted in greater politicization of risk assessments, and that has encouragedU.S. regulators to rely on conservative, formal models, which, in the case ofdioxin, have yielded more cautious risk assessments than informal safety fac-tors (see also Jasano¡, 1986). However, the fact that neither the public norregulators in either country have shown much inclination to focus on dioxinsin breast milk suggests that institutional di¡erences are not an importantexplanatory factor in this case.

On the other hand, institutional similarities may have shaped the terms ofdebate in ways that constrained access of this issue to the formal agenda in bothcountries. Both Canada and the U.S. have statutes authorizing regulation ofcontaminants in foods sold in commerce. This institutional venue may haveframed the issue as one of safety of the commercial food supply in bureaucrats'minds. The fact that breast milk is not sold commercially may have meant thatit was not top of mind for government o¤cials, even though, if breast milk wereregulated like infant formula, levels of dioxin contamination would in theoryprohibit its sale in both countries.28 While this institutional barrier thus may havebeen a contributing factor, the signi¢cance of this explanation is undermined bythe fact that both Canadian and U.S. regulators have issued non-bindingadvisories concerning dioxins to recreational and subsistence ¢shers, who donot sell their catch. Moreover, with the exception of ¢sh, violation of dioxinstandards for commercial foods is routinely ignored in Canada in any case.29

Agenda denial

Recent literature on political agenda setting focuses on the critical role ofproblem de¢nition in the political struggle between actors seeking to place an

47

issue on the agenda and those who would deny that issue agenda status.Following on the recent work of Cobb and Ross (1997), one can hypothesizethat `initiators,' in this case environmentalists, may have tried to rede¢ne anenvironmental problem (dioxins or persistent toxic substances more generally)by focusing on the powerful cultural symbol of breast milk, but that they failedto do so because `opponents,' those who stand to lose should the issue achieveagenda status, succeeded in reinforcing an alternative framing of the issue.

As discussed above, a handful for stories concerning dioxin in breast milkhave emerged in Canadian and U.S. newspapers over the last decade. In somecases, environmentalists played a role in alerting reporters to those stories,although in others, such as the New York Times' coverage of EPA's dioxinreassessment, the media appear to have been simply reporting on the govern-ment's own studies.30 Canadian newspaper coverage of the potential risksto breast-feeding infants discussed in the 1990 Canadian dioxin report onlyemerged several months after the report's release, after Greenpeace hand-picked a wire service reporter considered sympathetic and provided him withextensive background research.31 However, the Greenpeace campaigner recalled,`We hadn't planned to focus on breast milk. It was supposed to be about dioxinand food. It was [the reporter] who chose to emphasize breast milk.' In the U.S.,the Environmental Defense Fund held a press conference in 1987 to releasethe results of a university scientist's testing of breast milk for dioxins.32 It isnoteworthy that in that press conference, indeed in almost all cases whereenvironmentalists referred even in passing to contamination of breast milk,they invariably stressed that the bene¢ts of breastfeeding are believed to out-weigh any risks.33 In fact, on one occasion in 1997 when Greenpeace U.S.A.held a press conference at which an association between PVC and dioxin inbreast milk was expected to arise (though it was not the primary focus of thepress conference), they invited representatives of the La Leche league to presenttheir own press release simultaneously.34

The `opponents' in this case include politicians and government o¤cials,breastfeeding advocates, and the media themselves, all of whom were e¡ectivein using the cultural symbolism of breast milk to their advantage in containingthe issue. Although many of the studies suggesting cause for concern have beenprepared by government o¤cials, politicians and bureaucrats have not onlychosen not to highlight the issue of dioxin in breast milk, they have gone toconsiderable lengths to downplay it, using the strategies of placation, denial,and attack identi¢ed by Cobb and Ross (1997).When press coverage did occa-sionally arise, government o¤cials responded with a standard set of arguments.First, they expressed shared concern, but noted that various actions werealready being taken to reduce dioxin discharges at the source. Second, theydismissed the quantities as `minute,' and only detectable as a result of recentadvances in analytical techniques. Third, they dismissed the risks to infants asinsigni¢cant or, in the case of Canadian bureaucrats' statements, nonexistent.35

Their reassurances were picked up in headlines such as `Dioxin traces foundin breast-milk, samples from Quebec no cause for alarm: Scientists' and `No

48

problem' for infants in short term: Dioxin traces found in human milk inQuebec.' Fourth, like environmentalists, politicians and government o¤cialsstressed that the bene¢ts of breastfeeding outweigh any risks, and encouragedwomen to continue breastfeeding.

Finally, o¤cials and politicians occasionally resorted to accusations thatthose who raised the issue were fear mongering. The most extreme exampleof this strategy was when the Canadian Minister of Health, Perrin Beatty,responded to questions about his government's dioxin report from theParliamentary opposition, who also have been provided with Greenpleace'sbackground resarch on dioxins in foods,36 by accusing them of `attemptingto frighten nursing mothers into believing that they were poisoning their in-fants.'37 Interestingly, the Minister's comments indicate that he had been briefedto anticipate that the opposition would focus on dioxins in breast milk. Heappears to have followed through with his brie¢ng notes, even though theopposition had not even mentioned breast milk in their questions about thedioxin issue. Other examples of self-restraint by those would might seek topublicize dioxins in breast milk to further their policy objectives are discussedin the next section.

In responding to questions about contamination of the food supply and thegovernment's proposals to relax dioxin standards for foods, Beatty convenientlyhid behind the breast milk issue, raising the proposition that `If we were to takethe position that no dioxins whatsoever could be allowed in any food used inCanada, we would have to ban breast milk, for examle, and throw away literallyhundreds of millions of dollars worth of good, nourishing, nutritious food inCanada each year.'38 The government of Canada thus used rhetoric to placeitself squarely on the side of babies and mothers, arguing that anyone whowould even question the purity of breast milk was out to harm children.

