The Demography of a New Guinea Highlands Valley

Dados Bibliográficos

ID12888018
AutoresPaul Sillitoe (Durham University, autor correspondente)
Ano2006
Volume2
Fascículo3
Páginas271-294
Data de publicação2006-11-01
Peer ReviewedSim
Open AccessNão
TipoARTICLE
PeriódicoAsian Population Studies (JOURNAL)
Identificadores do periódicoISSN: 1744-1730 • E-ISSN: 1744-1749
EditoraTaylor & Francis (PUBLISHER • GB)
DOI10.1080/17441730601073862
OpenAlexW2070313808
IdiomaEN
Citações recebidas1
Referências citadas32

Relatively little is known about demographic processes in the Highlands of Papua New Guinea, although since the reports of the first ethnographers, population growth has been thought a significant issue. This paper analyses a series of population surveys conducted over 24 years in a Wola-speaking community of the Was valley of the Southern Highlands Province. It seeks to put a standard demographic investigation of births, deaths and migration in cultural and historical context. Polygynous marriage arrangements, male fears of menstrual pollution and attitudes to extra-marital affairs impact on birth rates. The periodic occurrence of El Niño induced food shortages and ‘tribal fighting’ affect death rates. Migration patterns are influenced both by traditional arrangements, such as families living on two territories, and recent changes, such as towns that offer new employment opportunities. Although the population is expanding, there is currently no problem with resources. Keywords: Papua New Guineamicro-demographycultural factors I thank Bryant Allen, Robin Hide and Kate Hampshire, and the three Asian Population Studies reviewers, for their generous help and advice with this paper. Notes 1. This is common in anthropological studies, contributing to a heavily criticized synchronic focus and inference of population stability (Zubrow Citation1976, p. 12). 2. The homestead holder is the senior resident person who people name when asked whose household it is, usually a man, often a father. 3. Burton (1988, p. 37) notes that there is a ‘deal of variability in the number of children born from one five year period to another’ for a small Wahgi community. 4. John Burton has furthered these ideas by developing a computer-based method of collecting and managing genealogical data, with sub-routines for estimating ages, birth rates, death rates etc., called Community Express (http://www.communityexpress.info/). 5. According to the Tari surveillance data, the male to female ratio at birth is 104:100. Intriguingly, van de Kaa (Citation1971) reports more males to females for the total population of the Papua New Guinea in the 1960s. 6. From a ratio of 89 males to 100 females between 1973 and 1983, to one of 76 males to 100 females in 1993 and 1997. 7. John Gordon-Kirkby, one time Assistant District Officer at Nipa station, kindly supplied me with a list of such significant events that included dates, used then by the administration. 8. In addition to regular return visits, I have kept up correspondence with friends in the community informing me of news including births and deaths. 9. The pyramids have five-year intervals and they plot age range as, for example, five to nine years nominally covers all those between five years, which and one day short of 10 years—although such accuracy was not possible with many persons, as birth dates are unknown. 10. Burton (1988, p. 27) reports similar anomalies in Wahgi valley census data, an eight-year-old recorded as a mother and another woman as having her first child at 46 years, and he proposes a genealogical reconstruction method to correct erroneous age estimations. 11. The Chimbu birth rate data ranged from 24 to 32 births and those for fertility from 116 to 145 (Brown & Winefield 1965, p. 181); the Baiyer valley birth rate was 27 births and fertility 116 per 1000 (Becroft et al. 1969, p. 50). 12. This is the average age at which a sample of 50 women >50 years gave birth to their last child. On the basis of this I take 40 years as the average age at which women cease child bearing in the following discussion (women evidence a noticeable decline in fertility when they are >30 years old). 13. This is not to deny the demographic argument that polygyny increases overall fertility rates by ensuring that all women of childbearing age are in sexual unions, outweighing on a population level the fact that individual women may have fewer children each (e.g. evidence from Africa shows that populations practising monogamy have lower overall fertility rates than those practising polygyny: Lesthaeghe Citation1989; Randall Citation1996). 14. This graph and accompanying calculations, and the following ones, include only women past childbearing age, which as noted previously, these demographic data suggest is on average 40 years of age. They also omit women married twice or more, one of whose marriages was monogamous and the other polygynous, as not possible to disentangle monogamous and polygynous effects on their reproductive histories. 