Chasing spirits: Clarifying the spirit child phenomenon and infanticide in Northern Ghana

https://doi.org/10.1016/j.socscimed.2010.04.022Get rights and content

Abstract

In the Kassena-Nankana District of Ghana, researchers and health interventionists describe a phenomenon wherein some children are subject to infanticide because they are regarded as spirit children sent “from the bush” to cause misfortune and destroy the family. This phenomenon remains largely misunderstood and misrepresented. Based upon both ethnographic research and verbal autopsy data from 2006 to 2007 and 2009, this paper clarifies the characteristics of and circumstances surrounding the spirit child phenomenon, the role it plays within community understandings of childhood illness and mortality, and the variations present within the discourse and practice. The spirit child is a complex explanatory model closely connected to the Nankani sociocultural world and understandings surrounding causes of illness, disability, and misfortune, and is best understood within the context of the larger economic, social, and health concerns within the region. The identification of a child as a spirit child does not necessarily indicate that the child was a victim of infanticide. The spirit child best describes why a child died, rather than how the death occurred. In addition to shaping maternal and child health interventions, these findings have implications for verbal autopsy assessments and the accuracy of demographic data concerning the causes of child mortality.

Introduction

What is a spirit child? It is a child that has a large head, is born with teeth or a beard, spies on its parents, and vanishes when the parents are not looking. Sometimes when you give birth, you don’t know you have given birth to it. A woman who gives birth, continuously falls sick, and doesn’t get well has given birth to a spirit child (Elder Nankani woman, 2007).

Although preventable diseases and, ultimately, the effects of poverty constitute the primary causes of infant and child mortality throughout the Kassena-Nankana District (KND) in Northern Ghana, local discourse suggests that a number of infant and child deaths are intentionally facilitated by family members. In these cases, deformed or ailing children, births concurrent with tragic events, or children displaying unusual abilities are regarded as spirit children sent “from the bush” to cause misfortune and destroy the family. From the Nankani perspective, spirit children are not human, but are bush spirits masquerading as such. From a biological perspective, many of these children have disabilities or are chronically ill.

The spirit child remains the subject of considerable speculation and is frequently misunderstood. Community members, population health researchers, and development workers have described the spirit child practice as infanticide, a cause of family misfortune, a way to cope with unwanted or sick children, and as a primitive practice grounded in ignorance needing complete eradication. Although there has been a successful history of child and maternal health research and intervention in the KND, few have studied the spirit child directly or have made the phenomenon the primary focus of their research; rather, it has remained a side interest within the context of larger studies on child and maternal health.

The few reports enumerating the spirit children are inconsistent. The Population Council estimated that four percent of child deaths in the KND are due to poisoning of spirit children (Lothian, 1996). Allotey and Reidpath, however, estimated that 15% of deaths of infants under three months in the KND could be due to the spirit child practice (2001). They concluded that the spirit child was a “significant health problem” and an “important risk factor for infant mortality within the district” (Allotey & Reidpath, 2001, p. 1006, 1010). Descriptive data from the Navrongo Demographic Surveillance System (NDSS) and verbal autopsies from 1995 to 2002 were used to estimate that 4.9% (n = 53) of neonatal deaths (N = 1068) were due to infanticide (Baiden et al., 2006). Baiden et al. remarked that in addition to preventing and treating neonatal infections and having a skilled attendant at the delivery, a further reduction in neonatal mortality is achievable with the elimination of infanticide (2006, p. 532).

Regardless of their accuracy, these estimates do not take into account the differences between how people talk about the spirit child, how it is reported in verbal autopsies, and what is happening in practice. Allotey and Reidpath remarked that, “works of fiction, anecdotes, and fireside stories remain the strongest indication of the existence of the spirit child” (2001, p. 1008). Indeed, the limited research into the spirit child phenomenon has been more of an exercise in chasing elusive spirits embedded within community discourse, since few scholars have offered significant evidence that it actually occurs outside of discourse. Understandably, researchers have not had opportunities to carefully examine the demographic data, to attend to the sociocultural and ethnographic context, and to gather first-hand accounts necessary for accurate description and analysis of the phenomenon.

Baiden et al. recommended anthropological studies to guide interventions and further validation studies to “ascertain the sensitivity and specificity of the use of verbal autopsy to diagnose infanticide” (2006, p. 536). The purpose of this paper is to begin filling in the “important gap” in information available on the spirit child phenomenon; its characteristics, the extent of the practice, and the role that it plays in child mortality (Allotey & Reidpath, 2001, p. 1008). Specifically, we aim to provide an in-depth description of the spirit child phenomenon from the stages of suspicion and diagnosis, to the varied circumstances surrounding death. We establish that infanticide is not always present in spirit child cases, that there is significant diversity in how it is used as an explanatory model for child mortality, and that the practice itself is not necessarily an “important risk factor” or a major community health issue as claimed, particularly when compared to mortality rates from malaria and other diseases.

