The research began with a review of available literature, organised on a regional basis. This had two justifications. In the first place, there appears to be no universal structure of ideas informing discussions on the commercial sexual exploitation of children, which is dispersed among a variety of agencies with a number of disparate objectives and activities varying from law and advocacy to welfare, and even arguments in favour of paedophilia. In the second place, it was clear that certain aspects dominate the discourse in specific regions of the world. The research thus began with a twofold purpose, examining existing discourses for what they might provide in the way of well argued, internally-consistent structures of ideas. In this respect, it has to be made clear from the outset that by 'discourse' we mean clearly distinguishable sets of ideas, publications, speeches and other social products that inform and construct the way people think and act. Any discourse on child commercial sexual exploitation will be related to other discourses -- on childhood, sexuality, exploitation and prostitution, for example. It will produce and reproduce these ideas in ways that tend to reinforce current structures of power and hierarchy. The first task for this review, therefore, was to examine not simply the evidence about the commercial sexual exploitation of children but, more importantly, how it is being produced, reproduced and presented.

1.1. The regional approach

South East Asia, particularly Thailand and the Philippines, is the key to any discussion of the commercial sexual exploitation of children because it was the situation in this area in the past two decades that raised public awareness of the commercial sexual exploitation of children and mobilised public opinion against 'child sex tourists'. The discourse derived from these two countries has set the parameters and tone of the debate. The issues debated with respect to Thailand and the Philippines have become key to the global discussion of child prostitution to the extent that it is impossible to talk about commercial sexual exploitation without reference to them. Thailand and the Philippines have provided much of the mythology and iconography of the commercial sexual exploitation of children so that it is important to look at these two countries in detail in order to understand the origins and the boundaries of the issue.

The role of non-governmental organisations (NGOs), in particular End Child Prostitution in Asian Tourism (ECPAT), in raising awareness cannot be underestimated. It is they who first brought the issue to public attention, set the agenda and have continued to dominate the debate. Although the countries for which they mainly campaign are Thailand and the Philippines, they are also concerned with Sri Lanka, Taiwan and Korea. Their work is widely disseminated through both the Western media and national English language press. These NGOs also publish articles widely in the popular press and ECPAT is responsible for two influential books, The Child and the Tourist (O'Grady, 1992) and The Rape of the Innocent (O'Grady, 1994), popular paperbacks designed to appeal to a non-specialist audience. Drawing on ECPAT's success, other groups now campaign against child prostitution but ECPAT remains a central information source. It is rare to find an article on child sexual exploitation in a Western newspaper that does not make reference to ECPAT.

The content of this information follows predictable formulae. A typical example would be a case study of a very young girl forced or tricked into a brothel where she is obliged to service 20 customers a night for very low remuneration. In the story she will be rescued by a welfare agency and sent back to her village, only to discover that she has contracted HIV and will shortly die. It will be stated or implied that it is demand from Western men that causes her to become a prostitute. Aspects of degradation and abuse are repeatedly emphasised, as is the youth of the girl. The language used is often emotive. For example, one report describes 'the lifeless body of an eight year old child, left in a Saigon hotel room after a night of sexual abuse' (ECPAT Newsletter, 1995), while a campaigner told the press 'I still remember vividly the tears in the eyes of the child rescued from a Bangkok brothel who told me how she begged a customer not to harm her, only to have her pleas mercilessly rejected' (Bangkok Post 6/10/93).

In the current campaigning literature, both within South East Asia and in the West, the image of child commercial sexual exploitation is of small children being sought out and exploited by Western tourists. Much press coverage now is concerned with finding and punishing these men (and it is assumed that all sex tourists are male in this geographical region, even though there are reports of female sex tourists in other areas). Yet around the slums of the port of Klong Toey in Bangkok, men who cannot not afford a 'real' (meaning fully grown) woman will find a young girl a reasonable substitute because she is cheaper and easier to control. While this is an aspect of child prostitution that many campaigning groups in Asia would prefer to ignore, it is likely that the majority of young prostitutes are not found in the bars of Bangkok or Manila but in the brothels in the rural areas or the back streets of cities. In many local brothels in Thailand and the Philippines, younger women are said to be prized for their innocence and freshness, while girls even younger are prized for their cheapness (Ennew, 1986; Black, 1994). Even among the better-off Thai men, there is a marked preference for younger girls. A 1994 survey conducted among students, office workers, and residents of a slum area, to assess the impact of HIV on children found that the most desirable age for prostitutes is under 18: 'Many males felt that child prostitutes between 15 and 18 were more desirable than adults, but that it was wrong to sleep with younger ones (under 14)' (Sittitrai and Brown 1994, p. 4).

In other parts of South East Asia, although the discourse is not as influential on the world stage, the issue is also often constructed on the basis of the idea of the foreign male exploiter. There is less information on Vietnam, Cambodia and China but these three countries share the language of 'social evil', which is behaviour or ideas that are contrary to and damaging for national culture. Child prostitution, like AIDS, is described as a problem imported from other cultures, not always the West, or considered in terms of trafficking so that Vietnamese girls, for example, are forced across the border to Cambodia or China without any acknowledgement of indigenous prostitution. China however, adds another component to the picture. Chinese girls are said to be the victims of traffickers, especially young girls from the Yunan region (Centre for the Protection of Children's Rights, 1991) but Chinese men, rather than foreigners, are often blamed for the problem. It is also repeatedly claimed that Chinese men will pay to have penetrative sex with a virgin because this is believed to be a cure for AIDS (Muntabhorn, 1992; O'Grady, 1992; 1994). This is a common assertion in the literature worldwide, even though there is no ethnographic evidence to prove it. In Vietnam and Cambodia in particular, anti-Chinese feeling leads to the repetition of this notion (Thang, 1996).

In addition to the accounts of NGOs and journalists, a relatively small amount of information is produced by anthropologists and sociologists who have worked on limited research projects (Truong, 1982; 1986; 1990; Hantrakul, 1983; Muecke, 1992; Mccaughy & Hou, 1994; Care, 1994; Lie, 1995). The resulting articles and books are concerned with the cultural background, particularly religion and societal norms that can be predisposing factors in encouraging or discouraging children to become prostitutes. In this respect a major issue is the apparent religious sanction given to prostitution by Buddhist values, which stress models of duty and sacrifice for children especially for girls. One argument is that by supporting her family through prostitution a girl gains merit rather than bringing shame on herself and her family, a justification that is frequently manipulated in some areas of the literature. However, these texts are usually published in the academic Western press, with a relatively small circulation among other academics.

One problem in the NGO literature is that the academic literature seems to be largely ignored or unknown. In addition, within the mainstream, campaigning literature, certain categories become blurred. Thus it is a feature of the reporting that:

€child prostitutes are often by implication only girls;

€pre-pubertal and post-pubertal children are often included in the numbers given for child prostitutes along with young women over the age of 18 years;

€numbers given for Western tourist clients are confused with numbers of Western tourists as a whole, with no account given of local clients;

€within the undifferentiated category of child prostitute, the origins of child prostitutes are hidden, obscuring aspects of origin, such as ethnic or socio-economic factors.