The second source of opposition was from the medical profession and thebreastfeeding advocacy community. This includes physicians who specialize inbreastfeeding, who are frequently quoted in newspaper stories about breastmilk contamination. As one reporter re£ected about Dr. Jack Newman, fromthe Hospital for Sick Children in Toronto, `If you write anything suggesting thatthere might be anything in breast milk that would cause anyone to consider notnursing, he ¢res o¡ a letter or calls you up.' In light of the well-documentedbene¢ts of breastfeeding, environmentalists' would-be allies within the medicalprofession tended to see the issue in the ¢rst instance not in terms of contami-nated vs. pure breast milk, but rather as one of discouraging vs. promotingbreastfeeding (breast milk vs. arti¢cial formula). They thus placed the burdenof proof on those who would suggest risks from contamination.39

Also included in this community are volunteer groups such as the La LecheLeague and the Infant Feeding Action Coalition (INFACT). One can imagineorganizations that passionately support breastfeeding might have quite di¡erentreactions to news of dioxin contamination of breast milk: Outrage at thedesecration of breast milk, or denial should any hint of risk discourage womenfrom breastfeeding. The overwhelming reaction, at least for public consump-

49

tion, has been the latter.40 Echoing goverment o¤cials, breastfeeding advocatesemphasized the bene¢ts of breastfeeding and the `minute quantities' of con-taminants detected. Building on previous political struggles with infant formulamanufacturers, some breastfeeding advocates also raised two other themes.First they emphasized that cow's milk and formula are polluted too. Andsecond, some authors have suggested that allegations of contamination ofbreast milk are the work of formula companies seeking to discourage breast-feeding.41

Interestingly, a third, though less consistent, source of opposition has beenthe press themselves. There have been almost as many stories criticizing `alarm-ist' coverage of pollutants in breast milk as there have been stories to criticize.42

One U.S. environmentalist interviewed said that although her group doesinclude occasional references to dioxins in breast milk in press releases, thepress won't cover it: `Reporters say that their editors say it's too in£ammatory,no one will believe you.'43 However, on the other hand, another U.S. environ-mentalist reported resisting encouragement by reporters to sensationalizestories by playing on the breast milk angle, and, as noted above, it was the1990 Canadian Press reporter who chose to focus on dioxins in breast milkfrom a broader Greenpeace report on dioxins in foods.

One example of the powerful backlash that references to contamination ofbreast milk can elicit emerged last year, when Theo Colborn, a scientist withthe World Wildlife Fund who was central to the emergence of the issue ofendocrine-disrupters on the public agenda with her best-selling book, OurStolen Future (Colborn et al., 1996), responded to a question about dioxins inbreast milk in an interview in Mother Jones magazine. Although Colborn'sresponse included the standard statement that the bene¢ts of breastfeedingappear to outweigh the risks,44 the fact that she suggested any uncertaintyabout that elicited what the magazine's editorial sta¡ described as an `enor-mous volume of mail ^ all negative,' a sample of which was reprinted in thefollowing issue.45 The letters to the editor included not only the argumentsnoted above, but also at least one personal attack on Colborn's scienti¢ccredentials.

As noted above, although a handful of stories about dioxin contamination ofbreast milk were published in both countries, in no case did the issue `take o¡.'With the exception of one or more stories the following day reassuring breast-feeding mothers, there would be no further coverage in the same newspaper formonths or even years.Whatever e¡orts environmental activists may have madeto focus public attention on the issue of dioxin in breast milk as a symbol ofmore widespread risks from persistent toxic substances were rebu¡ed by moredominant messages: That `breastfeeding is good' and that anyone who mighthint otherwise is at best irresponsible and at worst malicious. In contrast, whenGreenpeace alerted the public to a government study con¢rming the presenceof dioxins in milk in paper cartons, their message was met with a moremeasured response from both government and the a¡ected industry. Althoughgovernment o¤cials provided comparable reassurances that the risks were

50

minimal and that they were taking every action to eliminate them, there was nocomparable defense of the wholesomeness of milk nor depiction of the bearersof the news as somehow anti-children.

The literature on agenda denial thus o¡ers considerable insight into this casestudy. However, it falls short in two respect. First, it is noteworthy that, unlikemost of the case studies analysed in Cobb and Ross (1997),there is no industrywhose self-interest is immediately threatened by allegations of dioxins in breastmilk.46 None of the opponents had an interest in maintaining contamination ofbreast milk ^ quite the contrary. In each of the cases of dioxin in ¢sh and milkand Alar in apples, there was an industry selling those products that wasmaterially threatened by public concern about impurities, whether as a resultof intentional use of a product like Alar or unintentional contamination.(Despite the incentives for those industries to defend the reputations of theirproducts, it is noteworthy that, in contrast to the breast milk case, these issuesdid achieve agenda status, which in turn resulted in policy change.) In the breastmilk case, the `opponents' were not defending their economic self-interest.47

The question of why a pro-breast feeding lobby would emerge in this case wouldappear to turn more on culture than self-interested agenda denial. Second,although agenda denial can help to explain why the breast milk contaminationstory never `took o¡,' what is more striking is just how seldom environmentalistseven tried to make an issue of dioxins in breast milk, an issue considered in thenext section.

Self-restraint by the environmental community

The fact that the media have seldom focused on the question of dioxins inbreast milk is, in large part, because environmentalists themselves have maderelatively few e¡orts to date to draw attention to the issue, whether as a strategyto press for action on dioxins or toxic substances more generally. The oneinstance, discussed above, where Greenpeace Canada prompted, albeit inadver-tently, a story about dioxin in breast milk that was picked up nation-wide in1990 would appear to be the exception that proves the rule. An internalGreenpeace memorandum describes the media focus on breast milk contami-nation as a `negative feature' of the dioxin in food initiative that Greenpeacewas `not prepared for . . . and did not adequately deal with.' The campaignerresponsible for the initiative nonetheless was encouraged by the responseand developed a proposal for an ongoing campaign focused on the risks ofdioxin and other toxic substances to children as `the most defenceless andvulnerable members of our society.'48 He later explained, `I know people getmore concerned about damage to their children. They're willing to put up witha certain amount if it's their own bodies, but they get more concerned if it's athreat to their kids. So I thought this was a great environmental issue, becauseof that angle.' However, for reasons discussed below, his colleagues withinGreenpeace chose not to pursue that campaign, displaying uncharacteristic

51

restraint for an organization that has made its name with publicity-seekingprotests.49