Section snippets

Methods

A mixed method approach combining ethnographic fieldwork and analysis of demographic data provided valuable and more holistic insights into infanticide and the spirit child phenomenon. Ethnographic fieldwork was conducted between 2006 and 2007 and again in 2009 as part of the dissertation research conducted by AD in collaboration with the Navrongo Health Research Centre (NHRC) and AfriKids NGO. In addition to participant-observation, unrecorded interviews, and informal interactions, over 100

Context

This study focused on Nankani families living in the Eastern Sub-District of the Kassena-Nankana District in the Upper East Region of Ghana. The KND is a semi-arid, sub-Sahelian guinea savannah with one annual rainy season. As part of the Volta Basin, its topography and cultural characteristics are more akin to those living to the north in Burkina Faso, for example, rather than in the rainforests to the south (Mensch, Bagah, Clark, & Binka, 1998).

The primary occupation in the district is

Perspectives on infanticide

Infanticide is widely distributed culturally, historically, and geographically; “rather than being the exception, it has been the rule” (Williamson, 1978, p. 61). An inclusive definition of infanticide includes “any form of lethal curtailment of parental investment in offspring” (Hrdy & Hausfater, 2008, p. xv). In a literature review, Hill and Ball identified that the most common reasons for infanticide included low likelihood for survival, inadequate resources, twins, gender preference, or

What is the spirit child?

The spirit child was just in the olden days of the ancestors; if the child grows sick or lean then they say that such a child is a spirit child, and they will send the child away. But now, if a child is sick, they can send for treatment. There is no spirit child (Middle-aged mother, 2006).

The spirit child is like this example: My mother gave birth to a child. Just by looking at this child nobody believed that it could survive. Then some strange things started happening in the house. All the

Detection and diagnosis

The literature describing the spirit child often portrays the detection and decision-making process that surrounds suspected cases as urgent and unilineal, quickly proceeding from suspicion to poisoning and abandonment. Awedoba noted that “in the past” such neonates when detected were disposed of as soon as possible, often within 24 h of birth, as it was in the best interest of the family (Awedoba, 2000). Allotey and Reidpath also described a case where an infant was allegedly given an herbal

Determining the cause of spirit child mortality

The VA data, subsequent demographic results, and the available literature attribute all spirit child deaths to infanticide. Upon VA data analysis, however, it was determined that there is considerable variation in the use of the spirit child phenomenon as an explanation for a child’s condition or family circumstances, and how children are or aren’t treated. Not all spirit child cases are homogeneous and result in intentional death causing behaviors. In fact, when a family describes a child’s

Treating spirit children: concoctions and concoction men

While a portion of the spirit child deaths are not due to infanticide, a majority of the cases offer direct evidence of infanticide. When confronted with a spirit child, a family may attempt to appeal to the ancestors to take the spirit child away, but in most cases the ancestors are not powerful enough to facilitate its death. After a period of suspicion and help-seeking, a concoction man is sought by the family to treat the spirit child with the dongo. The concoction man is the ritual

Clarifying the spirit child

We have posited that a portion (36%) of spirit child deaths are not due to infanticide, since some diagnoses are post-mortem or the child dies naturally. We have also examined the preparation and administration of the concoction, showing that in some traditions a limited amount or no poison is given, bringing into question if infanticide actually occurred in some cases. This final section examines the VA and ethnographic data to further describe the circumstances surrounding the intentional

Acknowledgements

This research would not have been possible without strong collaborative relationships between the University of Alberta, Northern Arizona University, the Navrongo Health Research Centre, and AfriKids, a child rights NGO working to address the root causes of spirit child phenomenon. Special thanks to James Phillips, the Population Council, and the community members of the Kassena-Nankana District.

References (27)

  • K. Edmond et al.

    Diagnostic accuracy of verbal autopsies in ascertaining the causes of stillbirths and neonatal deaths in rural Ghana

    Paediatric and Perinatal Epidemiology

    (2008)
  • J. Einarsdottir

    Tired of weeping: Mother love, child death, and poverty in Guinea-Bissau

    (2004)
  • A. Gottlieb

    Under the kapok tree: Identity and difference in Beng thought

    (1992)
  • Cited by (42)

    • Filicide as a cultural practice in Ghana: The qualitative understanding of a family tragedy and its implications for child protection practice

      2022, Child Abuse and Neglect
      Citation Excerpt :

      Yet, these concepts provide a helpful way to understand the phenomenon and identify different but valid avenues for intervention. Studies by Allotey and Reidpath (2001) and Denham et al. (2010) showed that within the Ghanaian traditional culture, the cause for the birth of the so-called SC is placed on the mother. A number of conscious or unconscious activities engaged in by the mother was known to be the cause of the birth of a spirt child.

    • Global Neonatal Health

      2018, Avery's Diseases of the Newborn: Tenth Edition
    • Global Neonatal Health

      2017, Avery's Diseases of the Newborn, Tenth Edition
    View all citing articles on Scopus
    View full text