A second major geographical focus is South Asia -- India, Nepal, Sri Lanka and, to a lesser extent, Bangladesh and Pakistan. This has three components, the girl child, religious prostitution and trafficking for sexual purposes. The idea of the girl child arose largely through Indian attempts to assign a special space in social philosophy and policy to girls (Williams, 1991). 1990 was named The Year of the Girl Child by the South Asian Association for Regional Cooperation (SAARC) and the UNICEF India Office. Two main themes of the year were prevention of child marriages and rehabilitation of child prostitutes. One Indian woman's comment on the material published in the Year of the Girl Child shows the reproductive nature of the information purveyed in the resulting 'spurt of publications':

Many of the papers are clones of earlier ones. The self perpetuations occur through citations being made circular, until a body of knowledge is assumed to have been created. The sum and substance of much of this writing concern the low status of the girl, her limited opportunities for education and the gender bias in the home. The television screen has also sensitized people to the plight of the girl child as a drudge (Aanandalakshmi, 1991, p. 29).

This was not the intention of 'girl child' campaigners such as Sheela Barse, who makes it clear that she is opposed to the 'feminist method of carving out the female persons' group out of the human race and examining it in isolation, on the presumption that females are always wronged' (Barse, 1991, p. 99). Barse states that it is more important to establish that girlhood cannot be the entire province either of womanhood in the women's rights movement, or of childhood in the children's rights movement.

Nevertheless, the notion of the 'girl child' has been reified, as if there is something essential about all female children regardless of wealth or ethnicity, something that makes them more vulnerable than boys in general. This focuses on vulnerability to sexual violation, early marriage, unplanned teenage pregnancy and is a new manifestation of the old control over women's sexuality and fertility. It masquerades as concern for their vulnerability but actually implies that females cannot control their own sexuality, which should consequently be under male control. This is reflected in three approaches to development aid programming for girls:

The link between female education and child survival;

The emphasis on the employment of girls as prostitutes, which is a minor exploiter of girls compared to agriculture, domestic service and manufacture;

The immense silence about the specificity of certain kinds of exploitation that are exclusive to boys.

In India and in Nepal the 'girl child' discourse focuses on religious prostitution (of which devadasi is the best known form) and trafficking in children, while in Bangladesh, the rape of pre-pubertal girls hired as domestic servants is an important theme (Blanchet, 1996, p. 119) . In Pakistan, Baluchistan and the North West Frontier provinces of India, much of the literature on the girl child is legal (see for example Jahangir, 1986). A further concern is the fate of the daughters of prostitutes who on account of social opprobrium are said to be forced into prostitution (Patkar, 1991).

Religious prostitution is practised in various parts of India and Nepal. Devadasi cults are found in Southern India and also practised in other parts of the country such as Uttar Pradesh and Orissa. They derive customary sanction from oppressive upper-caste temple traditions. Pre-pubertal girls, aged between five and nine years, from poor, low-caste homes, are dedicated by an initiation rite to the deity in the local temple during full moon. After a girl is married to the deity by the tali rite, she is branded with a hot iron on both shoulders and her breast. She is then employed by the temple priest. Sometimes, even before menarche, she is auctioned for her virginity; the deflowering ceremony known as udilumbuvadu becomes the privilege of the highest bidder. The market value of a girl falls after she attains puberty, when she is said to have no recourse other than prostitution. Yellama is represented as the principal goddess who is worshipped but, as recent research has shown, the practice of devadasi is prevalent in many other temple towns and other deities such as Meenakshi, Jaganath and Hanuman are also propitiated. Religious prostitution is known by different names such as venkatasani, jogini, nailis, muralis and theradiyan (Bahni, 1989; Marglin, 1985; Mowli, 1992; Story, 1987).

In Nepal, particularly the western parts, religious prostitution known as badini and jhuma is also practised. Although little is known about these practices, they do not seem to vary significantly from devadasi. However, the main sexual exploitation issue in Nepal is the traffic with India, the open border between the two countries making it difficult to monitor (Human Rights Watch, 1995; O'Dea, 1993; Rozario, 1988). It is also stated that, because of corruption, official assistance is given to the sale and the trafficking of young girl (Agroforestry Report, 1990). In most accounts Bombay is given as the main destination. The international trade routes most frequently mentioned are from Nepal and Bangladesh to India, and from India and Pakistan to the Middle-Eastern countries.

In sharp contrast to the emphasis on the 'girl child' in the rest of South Asia, in Sri Lanka the discourse revolves around boys and sex tourism. ECPAT and other advocacy organisations have influenced the discourse here, claiming that male western paedophiles are targeting Sri Lanka because of the availability of boys. For instance, it is argued that, despite the civil war, the number of tourists who visited Sri Lanka increased from 102,000 in 1989 to 169, 000 in 1996, although of course this does not necessarily mean that all tourists are male, let alone sex tourists. According to the literature, the South West coast, Negombo and Hikkawuda are the main destination points. They are close to the beach and some hotels assist paedophiles in procuring boys mostly aged between six and 15 years, although most boy prostitutes in Sri Lanka appear tend to work independently of either brothels or pimps (Bond,1981; Goonesekere & Abeyratne, 1986; Seneviratne, 1991). Although there is evidence of sex tourism in Sri Lanka, the picture is not clear because of the tendency to reproduce information. One case in point is the frequent reference to the role of Spartacus guides, which provide male homosexual tourists with up to date information about the availability of sexual contacts in most countries, and were notorious in the 1980s, particularly with respect to both Sri Lanka and the Philippines, for giving locations where boy prostitutes could be encountered. Despite the fact that the magazine no longer explicitly refers to children, and the then publisher of Spartacus has died, the magazine's role in the sex tourism business is constantly reiterated in popular literature on the commercial sexual exploitation of Sri Lankan boys as if no changes had occurred.

A third major geographical region with a characteristic discourse that influences global debates might be designated the Anglophone West and consists the United States, Canada, Australia and Western Europe. From the late nineteenth to early twentieth century there was considerable public concern about the 'White Slave Trade'. Many aspects of this earlier discourse are clearly visible in current debates. Nevertheless, modern concern about what is always referred to in the academic literature as 'juvenile' rather than 'child' prostitution resurfaced in the late 1960s alongside the issue of runaway children. Journalists were quick to use labels like baby-pros or, with respect to pornography, 'kiddie porn'. Then, the concern was for children who joined the 'hippie' culture of America's big cities, in particular San Francisco and New York (Deisher et al, 1969; Weisberg, 1985). Britain and Canada, where many of the same social problems as America were reported, but a few years later, experienced similar concerns in the early to mid 1970s (Sereny, 1984; Donovan, 1992).

The earliest indications of a juvenile prostitution 'problem' can be found in the mass media, but academics were also involved from the very beginning (see for example Deisher et al, 1969). In addition, since the late 1960s there have been a number of moral panics about child abuse in North America and Western Europe. Thus interest in juvenile prostitution has come in waves, eclipsed at times, by concerns about 'battered children' or more recently 'satanic' or 'ritual' abuse (La Fontaine, 1990; Jenkins, 1992; Joseph, 1995). Thus, there have been few long term studies of juvenile prostitutes, but rather a flurry of papers and articles at times when interest in the issue is high. Currently, concern is focused on young male prostitutes, because they are seen as vectors for the spread of HIV and research into their lives is conducted along bio-medical models that are concerned with certain areas of behaviour and particular attitudes (National Consultation on Adolescent Prostitution, 1984; Pleak et al, 1990; Snell, 1995). Both academics and journalists have remained interested in the issue of delinquency and deviancy psychology, which it is assumed many young prostitutes share but, again, this is linked to bio-medical models with particular perspectives and limited range (Baizerman et al 1979; Davidson & Loken 1987). Western journalists have a major role in disseminating information about the commercial sexual exploitation of children. Articles on young male prostitutes, often referred to as 'rent boys' in the English language press, as well as young female prostitutes, remain an occasional feature of many newspapers and women's magazines and are presented as campaigning or investigative journalism, while revealing prurient details designed to shock and sensationalise. Popular books written on the issue of child prostitution and pornography are mostly written by journalists with a stated mission to reveal the truth. Playlands (Lloyd, 1977) and Child Pornography (Tate, 1991), two influential books on the subject were both written by journalists and one of the most famous books, H: Autobiography of a Child Prostitute and Heroin Addict, was written in collaboration with journalists from the German newspaper, Stern ('F', 1981).