It is noteworthy that environmental groups' self-restraint in this case is quiteinconsistent with Douglas and Wildavsky's depiction of sectarian environmentalorganizations using any hint of conspiracy or catastrophe to further theiragendas. It is noteworthy that the environmental community has not alwaysbeen so restrained, even with respect to health risks to children. Indeed, in itscurrent `toxic toys' campaign, Greenpeace is using threats to infants and childrenas a focal point for a larger campaign against PVC plastics. Similarly, as notedabove, Greenpeace emphasized risks to children in pressing Health Canada toaddress elevated levels of dioxins in cows milk in cardboard cartons. GreenpeaceU.S.A. also petitioned the U.S. Department of Agriculture to prepare an Envi-ronmental Impact Statement for its school lunch program, to re£ect the risksto children from dioxin contaminated cow's milk (Thompson and Graham,1997).50 Greenpeace's emphasis on cows milk is particularly noteworthy since,at the time of these activities, Greenpeace activists already had a copy of adoctoral thesis showing that levels of dioxin were much higher in breast milkthan cow's milk, and they were well aware of the contrast.51

Alar o¡ers another compelling counterexample. In 1989, U.S. environmentalgroups launched a very public campaign against Alar. The campaign wasorchestrated by a public relations ¢rm, and involved the release of a scienti¢cstudy commissioned by environmentalists estimating high cancer risks tochildren, and extensive reliance on a media-friendly group, Mothers andOthers for Pesticides Limits, fronted by Meryl Streep, that was created for thepurpose of the Alar campaign. The strategy was e¡ective in large part becauseit combined the credibility of science with `two potent symbols: The apple ^long mythologized as the quintessence of pure, healthy food ^ and the threat toinfants and young children ^ the most innocent and vulnerable members of thepopulation' (Harrison and Hoberg, 1994: p. 66). EPA subsequently bannedAlar, but not before the public reacted with alarm to environmentalists'campaign. Sales of apples and apple products plummeted and dozens of schoolboards stopped distributing apple products in cafeterias.

In contrast to these examples, interviews with Canadian and U.S. environ-mentalists con¢rm that they made conscious decisions not to sensationalize thebreast milk issue, though for di¡erent and, in some cases, multiple reasons. Therationales they o¡ered are grouped here into three categories: Anticipation ofdefeat and/or backlash, concern about unintended consequences, and personaldiscomfort with the whole issue.

With respect to the ¢rst of these, it is signi¢cant that evidence concerningdioxin in breast milk emerged in the late 1980s. In the last three decades,breastfeeding has experienced something of a renaissance in North America,with rates of breastfeeding doubling from 25% in 1970 to roughly 50% in 1987in the U.S.52 Popular culture has shifted from hostility to breastfeeding, where awoman would not even consider breastfeeding in a shopping mall, to the pointwhere numerous U.S. States have passed laws a¤rming women's right to

52

breastfeed in public places (Yalom, 1997: p. 142). At the same time, there hasbeen an international campaign against formula manufacturers, focused espe-cially on their e¡orts to promote bottle feeding in developing countries. As aresult of these factors, some environmentalists expressed concern about thepotential backlash should they question the healthfulness of breast milk. Anenvironmentalist active in dioxin campaigns in the late 1980s and early 1990sexplained, her voice full of frustration, `Everyone was ` shh, shh'' because` breast milk is better for you.'' You have to remember, there was the wholeWHO [World Health Organization] campaign [to promote breastfeeding] andthe Nestles boycott at the time.' The presence of a well-organized and favourablyconstructed breastfeeding advocacy community was a factor for another envi-ronmentalist, who explained, `It's extremely dangerous territory for us to wadeinto as political activists. There's been a huge investment from the La Lecheleague, nurses, and midwives, telling us the very best thing to do is to breastfeedyour baby.' Yet another observed that anytime she did mention breast milk, `theLa Leche league immediately started coming on to me and saying ``you can'tsay that!''' In other words, the anticipated backlash from the `opponents' wassu¤ciently strong that the `initiators' conceded the battle before it had evenbegun.

However, in other cases, environmentalists expressed con¢dence that theycould have `gotten mileage' from the breast milk issue, but that they chose notto for two di¡erent kinds of reasons. First, some environmentalists exercisedself-restraint out of concern that news of contaminants in breast milk mightcause undue alarm and prompt some women to discontinue breastfeeding. Theunanticipated strength of the public's reaction to the Alar campaign ^ e¡ectivelywithdrawing a nutritious staple from many children's diets ^ may have been afactor. One U.S. environmentalist re£ected that Alar was not our ¢nest hour,'and explained that `We consciously chose not to try to make [dioxin in breastmilk] an issue . . . I've tried very deliberately to frame this not as ``good grief,there's dioxin in breast milk'' . . . Breast milk is not the enemy, dioxin is.' Sheacknowledged that it was a di¤cult decision to forgo an opportunity to mobilizepublic concern using such a powerful symbol as breast milk, but nonethelessfelt comfortable that her organization had made the right decision. Otherenvironmentalists explained that, `There was incredible fear within Green-peace. People always said you can't not breastfeed,' and `When we look at thedata, it's really frightening.We are all very concerned that what we do doesn'tturn into a massive marketing campaign for formula.'

The desire not to provoke alarm was apparently a factor in GreenpeaceCanada's decision not to follow through on the 1991 dioxin in breast milk storywith a campaign focused on the risks of dioxin to women and children. Themale sta¡ member who had developed the campaign proposal explained,`I really hit a raw nerve. There was a big reaction [among Greenpeace sta¡,especially the women saying] ``We don't want people to panic.'' I hit the roof.That's paternalistic and condescending. If people's kids are at risk, they shouldget worried. I didn't expect the women in Greenpeace to be that blase.'

53

As his comment suggests, it is possible that women and men in the environ-mental movement reacted quite di¡erently to this issue. It is noteworthy thatthe leading environmentalists on toxics issues in both Canada and the U.S. tendto be women, many of them mothers.53 Their comments indicate that theirdecisions not to focus on breast milk were more than just strategic anticipationof backlash or self-restraint lest there be unintended consequences, but ratherhighly personal and emotional reactions to an issue that fundamentally chal-lenges our conceptions of our own bodies and our relationships to our children.Female environmentalists described how they `agonized' with women colleaguesover the issue of breast milk contamination. Interview subjects often invokedtheir own experience as mothers. As one noted, `Pregnant women have enoughother things to worry about. We've all been there.' Women environmentalistsstressed how `unthinkable' it is for themselves and other women to realize theyare passing on toxic substances from their own bodies to their babies. A U.S.environmentalist acknowledged that the apparent risks from dioxins in breastmilk are `much higher' than those associated with children's exposure to pesti-cides in foods, an issue on which her organization has actively campaigned, butexplained that, `People just block it out. . . . As a mother, I breastfed two infants.I remember thinking ``there's nasty stu¡ in there.'' But in the end, I just didn'twant to know.' Female environmentalists' self-restraint thus re£ected very per-sonal consideration of the impact such information can have on women likethemselves. A U.S. activist explained, `I say to myself, what would I tell apregnant woman? What advice would I give? If I don't know the answer, Ishouldn't raise the issue.'As a former Greenpeace activist summed it up, `It wasjust too close to home.'

These reactions suggest that women environmentalists in particular hadinternalized many of the same cultural values re£ected in a backlash againstallegations of dioxin in breast milk. Indeed, for them, the challenges posed totheir self-image and their relationships to their children by the issue of breastmilk contamination were both serious and profoundly personal. In consideringdioxins in breast milk, women are forced to think of their own bodies not onlyas contaminated, but as toxic to those they most love and seek to protect.54

In the end, the symbol of de¢ling a sacred relationship between mother andchild and of contaminating our own bodies may in some cases have been toopowerful for environmentalists themselves.

Looking to the future

Clearly, dioxins in breast milk has not been a prominent issue on either thepublic's or environmentalists' agendas. In the forgoing discussion, I have iden-ti¢ed several forces that have suppressed emergence of the issue. One need notimply, however, that such forces are immutable. In that regard, it is noteworthythat greater attention to this issue by environmentalists began to emerge in thepast year in the context of international negotiations of a treaty on Persistent

54

Organic Pollutants (POPs). The women's caucus of the International POPsElimination Network in particular has called for greater emphasis on breastmilk by environmentalists. A member of that caucus explained that `It becameclear to us that the breast milk issue and reproductive issues will help win theelimination debate. It's a straightforward and very moral argument, and God ison our side, whoever she may be.' Such a strategy is re£ected in recent reportsissued by both Greenpeace International (Allsopp, Stringer and Johnston, 1998)and WWF-UK (Lyons, 1999).55 It is nonetheless striking that these reportshave emerged only after levels of dioxin declined substantially in the last decade,though they remain far above international guidelines.

Environmentalists are proceeding with extreme caution. The question ofwhether to further pursue the breast milk angle is being contested amongWorld Wildlife Fund a¤liates internationally. As an environmentalist a¤liatedwith another organization considering whether to pursue a breast milk campaignexplained, `We have to be sure ^ really sure ^ that when we talk about thebreast milk issue we scope it out in a way that allows us to move aheadpolitically, rather than blow it. If we were to say levels of dioxin in breast milkare so alarming we have to question breast feeding in some parts of the world,we would get annihilated.'

How far environmentalists will proceed with this strategy of reframing theissue, and whether it will succeed in the face of anticipated opposition, remainto be seen. It is noteworthy that both the recent Greenpeace and World WildlifeFund reports received minimal media coverage in North America, failing toeven to elicit backlash stories on subsequent days. However, if, in the end,breast milk contamination does reach the popular agenda, the relative lack ofattention to the issue by environmentalists, the public and politicians for over adecade remains an important ¢nding that can inform our understanding ofagenda setting and cultural selection of risks.

Conclusion

The absence of a potentially explosive issue such as dioxin contamination ofbreast milk from the political agenda suggests a need to counterbalance recentconstructivist studies of political agenda setting with greater attention to thedi¤culty in reframing some issues. While studies of issue construction haveyielded important insights into how issues reach the political agenda throughreframing, one should not assume from a set of cases chosen for study becausethe issues reached (or returned to) the political agenda after being reframedthat successfully reframing an issue is a simple matter.

The emphasis of this paper on the absence of an issue from the politicalagenda is reminiscent of an older literature on `nondecisionmaking.' However,that literature was concerned with questions of competing class or economicinterests and political power. Indeed, Bachrach and Baratz (1970: p. 49) arguethat nondecisionmaking can occur only if there is a con£ict of interests or

55

values. Similarly, in more recent work, Cobb and Ross (1997) focus on the wayself-interested actors use rhetorical appeals to popular culture and symbols toprevail over their opponents. Yet it is striking that the breast milk case is notabout self-interested politics. None of the participants was in favour of breastmilk contamination nor opposed to breastfeeding. Rather, it reveals howshared conceptions of our selves and of our relationships with each other canshape ^ and suppress ^ political debates.

Thus, although there were instances in which political opponents reinforcedan alternative issue construction (`breastfeeding is good') to deny agenda access,or environmental groups chose not to press the issue in anticipation of such astrategy, in other cases environmentalists chose not to pursue the potentiallysensational issue of dioxins in breast milk simply because they shared the sameconcerns about provoking undue alarm and the same cultural values concern-ing breastfeeding as their would-be opponents. As Bosso (1994: p. 199) hasobserved, `Con£ict over problem de¢nition is not guaranteed. There are a greatmany times when prevailing values or the rules of the game simply screen outmost (and sometimes all) alternative de¢nitions of a problem.'

It is also noteworthy that environmentalists' self-restraint, re£ecting a mix ofsocial responsibility and personal discomfort, is inconsistent with both Douglasand Wildavsky's cultural theory of risk selection and a simple translation to thesocietal level of psychological studies of individuals' risk perceptions. Environ-mentalists may well have reacted with alarm to the hidden, irreversible, anddreaded characteristics of dioxin contamination of breast milk, but for themost part they elected to focus on other issues lest they provoke public alarm.This suggests that the environmental movement is not as marginal as Douglasand Wildavsky suggested, but rather shares many cultural norms that shape,and limit, the range of issues environmental groups choose to pursue. Moreover,the tendency revealed by some environmentalists to deny or disregard such adisturbing risk might not be discernable from surveys eliciting individuals'initial reactions to a list of risks. The psychological literature on denial andavoidance of information about risks such as HIV/AIDS and tobacco, as wellas the anthropological literature on taboos and sacred totems, may yield insightsfor further research on non-agenda setting.