Due to the ease of access and a history of journalistic interest, the material is very heavily urban based, with San Francisco (Weisberg, 1985), New York (Allsebrook & Swift, 1989), London (Sawyer, 1988), Birmingham (Donovan, 1992), Melbourne (Muntabhorn 1993),Amsterdam (Tate, 1991; Donovan, 1992) and Berlin ('F', 1981) over-represented. Other cities with known groups of child prostitutes, such as Dallas, Washington DC, Sydney, and in the United Kingdom, Brighton and Cardiff, are not discussed in nearly the same detail. In the USA especially, there are certain centres of expertise on child prostitution such as Huckleberry House in San Francisco and Covenant House in New York, with long campaigning and advocacy histories, so that it can be easier to gain access to both children and experts in the 'treatment' of abused children through these types of organisation. In Canada, particular emphasis has been placed on Ontario, not because of the higher concentration of juvenile prostitutes there, but because the Queen's University Social Program Evaluation Group has taken particular interest in the problems facing homeless and prostitute youth (Radford et al, 1989).

In the USA, the source of a good deal of the literature is part of the university sector that specialises in 'objective' non-judgmental studies of AIDS awareness and/or delinquent psychology (Nightingale n.d.; Baizerman et al 1979; Davidson & Loken 1987; Widom and Ames, 1994). In the USA much emphasis is placed on the 'runaway' phenomenon, first noted in the 1960s and now closely tied to the prostitution discourse (Lloyd, 1977; Sereny, 1984; Schaffer & Deblassie, 1984). The current prostitution problem continues to be seen as having its roots in the alternative culture movements of the late 1960s, with a special emphasis on the hippie communities in San Francisco (Weisberg 1985). It is claimed that many children ran way to join these communes but, on leaving them, found themselves unable to make any money other than through prostitution (Schaffer & Deblassie, 1984).

In Canada, the emphasis has been on AIDS prevention and protecting children for their own good and that of society (Lowman, 1987). It is notable that the Canadian literature concentrates on Canadians of European descent (National Consultation on Adolescent Prostitution, 1984). Native Canadians are seldom mentioned, despite the fact that they tend to show higher-than-average rates of what are regarded as predisposing factors such as coming from broken homes with a history of drug or alcohol use, or having been placed in state institutional care. Yet in the rest of the literature, the multiple problems that lead children into prostitution are emphasised (Lowman, 1987). This in direct contrast to United States research, where race or ethnicity are especially important markers. No study of juvenile prostitutes in the USA is complete without a breakdown of prostitutes' racial backgrounds and a discussion of what this might mean (Weisberg, 1985; Gibsonainyette et al, 1988).

In the United Kingdom, the stress is on boy prostitutes. 'Rent boys' have become a staple of the British media and even the more serious academic studies, have tended to concentrate more on boys than on girls, despite the smaller numbers of boy prostitutes. The West Midland Police, which covers Birmingham where there is a major red light district, commissioned a report on young male prostitutes and came up with a report based on a sample group of less than 20 (Donovan, 1992). While the police have concentrated on boy prostitutes, the advocacy groups discuss young children of both sexes, highlighting the problems they experience when they leave institutional care and the lack of support they are given. NGOs such as the Children's Society in the United Kingdom have published papers and articles suggesting that it is both lack of institutional care and the brutalising effect that many children's homes have on their inmates that contribute to their recourse to prostitution when they are discharged or escape (Lee & O'Brien, 1995).

A further contributory factor stressed in Western literature is the role of broken homes (Finkelhor, 1979; Sereny, 1984; Weisberg, 1985; Lowman, 1987; Gibsonainyette et al, 1988; Campagna & Poffenberger 1988; Allsebrook & Swift, 1989; Widom and Ames, 1994). Many of the juveniles surveyed have suffered sexual and physical abuse within the family and many are runaways from abusive situations (Finkelhor, 1979; Weisberg, 1985; Lowman, 1987; Snell, 1995). In Britain, the emphasis is placed more on children who have been in state institutional care rather than those from abusive families, yet the literature remains framed within the discourse on dysfunctional families (Lee & O'Brien, 1995). Throughout almost all the books and articles on the subject in the West, runs the theme that these children are outside society, and that reasons have to be found for their deviancy. Their life histories are presented in terms of theories of deviancy. There is a notable absence of views of the children themselves or, when literature does include their opinions, this is often countered by an authorial voice giving reasons why they are wrong.

The particular emphasis on boy prostitutes is related to a concern with the mental health of sexually exploited children in general. In the case of boys, there is considerable discussion about whether these boys are homosexual or 'really' heterosexuals whose commercial sexual activities are focused on financial gain. One concern frequently expressed is that heterosexual boys with homosexual clients may become gay by being prostitutes, resulting in a good deal of discussion about exactly what boys will let there partners do to them and whether they take an 'active' or 'passive' role in sexual activities. A parallel debate is entirely absent in the literature on female prostitution. Nor is there any information about young men who have female clients (Lloyd, 1977; Donovan, 1994). The only similar discussion in literature on female prostitution is whether or not girl prostitutes can become good mothers. The underlying implication is that, whether they are boys or girls, juvenile prostitutes will not have learned appropriate gender roles.

Literature on the poor psychological health of young sex workers also seems to take for granted they suffer from low self-esteem, suicidal tendencies and the inability to form relationships, generally without exploring scientifically the causal relationships involved (Finkelhor, 1979; Baizerman et al, 1979; National Consultation on Adolescent Prostitution, 1984; Davidson & Loken, 1987). The assumption is that low self-esteem results from prostitution, rather than that some juveniles become prostitutes because their self-esteem is low. One dissenting voice can be heard in the published words of a woman sex worker,

I also find it very interesting that [the authorities] look at the child prostitute, and they say the problem is prostitution. They forget the problems of theft, drugs, or just general exploitation of youth on the street... It's bordering on criminal for officials to try and say that prostitution is responsible for this....Prostitution is a symptom of a greater problem that these children have experienced that put them on the street in the first place (Bell, 1987, p. 26).

Many texts stress degrading aspects of child prostitution, including being forced into being prostitutes, raped by pimps, terrorised by gang members and becoming dependent on drugs (Sawyer, 1988; Tate, 1991). There is constant reference to the apparently inevitable links between prostitution and heroin use. Yet there is little information about the long-term effects in adulthood because there is no systematic research on the results of prostitution in childhood, simply the impression from reiterated assumptions that juvenile prostitutes end up either dead or living worthless and useless lives (Campagna & Poffenberger, 1988). Some accounts of the lives of boy prostitutes suggest that the average length of time as a prostitute is between two and seven years (Donovan, 1986; Snell 1995) but longitudinal studies seem not to be carried out.