The broader cultural context is clearly important in understanding whichproblem rede¢nitions are likely to succeed or even, as in breast milk case, to beattempted. The discovery of dioxins in breast milk happened to coincide with arediscovery of breastfeeding in North America. Yet it was not intuitively ob-vious whether the increased popularity of breastfeeding would propel the issueof contaminants in breast milk onto the political agenda or, as turned out to bethe case, repress it. The fact that the latter construction prevailed suggests thatsome cultural values may be so sacred that any hint of de¢lement, even by thosewho raise the threat in order to oppose it, may be just too unthinkable.

56

Notes

1. Research for this paper was funded by the Social Sciences and Humanities Research Councilof Canada. The author gratefully acknowledges the research assistance of Will Amos, andhelpful comments received from Lynda Erickson and participants of the ECPR workshop onthe Politics of Food.

2. With respect to documentary evidence, I reviewed all newspaper stories in which dioxins andbreast milk are mentioned in the NewYork times and the Canadian Newspaper Index between1985 and May 1999, Canadian and U.S. government reports on dioxins, scienti¢c literature onbreast milk contamination, and publications, web pages, and newsletters of interest groupsactive on toxics and/or breast-feeding issues in either country. In addition, I conducted open-ended con¢dential interviews with the following: 11 representatives of national environmentalgroups active on toxics issues and/or environmental health risks to children (6 Canadian,5 U.S.); 1 physician specializing in breast feeding; 2 reporters; and 3 representatives of nationalbreast-feeding advocacy organizations (1 Canadian and 2 U.S.).

3. There are two exceptions to this. First, the body burden can decrease during childhood as therelatively high dose received during breastfeeding is diluted by growth (Ayotte et al., 1996).Second, an adult woman's body burden declines during breast feeding as she transfers lifetime-accumulated contaminants to her infant through her breast milk.

4. The Canadian approach applies a `margin of safety' of 1000 to the lowest dose at which noexcess tumours have been observed in test animals. This presumes that there is a thresholddose below which exposure presents no risk to human health. In contrast, the U.S. EPAassumes that there is a linear relationship between dose and cancer risk, and thus relies onmathematical models to extrapolate from high dose animal tests to the lower doses relevant toregulatory goals of health protection and current human exposure. Although this approachassumes that there is always some residual risk at any non-zero dose, EPA typically de¢nes the`virtual safe dose' corresponding to a risk of one in a million. It is noteworthy that the moreopen and contested risk assessment process in the U.S. prompted EPA to prepare a draftdioxin reassessment several thousand pages in length, compared to the 55 page dioxin assess-ment issued by the Canadian government.

5. Health Canada estimates average exposure to be 2 to 4 pg TEQ/kg/day (Government ofCanada, 1990), while the U.S. EPA estimates background exposure to be 1 to 3 pg TEQ/kg/day,3 to 6 if dioxin-like PCBs are included (U.S. EPA, 1994b: pp. 9^15).

6. The daily dose per unit body weight decreases over time as an infant grows and begins toconsume other foods, and because dioxin in breast milk declines over time as dioxin reservesare £ushed from the mother's body. This also explains why cow's milk tends to be much lesscontaminated than human milk, since dairy cows are continuously lactating.

7. The authors did not have the capacity to measure prenatal exposures to dioxins, but observedthat, because levels of PCBs and dioxins tended to be correlated in the subjects' breast milk, itis di¤cult to ascertain the extent to which the observed e¡ects were caused by PCBs or dioxinsor both.

8. To the extent that the Canadian risk assessment is premised on the existence of a threshold,compared to the U.S. reliance on a model that is insensitive to the distribution of exposure overtime, one would expect greater concern about a high dose concentrated over a short period inCanada.

9. However, at the same time the agency curiously omits breast milk in summarizing lifetimesources of exposure, even though by the agency's own estimates it contributes 4^12% ofexposure. This is particularly noteworthy since EPA's summary of exposure sources includessoil ingestion by pre-schoolers as a route of exposure, but not breastfeeding, even though theformer constitutes 1/10th the lifetime contribution of dioxin from breast milk.

10. Although it would have been preferred to compare indexes with comparable groupings ofnewspapers, or comparable individual newspapers, such as the (Toronto) Globe and Mail andNew York times, the New York Times (via Lexus-Nexus) and the Canadian Newspaper Indexwere the only indexes to which I had access.

57

11. `Quebec ¢rst to ban furans and dioxins in milk cartons: Page,' Montreal Gazette, 17 March1989, A2. The U.S. FDA promised action in May 1990 (NewYorkTimes, A17, 1May 1990).

12. Canadian Press, `Dioxin traces found in breast-milk samples from Quebec no cause for alarm:scientists,' Montreal Gazette, 3 April 1988; and `` No problem'' for infants in short term:Dioxin traces found in human milk in Quebec,' Globe and Mail, 4 April 1988, A3.

13. Dennis Bueckert, `Dioxins poisoning mothers' milk, Study ¢nds deadly chemicals exceedlimits in nearly all foods,' Winnipeg Free Press, A1, 29 October 1990; `Unacceptable levels oftoxic chemicals found in breast milk,' A7, 29 October 1990; `Mothers' milk high in dioxins,study says,' Toronto Star, 29 October 1990, A1; and Mary Lynn Young and CP, `Breast milkhas high dioxin levels,' Vancouver Sun, A1, 29 October 1990.

14. Keith Schneider, `Fetal harm is cited as primary hazard in dioxin exposure,' New York Times,A1, 11May 1994.

15. See, for instance, Philip Shabeco¡, `Government says dioxin from paper mills poses no majordanger,' NewYork Times, 1May 1990, A17.

16. In addition to the stories cited above, see Dennis Bueckert, `Nurse babies despite dioxin,Beatty advises,' Winnipeg Free Press, 30 October 1990, 15; and `Breast milk is safe, Beatty tellsmothers,' Montreal Gazette, 30 October 1990, B1.

17. Young, `Breast milk has high dioxin levels.'18. `Dioxin scare unsettles many Japanese,' Vancouver Sun, 3 June 1998, A14.19. Average exposure is summarized in EPA, 1994b, Table II-6. EPA (1994b, 9-20) indicates that

high end exposure can be 20 times higher.20. This is based on the proportion of EPA's estimated lifetime exposure from each source, and

assumes a linear dose-response function and that age of exposure is irrelevant (i.e., that infantsare no more sensitive to exposure than adults). The percent of exposure from contaminated¢sh, based on Table II-6 (EPA, 1994b) is calculated without accounting for breast milkexposure, and would thus be somewhat less corrected. This comparison neglects both the factthat some individuals consume more ¢sh than the average person and that some infants arebreastfed longer or receive milk with a higher fat content.

21. See, for example, `Dioxins discovered in milk cartons,' Vancouver Sun, 26 October 1988, A1-2;and Philip J. Hilts, `Cartons found leaching dioxin to milk,' NewYorkTimes, 2 September 1989, 8.

22. Hilts, `Cartons found leaching dioxin to milk.' Data on levels of breast milk contamination inindustrialized countries can be found inWorld Health Organization (1988).

23. FDA's estimate is reported in Hilts, `Cartons found leaching dioxin.' The estimate of lifetimecancer risk from breast milk is based on 4 to 12% (i.e., the proportion of total lifetime exposurefrom breast milk) of EPA's lifetime cancer risks from dioxin of 1E-3 to 1E-4. It is noteworthythat FDA has historically adopted a potency estimate for TCDD that is a factor of 9 less thanEPA's; in other respects, the two agencies risk assessment methodologies for dioxin arecomparable. This would have the e¡ect of rendering the upper bounds of these risk estimatescomparable if they were conducted by the same agency.

24. Not all risk estimates cited above were prepared by the same agency. However, an e¡ort hasbeen made in each pair-wise comparison (e.g., dioxin in milk vs. dioxin in ¢sh) to cite estimatesfrom the same agency or methodology.

25. Rogan et al. (1991: p. 235) compare breast milk (taking into account both bene¢ts and risksassociated with contamination) and formula, and conclude that `there is no clear advantageto avoiding breast feeding; there may be a disadvantage.' Similarly, Hoover (1999) reportscomparable risks and bene¢ts of contaminated breast milk.

26. In practice, however, one cannot count on consumers substituting nutritious alternatives,suggesting that the breast milk case may not be unique. This is particularly true when apopulation is heavily reliant on a particular food, as in the case of children's high consumptionof Alar-contaminated apple products and some Aboriginal communities' heavy reliance ondioxin-contaminated ¢sh.

27. The milk testing had been conducted by the federal government and reported at an inter-national dioxin conference in August 1989, which was attended by Greenpeace sta¡. The

58

report received no attention until Greenpeace held a press conference in October to publicizethe government's ¢ndings.

28. Under the Food and Drugs Act, Canada has prohibited any level of chlorinated dioxins in anyfood sold commercially, with the one (quite illogical) exception of allowing up to 20 ppt of onlythe most toxic dioxin, 2378 TCDD, in ¢sh. With respect to the U.S., Rogan (1996, 981) hasnoted that `breast milk, if regulated like infant formula, would commonly violate [U.S.] Foodand Drug Administration action levels for poisonous or deleterious substances in food andcould not be sold.'

29. Indeed, as discussed in Harrison (1991), the economic relevance of the commercial ¢sheryarguably contributed to less, rather than more, stringent regulation of TCDD in ¢sh.

30. Similarly, a series of stories on PCBs in breast milk in the early 1990s was initiated by theInternational Joint Commission. See IJC, `Contaminant levels in mother's milk amongindicators of ecosystem health recommended by Great Lakes Science Advisory Board,' mediarelease, 3 September 1991; Martin Mittelstaedt, `Lakes report cites risk to infants: Mothers'milk causes concern,' Globe and Mail, 31 August 1991, A1; Martin Mittelstaedt, `No need tostop nursing, MDs say: PCBs detected in human milk,' Globe and Mail, 4 September 1991;Stephen Strauss, `How some politically compromised scientists soured scienti¢c truth,' Globeand Mail, 21 September 1991, D10.

31. Personal interview.32. Philip Shabeco¡, `Dioxin in breast milk is evaluated in private study,' New York Times,

18 December 1987, A17.33. See also Gibbs (1995).34. Greenpeace U.S.A., `PVC/vinyl the ` worst plastic'' for environment,' media release, 22 April

1997; La Leche League, `Earth day statement,' 21 April 1997 and `Earth days news is thatbreastfeeding remains best choice in a polluted world,'April 1997.

35. For instance, when dioxins were reported to have been detected in the breast milk of Canadianwomen for the ¢rst time, a Health and Welfare Canada o¤cial stated `Certainly, for an infantin the short term, there is no problem in terms of toxicity.'

36. Greenpeace campaigners met with sta¡ from M.P. Jim Fulton's o¤ce before releasing theirstory to Canadian Press. Con¢dential interview.

37. House of Commons, Debates, 30 October 1990, 14882.38. House of Commons, Debates, 29 October 1990, 14807. It is by no means clear what statutory

authority the Minister would have relied on to ban breast milk, even if he were serious.39. One physician interviewed who specializes in breastfeeding stressed that `There is zero research

to show even dioxin-laced breast milk is worse than formula. . . . Before we make any changes,we have to be sure the alternative is better.'

40. See, for instance, Palmer (1988: p. 51); La Leche League, `Earth day statement'; NancyMohrbacher, `Breastfeeding and contaminants,' New Beginnings, 2: pp. 128^130, reprinted atwww.lalechleague.org/llleaderweb/LV/LVNBSeptOcto86.text.

41. Con¢dential interview, also Baumslag and Michels (1995: p. 96).42. See, for example, Ken Pole, `Report of dioxins/furans in breast milk a media ¢asco,' Medical

Post, 13 November 1990; and BC Report, `Crying wolf once too often: experts debunk Green-peace's warning about contaminated mother's milk,' 19 November 1990, 24. Although focusedon PCBs rather than dioxins, a similar story about breast milk contamination prompted by anIJC report and press release elicited a scathing response from columnist Stephen Strauss, whoaccused the IJC scienti¢c advisory board of `cynically manipulat[ing] [nursing-mothers'] fearsand feelings' (Martin Mittelstaedt, `Lakes report cites risk to infants: Mothers' milk causesconcern,' Globe and Mail, 31 August 1991, A1; Stephen Strauss, `How some politically compro-mised scientists soured scienti¢c truth,' Globe and Mail, 21 September 1991, D10).

43. Con¢dential interview.44. Colborn responded to a question about EPA's estimate that a breastfed infant receives 4^12%

of its lifetime exposure to dioxins in its ¢rst year of life by saying, `We don't have enoughevidence yet. But I'll tell you quite frankly that I would not want to have to make the decision

59

myself today. It appears that breast-feeding strengthens a baby's immune system, but we alsowonder how these chemicals might be interfering with immune competency in these children.So far, the bene¢ts seem to outweigh the risks, but we just don't know.' `Theo Colborn,'Mother Jones, March/April 1998.