A further concern for those writing about young girl prostitutes is the part played by males who live off their earnings, with far more attention paid to this than to the role of female adult exploiters. Pimps are almost always portrayed as vicious and evil psychopaths, and the fact that many girls speak fondly of their 'protectors' is explained as co-dependency. Even though not all prostitutes work for pimps and some men living with prostitutes are part time prostitutes themselves, the overwhelming impression given of pimps is that they are older, manipulative men (Lowman, 1987). The assumption that all girl prostitutes must be controlled in this way is, of course, a reflection of overall societal assumptions about the vulnerability of women and the need to police their sexuality.

These three discourses, from South East and South Asia and the West represent the majority of studies on the commercial sexual exploitation of children. In each case the discourse on children and prostitution is constructed within structures of ideas about society, childhood, gender and sexuality. In South East Asia the literature on children and prostitution tends to dominate current discussion of childhood. There is no developed literature on children and the social phenomenon of childhood, but a long tradition of constructing gender and sexuality on the basis of an image of passive, childlike women as well as defining cultural norms in opposition to external or foreign evils. The current debates about female children in South Asia, subsumed in the incorrect, essentialist notion of the 'girl child' have been providing a trenchant feminist critique of the unequal status of women in society. This stresses the more vulnerable aspects of female sexuality and thus not only emphasises the oppressive aspects of tradition, through concentrating on the image of girl children as temple prostitutes, but also pivots on the notion of women and girls as objects, as in trafficking. Western discourses on the commercial sexual exploitation of children, on the other hand, are more individualistic and thus concentrate on the deviant behaviour and mental health of individual cases, as well as a concern with the causal relationship between deviant, abusive families and prostitution, showing overall societal concerns only in with respect to the proper assimilation of gender roles.

In contrast, the literature on child sexual exploitation in Africa and in Latin America is far less developed and cohesive. In both areas, the literature on childhood has different emphases.

In so far as they exist at all, child studies in Africa, focus on child health and construct the idea of children as victims, without any particular focus on commercial sexual exploitation of children. Indeed, it is often claimed that this term cannot be applied in most African contexts because the distinction between sexual abuse, sexual exploitation and commercial sexual exploitation can not be clearly drawn, not only in analytical terms within studies, but also within cultural understandings.

Africa's diversity makes the definition of childhood in itself a research issue. The African social science community, includes very few specialists on children's issues outside traditional concerns with health, education and psychology. The main themes in academic research on children in Africa carried out by Western researchers have traditionally been socialisation and initiation or puberty rites. In the nexus between academia and programme makers, a good deal of more recent research has concentrated on medical anthropology, with an interest in traditional health practices in child care and nutrition.

African social science researchers with an interest in child studies now tend to focus on the broad area of child abuse, although, as will be argued below the definition has a particularly African texture. A further major interest in all circles, in research largely dominated by Western researchers, is HIV/AIDS. In the African context this concerns children rather more than it does in other regions of the world, partly because the pattern of infection has long been recognised to be heterosexual, thus affecting children through vertical transmission; partly because of the existence of a relatively large number of 'AIDS orphans', principally in East Africa. A number of mostly Western psychologists have also been studying children affected by armed conflict, concentrating on aspects of traumatisation and victimisation (see for example, Dodge & Raundalen, 1987).

The considerable upheavals due to wars, conflicts, natural and man-made disasters, together with mounting impoverishment among African populations, increasingly affect African children. Yet very few socio economic studies of these effects have taken place. Research on children appears to be fragmented and there are few outlets for publication of research on children's issues. Two recent annotated bibliographies of studies of African children and childhood provide no references to published work on sexual abuse, much less on commercial sexual exploitation of children (Gueye 1995; Ross 1995).

Within this context, the topic of the sexual exploitation of children is part of an overall emphasis on children as victims, fitting within the concerns of a relatively-well developed discourse on child abuse and neglect. This latter is largely the outcome of the activities of ANPPCAN, the African member of the International Society for the Prevention of Child Abuse and Neglect. Based in Nairobi and with national chapters in many African countries (largely Anglophone), ANPPCAN has been active in promoting research in this area as well as advocacy for children's protection rights.

It has to be said, however, that the concept of 'abuse' used by ANPPCAN, and entering the African literature, is not structured in the same way as used in the general Western literature, which can lead to confusion between researchers from different regions. The presentation given by an ANPPCAN functionary in Nairobi cited by Dallape (1988, pp. 104-8) makes this very clear. Child abuse is seen as a feature of other social phenomena or situations, rather than as a phenomenon in its own right:

'The following are the areas where child abuse is commonly evidenced:

Child labour

Children in prison

Handicapped children

Battering of children

Children under psychological stress

Abandoned children

Children in war situations.' (ibid, p. 104)

The field of sexual abuse in Africa cannot be separated from the literature of largely-Northern, feminist campaigners against female excision and infibulation (see the bibliography by Passmore Sanderson, 1986). This is likewise implicated in the anthropological literature on initiation (see La Fontaine, 1985, for an over-view). Although campaigners sometimes like to include these debates within the general field of sexual exploitation of children, this is conceptual nonsense. It does not advance the purposes of advocacy to obscure issues by merging disparate concerns with different root causes.

The study of child sexual exploitation in Africa cannot be separated conceptually from constructions of sexuality and sexual morality. It is notable that the literature shows a particularly marked distinction not only between the behaviours expected of boys and girls, but also the expressed attitudes towards sexual morality of both groups. A further focus of study in this area is the migration nexus, together with the contrasts drawn between rural morality and town morality, and between generations. This is also situated within the overall massive population movements within the continent in the face of natural disaster, war and impoverishment. Nevertheless, although there is considerable anecdotal evidence of child sexual exploitation resulting from these types of situation in which children are rendered particularly vulnerable to the misuse of adult power, hard facts and properly-conducted research are difficult to come by.

In terms of academic discourse, or even within NGO and IGO literature, the topic of child sexual exploitation in Sub-Saharan Africa, consists of an almost total vacuum, in which dispersed and disconnected items of journalistic and project-oriented text are floating aimlessly. The vast majority of the latter material is unpublished.

Although there is a considerable body of literature on children in Latin America it is dominated by the discourse on so called 'street children' (Rizzini, 1996). It is worth noting that the Latin American model of street children has tended to dominate work in this area in all developing countries, largely because the ideas (but not the texts) have been disseminated by Western aid agencies (Ennew, 1996). Children who live on the street often engage in sexual relationships, with each other and in the course of prostitution, which may be occasional or virtually full time for both boys and girls. Thus the literature on the commercial sexual exploitation of children often seems to be merged with that on street children. The street children discourse in Brazil and Colombia in particular is characterised by ambivalence (Aptekar, 1988). Street children are worthy of pity if young, and feared if older adolescents. In either case they are often stigmatised and the literature, frequently dominated by writers from outside the region, is often lurid and lacking in any kind of academic rigour (see for example Meunier, 1977; Agnelli, 1986; Bridel & Collomp, 1986; Dimenstein, 1991). The HIV/AIDS complex has largely overwhelmed the literature on sexuality since the late 1980s, together with a specific interest in street girls in Brazil, due partly to the influence of two charismatic female project directors. The literature on this issue is on the one hand largely tied to project publicity and on the other to medical discourses.