45. `Backtalk,' Mother Jones, May/June 1998. See also Colborn et al. (1996: p. 215).46. Although the pulp and paper industry, as the most signi¢cant and visible industrial source of

dioxins, was indirectly threatened to the extent that concern about dioxin in breast milk mightprompt more aggressive regulation of the industry, their voice was noticeably absent from anystories on dioxin in breast milk. Reporters covering the stories do not appear to have made theconnection, and the pulp and paper industry presumably chose not to weigh in on a potentiallyexplosive debate by dismissing the risks of dioxins, which they had already committed toreduce to non-detectable levels voluntarily.

47. One might argue, however, that politicians were defending their political self-interest byde£ecting attention from an issue that might re£ect badly on their performance in protectingthe environment.

48. A proposal,' December 1990.49. Con¢dential interview. Three years later, when Greenpeace Canada held a press conference to

try elicit interest in dioxin to coincide with the release of the U.S. dioxin reassessment, thepress release made only a passing and indirect reference to breast milk, even though infants'exposure was arguably one of the most explosive ¢ndings in the EPA report.

50. The petition, based on Health Canada's test results, prompted the U.S. FDA, at the Departmentof Agriculture's urging, to conduct its own testing of milk in cartons, and subsequently todemand risk reduction measures from the dairy industry.

51. Personal communication. The thesis was by Lindstrom (1988).52. However, Yalom (1997: p. 141) notes quite disparate rates of breastfeeding among di¡erent

racial groups in the U.S.53. Although obviously a very small sample, the two male environmentalists interviewed were

much more inclined to `go for it' and revealed frustration with the hesitation of their femalecolleagues. One U.S. environmentalist seemed unaware of how gender-loaded was his recounting(in a sarcastic falsetto) of a meeting at which the female Administrator of the EPA asked TheoColborn, `But Theo, what should I tell nursing mothers?' He reported that, while Colbornhesitated, he interrupted with `Just tell them to write a letter to Congress!'

54. An analogy might be made to genetic testing for hereditary diseases.55. Interestingly, the Greenpeace Canada press release did not place much emphasis on breast

milk in comparison to the report itself. However, a Montreal Gazette story based on the reportdid emphasize the breat milk angle. (Greenpeace Canada, `Greenpeace Report Shows ToxicPollution Knows No Boundaries,' media release, 25 June 1998; Montreal Gazette, `Breast milkin some Canadians among world's most toxic: Greenpeace,' Calgary Herald, 26 June 1998,A21). In contrast, WWF-UK's press release is the one example found by the author of arhetorical strategy that seeks to provoke outrage about contamination of breast milk, on thegrounds that women have a right to breastfeed their infants with con¢dence, while simulta-neously stressing the net bene¢ts of breast feeding. (WWF-UK, `Government urged to tackle` hand-me-down'' poisons,' 12 July 1999.)

References

Alaluusua, S. et al. (1999). `Developing teeth as biomarker of dioxin exposure,'Lancet 353: 206.Allsopp, M., R. Stringer and P. Johnston (1998). Unseen Poisons: Levels of Organochlorine

Chemicals in HumanTissues. Greenpeace International.Ayotte, P., G. Carrier and E. Dewailly (1996). `Health risk assessment for Inuit newborns exposed to

dioxin-like compounds through breast feeding,'Chemosphere 32: 531^542.Bachrach, P. and M. S. Baratz (1970). Power and Poverty: Theory and Practice. Toronto: Oxford

University Press.

60

Baumgartner, F.R. and B.D. Jones (1993). Agendas and Instability in American Politics. Chicago:University of Chicago Press.

Baumslag, N. and D. L. Michels (1995). Milk, Money, and Madness: The Culture and Politics ofBreastfeeding.Westport, CT: Bergin and Garvey.

Birkland, T. (1998). `Focusing events, mobilization, and agenda setting,' Journal of Public Policy 18:53^74.

Bosso, C. (1994). `The contextual bases of problem de¢nition,' in D.A. Rochefort and R.W. Cobb,eds.,The Politics of Problem De¢nition: Shaping the Policy Agenda. Lawrence, Kansas: UniversityPress of Kansas.

Canadian Paediatric Society and Health Canada (1998).Nutrition for Health Term Infants. Ottawa:Minister of PublicWorks.

Cobb, R.W. and C.D. Elder (1972). Participation in American Politics: The Dynamics of AgendaBuilding. Boston: Allyn and Bacon.

Cobb, R.W. and M.H. Ross (1997). Cultural Strategies of Agenda Denial: Avoidance, Attack, andRede¢nition. Lawrence, Kansas: University Press of Kansas.

Colborn, T., D. Dumanoski and J. Peterson Myers (1996).Our Stolen Future. NewYork: Dutton.Davidson, D. and W.R. Freudenburg (1996). `Gender and environmental risk concerns: A review

and analysis of available literature,'Environment and Behaviour 28: 302^329.Douglas, M. and A.Wildavsky (1982). Risk and Culture: An Essay on the Selection of Technical and

Environmental Dangers. Berkeley: University of California Press.Flynn, J., P. Slovic and C.K. Mertz (1994). `Gender, race, and perception of environmental health

risks,'Risk Analysis 14: 1101^1108.Gibbs, L.M. (1995).Dying from Dioxin: ACitizen's Guide to Reclaiming Our Health and Rebuilding

Democracy. Boston: South End Press.Government of Canada (1990). Canadian Environmental Protection Act, Priority Substances List

Assessment Report No. 1: Polychlorinated Dibenzodioxins and Polychlorinated Dibenzofurans.Ottawa: Minister of Supply of Services.

Gustafson, P. E. (1998). `Gender di¡erences in risk perception: Theoretical and methodologicalperspectives,'Risk Analysis 18: 805^811.

Great Lakes Science Advisory Board (1991). 1991 Report to the International Joint Commission.Windsor: International Joint Commission.

Harrison, K. (1991). `Between science and politics: Assessing the risks of dioxins in Canada and theUnited States,'Policy Sciences 24: 367^388.

Harrison, K. and G. Hoberg (1991). `Setting the environmental agenda in Canada and the UnitedStates: The cases of dioxin and radon,'Canadian Journal of Political Science 24: 1^27.

Harrison, K. and G. Hoberg (1994). Risk, Science, and Politics: Regulating Toxic Substances inCanada and the United States. Montreal: McGill-Queen's University Press.

Hoover, S.M. (1999). `Exposure to persistent organochlorines in Canadian breast milk: A proba-bilistic assessment,'Risk Analysis 4: 527^545.

Huisman, M. et al. (1995). `Neurological condition in 18-month-old children perinatally exposed topolychlorinated biphenyls and dioxins,'Early Human Development 1995: 165^176.

International Agency for Research on Cancer (1997). Polychlorinated Dibenzo-para-Dioxins andPolychlorinated Dibenzofurans. IARC Monographs on the Evaluation of Carcinogenic Risks toHumans, volume 69. Lyon: IARC.

Jacobson, J. L. et al. (1984). `Prenatal exposure to polychlorinated biphenyls and dioxins and itse¡ect on neonatal neurological development,'Early Human Development 41: 111^127.

Jacobson, J. L. and S.W. Jacobson (1996). `Intellectual impairment in children exposed to poly-chlorinated biphenyls in utero,'New England Journal of Medicine 335: 783^789.

Jasano¡, S. (1986).Risk Management and Political Culture. NewYork: Russell Sage Foundation.Kingdon, J. (1984).Agendas, Alternatives, and Public Policies. Boston: Little-Brown.Koopman-Esseboom et al. (1996). `E¡ects of polychlorinated biphenyl/dioxin exposure and feeding

type on infants' mental and psychomotor development,'Pediatrics 97: 700^706.Lindstrom, G. (1988). Polychlorinated Dibenzo-p-Dioxins and Dibenzofurans: Analysis of and Oc-

currence in Milk. Umea, Sweden: Department of Organic Chemistry, Umea University.

61

Lyons, G. (1999).Chemical Trespass: AToxic Legacy. Surrey: WWF-UK.Milly, P. and W. Leiss (1997). `Mothers milk: Communicating the risk of PCBs in Canada and the

Far North,' in D. Powell and W. Leiss,Mad Cows and Mother's Milk. Montreal: McGill-Queens.Palmer, G. (1988).The Politics of Breastfeeding. London: Pandora Press.Patandin, S. et al. (1998). `E¡ects of environmental exposure to polychlorinated biphenyls and

dioxins on birth size and growth in Dutch children,'Pediatric Research 44: 538^545.Patandin, S. et al. (1999a). `E¡ects of environmental exposure to polychlorinated biphenyls and

dioxins on cognitive abilities in Dutch children at 42 months of age,' Journal of Pediatrics 134:33^41.

Patandin, S. et al. (1999b). `Dietary exposure to polychlorinated biphenyls and dioxins from infancyuntil adulthood: A comparison between breast-feeding, toddler, and long-term exposure,' Envi-ronmental Health Perspectives 107: 45^51.

Ralo¡, J. (1999). `Dioxin can harm tooth development,' Science News Online, http://www.science-news.org.

Rochefort, D.A. and R.W. Cobb, eds. (1994).The Politics of Problem De¢nition: Shaping the PolicyAgenda. Lawrence, Kansas: University Press of Kansas.

Rogan,W. J. (1996). `Pollutants in breast milk,'Archives of Pediatric and Adolescent Medicine 150:981^990.

Rogan,W. J. and B. C. Gladen (1991). `PCBs, DDE, and child development at 18 and 24 months,'Annals of Epidemiology 1: 407^413.

Rogan, W. J. and B. C. Gladen (1993). `Breast-feeding and cognitive development,' Early HumanDevelopment 31: 181^193.

Rogan,W. J., P. J. Blanton, C. J. Portier and E. Stallard (1991). `Should the presence of carcinogensin breast milk discourage breast feeding?'RegulatoryToxicology and Pharmacology 13: 228^240.

Salter, L. (1988). Mandated Science: Science and Scientists in the Making of Standards. Boston:Kluwer Academic Publishers.

Shecter, A., J. J. Ryan and O. Papke (1998). `Decrease in levels and body burden of dioxins,dibenzofurans, PCBs, DDE, and HCB in blood and milk in a mother nursing twins over athirty-eight month period,'Chemosphere 37: 1807^1816.

Schneider, A. and H. Ingram (1993). `The social construction of target populations: Implicationsfor politics and policy,'American Political Science Review 87: 334^347.

Slovic, P., B. Fischo¡ and S. Lichtenstein (1979). `Rating the risks,'Environment 21: 14^20, 36^39.Stone, D.A. (1988). `Causal stories and the formation of policy agendas,'Political Science Quarterly

104: 281^300.Stone, D.A. (1997). Policy Paradox: The Art of Political Decision Making. NY: WW Norton and

Company.Sullivan, M. J., S. R. Custance and C. J. Miller (1991). `Infant exposure to dioxin in mother's milk

resulting from maternal ingestion of contaminated ¢sh,'Chemosphere 23: 1387^1396.Thompson, K.M. and J.D. Graham (1997). `Producing paper without dioxin pollution,' in J.D.

Graham and J.K. Hartwell, eds., The Greening of Industry: A Risk Management Approach.Cambridge, MA: Harvard University Press.

U.S. Environmental Protection Agency (1994a). Health Assessment Dopcument for 2,3,7,8-Tetra-chlorodibenzo-p-dioxin (TCDD) and Related Compounds. External Review Draft.

U.S. Environmental Protection Agency (1994b). Estimating Exposure to Dioxin-Like Compounds.Vol. 1: Executive Summary. External Review Draft.

Van Strum, C. and P. Merrell. No Margin of Safety: A Preliminary Report on Dioxin Pollution andthe Need for Emergency Action in the Pulp and Paper Industry. Greenpeace U.S.A.

World Health Organization (1988). Assessment of Health Risks in Infants Associated with Exposureto PCBs, PCDDs and PCDFs in Breast Milk. Copenhagen: WHO.

World Health Organization European Centre for Environment and Health (1999). `Contaminationwith dioxin of some Belgian food products,' http://www.who.dk/envhlth/dioxin/dioxin.htm,updated 15 June.

Yalom, M. (1997). AHistory of the Breast. NY: Ballantyne.

62