More recently, psychologists in several Latin American countries have shown an interest in studying child sexual abuse, which now constitutes a proper field of study and is resulting in some interesting publications. In general, but to a less hysterical extent, Latin America is going through the same kind of discovery of child sexual abuse that occurred in the USA and Europe in the late 1980s.

Prostitution itself is enmeshed in a series of ideas about men and women. These include the idea that men's sexual appetites are uncontrollable and have to be satisfied by women who enjoy sex (prostitutes) rather than by 'good women' (mothers, daughters, wives) who do not enjoy sex. This is related to ideas of honour and family. A man's honour is tied up in the chastity or purity of the females in his family, wives, sisters and daughters. In its most macho form, masculinity entails protecting the honour of the women in your own family while proving your virility by manifestations of virility (see the papers in Pescatello, 1978). This means that female prostitutes perform an important role with respect to families. Both journalism and research tend to reproduce the same historical accounts that justify the existence of prostitution as a necessary social evil that protects the purity of mothers, daughters and wives and thus ensures the continuing existence of the family (Arnold, 1978). However, it should be noted that this role is exclusive to female prostitutes, many of whom, according to most accounts, begin this work around the age of 15 years (see for example Cairo, 1967; Alves-Milho, 1977). As the majority of street children are boys, the prostitution in which they are engaged belongs within an entirely different complex of social ideas, which does not seem to have been researched in depth. The overwhelming majority of studies of children and prostitution in Latin America is concerned with female children.

There are two resounding absences within the literature on the commercial sexual exploitation of children on the world scene: Eastern Europe and Arab/Islamic countries. However these absences occur for different reasons. Such literature as does exist in Islamic societies is dominated by legal considerations, which is to a certain extent replacing a culture of denial. Whereas in the past the tendency was to comment that child abuse and exploitation is forbidden by the Koran and therefore does not take place, there now seems to be evidence of considerable reflection on children's issues in general and exploitation and abuse in particular. This remains within religious paradigms, but is adding considerably to understanding of the attitudes of Islam to both children and sexuality (see for example Risaluddin, 1996)

In so far as any link might be said to exist between the literature on child sexual exploitation in Arab countries and Eastern Europe and the former Soviet Union is the dominance of lurid journalism on the topic of trafficking in girls and women. There is an established link in sex trafficking between Arabs and South Asian countries such as India and Bangladesh. One estimate suggests that 200,000 girls from Bangladesh were taken against their will to work in Arab brothels (International Children's Rights Monitor 5 (2-3); 10 (1 2)). However the data are anecdotal.

With the sudden economic and social changes of 1989 families in all the former socialist countries of Eastern Europe and the former Soviet Union have been faced with the collapse of welfare systems and the burden of child support has largely shifted to families, many of which are undergoing severe hardship. Children have been particularly badly affected and accounts by journalists suggest that there has been a sharp increase in child sex exploitation (see for example June 21, 1993, Time). The evidence is extremely patchy and usually sensational, telling for example of the longest in the world situated between Berlin and Prague called the 'Highway of Cheap Love' managed by the 'Chechen Boys,' where 'kids are sold like a kilo of bread'.

Box 1: The press view of the commercial sexual exploitation of children worldwide AMSTERDAM, Mar 26 (IPS) - Beneath New York's endlessly high buildings, in the glare of the neon lights and fancy storefronts which carve that city's image, is a hauntingly sombre reality: thousands of children selling their bodies to survive. Halfway across the globe, in Kenya, a police raid in a Nairobi hotel in January found West African and Japanese men frolicking with nine girls aged between 12 and 16. In Sierra Leone, a nation ravaged by conflict, few tourists come to visit these days. Those who do find themselves being solicited by legions of young victims of civil war and poverty willing to provide sex for the price of a meal. Economic need and the insatiable appetite of a new breed of tourists are pushing downwards the age of participants in the flesh trade. No longer is sex an 'attraction' promoted through subtle messages: it has moved up front, and children play a conspicuous role.

1.2. The nature of evidence in current literature

The single most important factor uniting the literature on the commercial sexual exploitation of children is the way in which information is managed. There are few examples of rigorous research or of data presented within a comprehensible cultural context. Rare examples of good research practice do exist, such as the study of girl prostitutes in Costa Rica carried out by Tatiana Treguear and Carmen Carro in which both methodology and method are laid bare and it is possible to judge the quality of the data by the way they are presented (Treguear & Carro, 1994), and the exploration of cultural understandings of child abuse in Zimbabwe, carried out by a team of researchers working with participatory methods (Loewenson & Chikamba, 1994). But these are exceptions. In general terms, the available global discourse on this theme is characterised by a poor understanding and use of quantitative information, lack of attention to research techniques, the reproduction of myths and unsubstantiated facts, as well as the use of assumptions and campaigning imperatives in place of established bodies of theory. If, as is frequently stated, children indeed deserve the best we (adults) have to give, they are not receiving their just deserts in a field in which they are particularly vulnerable. Or, as has been said with respect to the discourse on street children and its effects, it is not acceptable

that international organisations, policy makers, social institutions and individuals who feel entitled to intervene in the lives of children with problems, do so on the basis of obviously unclear and arbitrary knowledge about the reality of these children's lives. (Glauser, 1990, p. 144).

Numbers

Perhaps the most serious of these aspects is the way in which numerical data are manipulated and reproduced. International interest in children gained momentum in the United National International Year of the Child (1979) and was given further impetus through the adoption and entering into force of the United Nations Convention on the Rights of the Child (1989/90) yet, in the space of nearly two decades, little has changed in terms of the way research is carried out and used by child welfare and advocacy organisations, despite considerable advances in theories of childhood and methods of researching children's issues within the academic community. What this amounts to is that the numbers provided for all groups of children in need of special care and protection have tended to remain the same, based on guestimates rather than research. Whereas guestimates have their place in the early stages of research, provided that they are based on sound reasoning, the role they should play is that of baseline hypotheses, to be proved or disproved so that the true scale of a problem can be understood and children protected using programmes grounded in a real understanding of their situation. In the case of child sexual exploitation, however, guestimates have become fact, partly because they have become inscribed in rhetorical discourses aimed to raise awareness. The objective appears to be to heighten public and policy interest in the issue by stressing the scale of the problem. Yet this is neither ethically acceptable nor logical. In the first place, as stated as far back as 1983 by the United Nations Special Rapporteur on the Suppression of the Traffic in Persons and the Exploitation of the Prostitution of Others, 'The important point is not the scale of the problem but its degree of seriousness as a violation of the fundamental rights of the human person' (Fernand-Laurent, 1983, p. 14). In the second place, the normal practice within the literature of providing a raw number, such as 'There are 1 million sexually abused children in Asia' (Narvesen, 1989), does not actually provide an idea of scale. To do this would require some knowledge of source, time, relative location and proportion. A correct statement of scale would be something like 'According to estimates made by A on the basis of B type of calculation, in 1996 there are 1 million sexually abused children (under 18 years of age) in Asia. This is C% more than the calculation made on the same basis in 1994, D% more than estimated by the same method in Africa in the same year, and represents E% of the total population of Asian children in this age group.' Scale can only be understood in these terms. Moreover it is a poor excuse for adult society to claim that, after nearly two decades and with information technology that can provide minute details of weather, money markets and economic cycles for every country in the world, to claim that guestimates are the only data available. Some relevant data do exist and more could be collected. We have the technology. But agencies 'who feel entitled to intervene in the lives of children' fail to use it.

The current situation on the quantification of the commercial sexual exploitation of children is characterised not only by widely differing figures but also by very different definitions. Data on children are frequently hidden within and confused with data on adults. One problem is caused by the different age groupings used. In many studies the juvenile population is defined as under 21, rather than 18, years (see for example Lowman, 1987). Combined with the vague bases for guestimation, this makes it difficult to judge the scale of commercial sexual exploitation of children even within a single national context. In the case of the USA, for example, one study provides an extraordinarily broad estimate of between 100,000 and 300,000 of young prostitutes (cited elsewhere as 'child prostitutes') on the basis of a sample of people aged 14 to 34, with a mean age of 22 years (Snell, 1995). Another study, of the USA, suggested that there are 2,400,000 child prostitutes and estimated a total of 100,000,000 sexually abused children in the world, but did not distinguish adequately between the categories of child prostitute and sexually abused children (Joseph, 1995). Gibsonainyette and colleagues estimated that the number of prostitutes under the age of 18 in the USA in the 1980s and also stated that there had been a 242% increase in underage prostitution between 1967 and 1976 (Gibsonainyette et al, 1988). Campagna and Poffenberger claimed in the same year that 1.2 million children were being sexually exploited in the USA but did not define sexual exploitation (Campagna & Poffenberger, 1988 ). Combining these figures results in a highly confusing picture (Table 1; see also Box 2).

Table 1: (Gu)Estimates of child abuse and prostitution in the USA, various sources

Year	Estimate	Definition	Age	Source
1967	247933	child prostitutes	under 18?	Gibsonainyette et al, 1988
1976	600000	child prostitutes	under 18	Gibsonainyette et al, 1988
1977	600000	boy and girl prostitutes	?	Densen-Gerber (Lloyd)
1988?	1200000	abused children	?	Campagna & Poffenberger
1995?	100,000 to 300,000	young prostitutes	14-34?	Snell
1995?	2400000	child prostitutes	?	Joseph 1995

Box 2: Creation of a statistic ...even the best estimates of the numbers of juvenile prostitutes may have poor statistical foundation. Nevertheless, because they are the only figures available, they enter official records and become facts which may be quoted confidently by anyone. One example of this is the figure of one million child prostitutes in the United States, which was given in evidence to the House of Representatives in 1977. The expert in question was Dr Judianne Densen Gerber, Director of the Odyssey Institute which operates rehabilitation programmes for children with various kinds of deviancy problem. The figure was a guestimate based on the number of 300,000 boys prostitutes given in what Dr Densen-Gerber refers to as 'the research of Robin Lloyd'. Because she assumed that there would indubitably be more girl prostitutes, she added 600,000 to this figure without providing any evidence to support her claim. Indeed, historical evidence about man-boy preferences, and some figures given by other authorities in other countries, might contradict her assumption. But the real problem is that Lloyd is not a social scientist working on any established methodology for gathering statistical information, but a journalist researching a book for the popular market. Here is his own account of how he arrived at the figure Dr Densen-Gerber quoted: In the early stages of research for this book, I approached police officers and leaders of the gay community with a working figure of 300,000 boy prostitutes in the United States alone. Deputy District Attorney James Grodin, in Los Angeles said, 'You wont get any argument from this officer for that figure'. During a television interview I offered the same figure to Morris Kight, the West Coast gay activist. Said Kight, 'It might well be double that amount'. But what Kight and Grodin were agreeing to was -- at its best -- a gut hunch. (Lloyd 1979, p. 202). None of these experts ever consider the alternative premise that the figure 'might well be' considerably less. Source: Ennew, 1986, pp. 4-5

Similar difficulties occur in other parts of the world, both the figures themselves and and the range of the guestimate tend to be large (Table 2). As many figures for child

Table 2: (Gu)Estimates of child prostitution in the Philippines (various sources)

Date of publication	Number	Source
not given	3,000 to 20,000 in Manila	Gearin
1993	100000	ECPAT
n/d	20000	Salenlahi
1989	30000	Moorehead
1995	60000	Sachs

prostitution are based on either female or male populations, depending on the discourse involved, and the gender of the 'category' is not always made explicit, it would be difficult to compare information from different countries. Thus, given that the emphasis in India is on the 'girl-child' and in Sri Lanka on boy prostitutes, juxtaposition of figures from these two countries would be fairly meaningless.

One characteristic of the manipulation of numbers is that they are always large and always said to be 'increasing', despite the fact that the same numbers are repeated year on year, a manipulative technique that would not work with company accounts. Sources are rarely provided, the credibility of an organisation being sufficient verification (Figure 1).

Figure 1 This map appears in a widely distributed report on the activities of ECPAT and has been reproduced in various publications. The dates of the estimates are not provided, nor are the definitions and age groupings standardised. The sole source of information cited is the Government of Vietnam, presumably because this provides extra credibility. Source: The international campaign to en child prostitution in Asian tourism, special edition of ECPAT Newsletter, NO. 9, July 1993.

Even greater difficulties become apparent in attempts to disaggregate data on child prostitutes by ethnicity. Most studies give percentages of whites, Asians, Afro-Caribbeans and so forth but with no commentary and little additional information. The reader is not told what is meant by the category 'White'. Does this include Hispanics or not? Are Chinese and other East Asians placed in a different category to South Asians or are they all simply classed as Asian? If the reader is told that, for example, 60% of a sample of prostitutes are 'Non-white' this might imply (especially in a journalistic account) that Caucasians are under-represented. Yet, if the 60% were disaggregated one might discover that 15% were Afro-Caribbeans, 15% South East Asian, 15% Latino and 15% South Asian, which would mean that Caucasians were over represented in comparison with other designated 'racial' groupings. Blurred categories provide perfect fodder for stigmatisation, particularly if the discourse involved focuses on ethnicity without taking into account other markers of social status and bases of exploitation, such as age, gender and socio-economic class (Ennew, 1986, pp. 9-10).

Reproduction and repetition of myths

Numbers are an important part of any campaign because they endow it with urgency and a overwhelming sense of importance. Campaigns, journalism and academic literature tend to use the same unreliable statistics, either without reference to the source or citing one of the texts that have attained credibility. A case in point is Narvesen, The Sexual Exploitation of Children in Developing Countries (1989), which claims that there are one million sexually exploited children in Asia, although other texts take this to mean one million prostitutes. It is normal for the highest figures to be given. They are always 'increasing', usually at an 'alarming' rate, reaching 'epidemic' proportions. Yet no reason is given for this except for the 'fact' that customers are turning to ever younger prostitutes because they believe them to be AIDS free. There is no evidence either way for this claim, but it has been endlessly repeated (Jubilee Action Trust, n.d.; Lee-Wright, 1990; Muntabhorn, 1992; O'Grady, 1992; 1994) until it has become a 'fact'. This reproduction of inauthentic or innacurate information is a characteristic of the literature on the commercial sexual exploitation of children. Even before the AIDS pandemic, the literature on child prostitution from all parts of the world repeated, without evidence, that child prostitutes were sought after because they were believed to be free of sexually transmitted diseases (STDs), including the apparently worldwide belief that penetrative sex with a virgin is a cure. Other commonly repeated myths include female customers injecting boys' penises (sometimes with an unspecified substance, sometimes with a variety of named but unlikely liquids) in order to produce or maintain erections. The evidence for these myths is may simply not be provided, or given as a citation to another text that itself provides no evidence, or take the form of hearsay evidence, the unverified statement of someone who knows somebody who knows these things are true.

Many of these myths are employed in order to demonise the clients of child prostitutes, as if the violation of children's rights involved were not sufficient condemnation. At other times myths are used as distancing devices as part of the construction of the client as outside the society in question, a foreigner, or a tourist. Unresolved question raised by this literature review include how to explain the conditions of production of these myths. What explains the prevalence of particular myths and figures and the dominance of oral over written discourse in this area? What are the 'politics of hype' that result in repetitive, shocking (and titillating) information in this field?

Tourism, xenophobia, construction of the other

Just as ECPAT dominates the overall discourse, so the single idea of sex tourism, with which ECPAT is associated, captures media and other attention (Lorayes n.d., Salinlahi et al, n.d; Asia Partnership for Human Development, 1985; 1992; Miralao et al, 1990; Lee, 1991; Hall, 1992; Anglican Synod of Australia, 1993; O'Grady, 1992; 1994). ECPAT began as a sub division of ECTWT (Ecumenical Coalition on Third World Tourism) apparently perceiving child prostitution as an inevitable consequence of tourism (ECTWT, 1983; 1990). One of the leading anti-tourism campaigners once wrote "too much tourism is the rape of culture, the environment, women and children" (Srisang et al, n.d). The success of the campaigns now apparent in the growing interest in sex tourism in other parts of the world, but is particularly acute in Asia, where it often takes on an expressly anti-Japanese tone. ECPAT's original statement of intent makes this clear:

The conduct of the tourists destroys all attempts to heal the wounds incurred during the Second World War. We would prefer to forget Japanese military imperialism, but, instead of the uniform, the Japanese come today in suits and violate the dignity of the people of Asia with a particular malicious form of socio-economic imperialism. (ECTWT, 1983, p. 14)

Particularly in the Philippines, memories of the Second World War run deep, Japanese are still disliked and much of the blame of child sex tourism is laid at their door. One study of child prostitution in the Philippines has an entire section on Japanese involvement in the trade (ECPAT Philippines, 1994).

Although the language of women's rights and children's rights is sometimes used, child prostitutes are seen as a symptom of the wider problem of foreign (often Western) influence, as in the use of the term 'social evils' in countries such as Vietnam. Although legislation to combat social evils came into force in Vietnam in February 1996, even the Ministry of Social Affairs, which is responsible for implementation, does not have an official definition of the term, which is used as a blanket phrase to mean all that is contrary and harmful to Vietnamese culture, yet associated with gambling, theft and prostitution, none of which are exclusive to non-Vietnamese. As in other 'sex tourism' countries, this mechanism parallels the distancing mechanism in the West that places the emphasis on 'stranger danger', so that sexual abuse within the family is played down in favour of fear of child rape by unknown, asocial men. As Jean La Fontaine states in her seminal study of child sexual abuse in the United Kingdom,

If people do think of the possibility of sexual assaults on children they see them as a risk from casual encounters in the street. The belief in the natural relationship between parent and child is the basis for the firm conviction that if danger threatens any child, it can only come from 'outside' the family. When the damage inflicted is sexual, the offending person must be an unknown, a shadowy and frightening stranger, not anyone with whom one has daily contact, let alone someone who is part of the familiar circle of family and friends....Newspapers report the rarer and more dramatic cases of children who disappear and in doing so reinforce the general idea that strangers carry a risk to children involving acts of the greatest perversion, serious damage and even death (La Fontaine, 1990, p. 109).

Figure 2 Demonisation of the figure of the foreign, male tourist. Distinguished by his camera and with his eyes hidden behind dark glasses, the tourist confronts children on a wasteland far from the high rise buildings of the modern city. The bicycle is presumably a bribe. One campaign slogan in the Philippines was 'He may look like a friend but he could abuse your child.' Source: The international campaign to end child prostitution in Asian tourism, special edition of ECPAT Newsletter, NO. 9, July 1993.

Campaigners discuss child prostitution using the language of the market place and talk about it as if it is a question of supply and demand (Good Shepherd Sisters 1994). Foreign men want children therefore they are supplied (O'Grady 1992; 1994; O'Connell Davidson & Brace, 1996). If the demand could be stopped, then there would be no prostitution, an idea that unhelpfully confuses moral and economic discourses (see for example ECTWT, 1990).

Thailand and the Philippines do have some of the best recorded cases of actual commercial sexual abuse by foreigners. Despite the sensationalism of some of the reports, there has been considerable documentation of individual cases. The Jubilee Campaign for example, published full trial transcripts of the trial of a paedophile accused of the rape and grievous bodily harm of a 12 year old girl who later died, giving a full account of the injuries she had suffered (Jubilee Campaign 1992). In this case the transcripts were presented without commentary and this restraint makes the account more shocking than might be the case with the use of emotive writing, which is more usually the case (Figure 3).

Figure 3 Newspaper headlines, collected by ECPAT and typical of the genre, show the emotive language commonly used to report child sex tourism. Source: The international campaign to end child prostitution in Asian tourism, special edition of ECPAT Newsletter, NO. 9, July 1993.

Other writers have described the situation of young prostitutes within the overall social and economic contexts of their lives, which contextualises the life choices and survival strategies that children use (see for example Black, 1991; Black, 1995)

In contrast to this, there is some first hand information from Cambodia that examines child prostitution in terms of local clients and a video made by the Cambodian's Women's Association video examines the role of local men. While acknowledging that certain foreigners, in particular members of the United Nations Transitional Force in Cambodia, were clients, the organisation sees this primarily as an issue of exploitation of the girls involved and is less interested in the nationality of the abusers. In Cambodia, the role of Vietnamese child prostitutes is given special attention as Vietnamese prostitutes are reported to make up over 50% of the prostitute population. Working illegally and not speaking the local language makes these children vulnerable to abuse (Care International, 1994; Cambodian Women's Development Association 1994; Cambodian Women's Association 1994; Thang, 1996).

The concentration on child prostitutes who service foreign clients raises the concern that other child prostitutes catering for local customers are being neither counted nor provided for within current policies and programmes. The concern with sex tourism may well be obscuring a large part of the child prostitute population, just as focus on street children forces attention away from the far larger numbers of deprived and marginalised children living and working in urban and rural areas. By presenting information in the way they do, many organisations committed to the eradication of the commercial sexual exploitation of children make it appear that child prostitution to has a single root cause and a single solution -- foreigners, the external enemy. This means that questions about national structural causes of poverty and marginalisation need not be asked. The approach taken by Treguear and Carro in their study of girl prostitutes in Costa Rica is thus all the more refreshing (Treguear & Carro, 1994). Even though Costa Rica is a major tourist destination and gender relations are implicated, they lay the blame for child prostitution on structured, economic and political violence against the poor, who are defined not simply in terms of lack of access to employment, goods and services, but also more specifically through their lack of access to power. These authors end their theoretical analysis by quoting Schibotto (1990, p. 163): 'each one of these girls is a reflection of the violence, not only of her lovers or her clients, but also the entire social formation. Because, in the last analysis, everyone of us -- at some time or in some way -- has gone to bed with them' (our translation).

How evidence is gathered

Perhaps the most disquieting aspect of the literature is the generally poor quality of research. The overwhelming majority of publications and 'grey literature' in the field of the commercial sexual exploitation of children is characterised by muddled, low level or misunderstood theories, badly thought out and applied research methods, poor data and inadequate analysis. Most of the literature consulted for this review was so poor that it was not worth including in the annotated bibliography and, indeed, inclusion should not be taken as a sign of high, or even adequate, quality material. It seems that, in this field, the burden of proof about the truth of a statement does not rest with the research witness who is thus enabled to make a priori claims about the existence and extent of facts that people in general find so desperately uncomfortable that they would rather accept the incredible than ask questions.

It is worth listing the more common errors of research method and analysis, because they contribute to the reproduction of unreliable or mythological information within the literature:

€Much so-called research is carried out by lawyers or activists with no background in the social sciences, whose activities are best described as 'fact-finding' and who accept as fact what would be thrown out of court as hearsay evidence;

€Data are probably biased when (as frequently happens) researchers gain access to research subjects by means of institutions, projects and programmes. Information may thus reflect what the children and others think the project would like them to say, fear of reprecussions from institutional staff, or exaggeration in order to attract greater project advantages.

€Samples are also likely to be skewed because they are drawn from what might be called the 'unsuccessful' prostitutes whose activities have attracted the attention of helping or controlling agencies.

€Researchers seldom use control groups when designing their research.

€Samples tend to be small, yet the information is frequently stretched extremely thin by being subjected to inappropriate quantitative analysis.

€Results of research with small-scale samples are generalised to represent large populations.

€Research is often divorced from local and cultural contexts. Little or no detail about research subjects is given.

€Researchers rely on single-method studies, often on anecdotes that are passed off as case studies. Information gathered is seldom cross-checked (sometimes called 'triangulation') by using other methods, or by comparison with other studies and secondary data.

€Far too often the only social science method employed is the questionnaire survey, which is at best a poor method when used alone, at worst a bad tool to use with children, particularly where sensitive subjects such as sexuality and abuse are concerned.

The language of evidence

One factor in the reproduction of inadequate information in the literature of commercial sexual exploitation of children is the style of language used. Language is the means by which ideas are reproduced. The way it is used in any one context reflects the structure of the paradigm or overall theory that has given rise to these ideas. Language is never neutral. If a structure of ideas is repetitive, so the words and phrases that it gives rise to will bear the same repetitive characteristics. Literature on the commercial sexual exploitation of children is characterised by assertiveness. Uncertain verb forms such as the conditional or the subjunctive mood are seldom used. Thus writers seldom say that something 'might be the case' or 'it is reported that...'., preferring to present their case in the positive indicative, 'is', 'was, 'were'. Thus sentences often begin 'There is evidence...', or 'It is claimed..' or 'It has been found...' although the nature and source of such evidence, findings and claims are not presented. A further immediacy is given to texts by the use of what is called 'the ethnographic present' which is the device of writing about past events in the present tense. A pastiche might be: 'Caroline stands outside the nightclub waiting for a client. She is shivering with the cold and hopes that the next man will not be violent like the man who blacked her eye yesterday. She is twelve years old, but she has the eyes of a much older woman.' This often obscures the fact that the case study was gathered long ago by someone else and that Caroline, if she ever existed, for it is common for writers to invent a composite or typical person, probably is a much older woman by the time the reader encounters her in the text.

Further linguistic devices are implicated in the repetitions and confusions of quantitative data. For example, in the chapter that is devoted to child exploitation in a highly respectable publication on children's rights, the author writes:

Undoubtedly most international trafficking is in women over the age of 16, but some children get swept up in the tide. Specific ages are rarely reported, but it seems reasonable to infer that at least some under-16s are included (Kent, 1992, p. 325).

'children as young as...' which usually refers to one exceptionally young child in a larger sample of children considerably older;

'up to 20 times a night'.. when referring to numbers of clients, which gives no idea of the average number of clients or the type of sex work under discussion.

Nevertheless, journalists do tend to quote children directly more often than academics although, when children are allowed a voice, it seems they are muted or speak according to predetermined scripts. Adult advocates are usually given much more column space, and their opinions are given far greater weigh,t so that it can sometimes seem as if journalists are merely quoting a child to prove the point that adults are making.

Within the non-governmental sector in general, data tend to be collected and/or collated in an extremely patchy and haphazard manner. There is a great reliance on newspaper reports, individual stories, other NGOs and social workers, child rescuers and other 'experts' who guess at numbers and statistics. There is almost no information from the children who work as prostitutes and little exchange of information with the academic sphere. What counts as evidence can be seen in the following passage from The Child and the Tourist, in which the author quotes 'an anonymous' Thai friend's description of 'paedophiles' he saw at a beach in Pattaya, Thailand. The 'friend's words apparently were:

"I sensed an insincere, almost sinister smile on the faces of the men. Many were giggling like school children and talking in the uninhibited manner of young children. In the water, they wore skimpy swimming gear while the boys were fully dressed. Frolicking in the water they took their shorts off and bobbed up and down and removed the boys clothing. On the beach, they played with the boys in mock fighting showing their superior strength. their communication with the boys was entirely physical (much of it with sexual references and some of it genital). They had total lack of awareness of Thai customs" (O'Grady 1992: 98).

The overall impression gained from carrying out this review of the literature on the commercial sexual exploitation of children is a permanent sense of dejá vu, because the material is so repetitive and the methods of data collection, analysis and presentation reinforce the way information tends to be reproduced. There are interesting debates and pockets of verifiable data, but these are obscured by the overwhelming weight of sensationalism, pressurised advocacy and refusal to examine taken-for-granted assumptions. The main reason is that data in this field generally arise in the context of campaigns so that knowledge is organised around adult requirements for particular kinds of fact, rather than the actual lives and needs of sexually exploited children.

1.3 In search of a framework

Facts simply do not lie around on the ground to be picked up as pigeons pick up peas. An organisational framework of ideas is necessary in order to measure and monitor any social phenomenon. Within the field of the commercial sexual exploitation of children several clusters of ideas and theories might be considered for their usefulness as organising principles for the development of concepts and the collection of data. Each is related to a particular campaigning stance, and we have chosen to present them here in schematic form (Table 3).

Table 3: Campaign theories and related theories in the field of the commercial sexual exploitation of children

Campaign	Theory
Feminism	Patriarchy, the 'girl-child'
Morality	Religion, sexuality, blaming perpetrators, rescuing children
Child survival and development	Psychology, medicine
International development aid	Poverty, demand and supply (economics), community development
Children's rights (including sexual rights)	Power, childhood, human rights

Although all these theoretical structures have their merits and explanatory value, the most coherent and (more importantly) children-focused framework for the purpose of measuring and monitoring the commercial sexual exploitation of children seems to us to be children's rights. This is because consideration of children's rights entails a discussion of the nature of childhood, which is inscribed in the power differential between adults and children. Discussion of this unequal relationship opens the possibility of discussing other inequalities that exist universally, while taking different cultural and historical forms. This would mean that a framework for monitoring based on these ideas would be both stable and flexible. The relationship between children and adults within families, has a parallel in the relationship between children and states, which, in their modern forms, are ultimately responsible for policing parenting, schooling and work, the main socialising institutions of childhood. Likewise, the trio of child/family/state exists within structures of regional and global domination, which include economic disparities, political inequalities and, last but not least, tourism.