15. These numbers of births indicate a mean birth interval of 4.1 years, which compares with Lehmann et al. (1993, p. 96) who report a mean birth interval of 3 years (or 1.8 years if infant dies) for the Tari basin, Becroft (Citation1967) who reports a mean interval of 5 years (or 1.5 years if infant dies) for the Baiyer valley, and Wood (1992, p. 104) with an interval of 3.7 years in Takwi valley. 16. Wood (1992, pp. 105–110) presents persuasive evidence to support his argument that it is length of lactation among the Gainj—in excess of three years—that primarily accounts for their low fertility as the ‘most powerful birth-spacing mechanism’. Nursing is a significant contraceptive with high frequency suckling (he reports 24-hour suckling rates of 160 minutes in newborns declining to 80 minutes in three-year-olds) and it decline slowly as children age. Women experience comparatively few regular menstrual cycles during their reproductive lives—they are either pregnant or lactating and so acyclic much of the time (only 13 per cent of married women 20–39 years of age were not lactating or pregnant). Another factor affecting fertility is that women's nutritional status declines cumulatively with each reproductive episode—he attributes this in part to effects of parity, notably the energetic demands of pregnancy and lactation (p. 111). 17. Adoption does not confound this assumption. It is not widespread in the Was valley, unlike some regions such as Goroka (Collins 1988, p. 186). Adoption occurs largely with orphans, close relatives usually taking over their care. The Was valley surveys noted such cases, and also adolescents who may be mobile and reside with kin other than their parents; this analysis associates them with their social-biological parents and not adoptive kin. 18. It is probable that these data omit several infant deaths but such individuals are also omitted from the birth record (i.e. they have gone completely unrecorded). 19. The Chimbu death rate data ranged from six to 25 persons (Brown & Winefield 1965, p. 181); for the Baiyer valley population it was eight per 1000 and life expectancy was 58 years (Becroft 1969, pp. 50, 54). 20. It is difficult to date this event accurately; I estimate that it occurred sometime in the late 1920s to early 1930s. 21. The severity of food shortages can vary considerably from one occasion to another, as evidenced by the impact of the 1972 and 1997–98 droughts. An evaluation of the 1997–98 event (AusAID Citation1998) found that a significant part of the food aid did not reach intended recipients and/or was sold. It attributes few, if any deaths, directly to starvation, finding that water shortage was more of a threat, although there was some increase in mortality with a breakdown in health services. It concludes that increased population mobility and traditional coping mechanisms (such as resorting to bush foods) averted some deaths. 22. It is probable that the medical cause of death in some cases was not literally starvation but that hunger fatally weakened people so they succumbed to other conditions (Bourke 1988, p. 86)—nonetheless, so far as their descendants are concerned, it was the famine that killed them. 23. I did this by working out the ages in 1930 of those censused in 1973 who were alive in 1930 and, noting these persons on the genealogy, drew an age isobar line cutting off all those not born then. I plotted the probable composition of sem (families) at that time, knowing these persons’ ages in 1930, and also having data on marriages and deaths tied to events with approximately known dates (such as outbreaks of armed hostilities), which allowed me to establish if deceased individuals were alive or not in 1930. 24. These violent death statistics are modest compared to those reported for some other regions of New Guinea, such as on the nearby Papuan Plateau where among the small Gebusi population the rate is five to six homicides per 1000 persons a year (Knauft 1985, pp. 376–377). Feil (1987, p. 71) gives some statistics for elsewhere in the Highlands, where the reported death rates in armed hostilities are considerably higher than reported in the Was valley, accounting for 20 per cent of male deaths among the Huli, 25 per cent among the Enga, 30 per cent among the Tauna-Awa and 32 per cent among the Usurufa; although Feil notes that he thinks these ‘figures on deaths are highly unreliable’ (p. 71); Riley and Lehmann (Citation1992, p. 76) agree that the Huli figures are ‘an overestimate’. 25. In tallying up violent deaths, the Wola would wish to include those persons killed nefariously through sorcery (woktoiz or hultort) and poison (mogtomb), which amounts to a total of 13 further men's deaths in the period under consideration (women unknown); a 40 per cent increase in deaths attributable to armed hostilities. 26. These migration figures include only those persons resident at the time of a census who had either moved into or out of the surveyed communities since the previous census; individuals and families who moved in and back out or out and back in between censuses were not picked up. 27. Although elsewhere in this region the rates were more similar to those reported here, at 4 per cent males absent (Epstein & Epstein 1962, p. 73). See Riley and Lehmann (1992, pp. 69–75) for a history and summary of migration patterns for Papua New Guinea. 28. The percentage migrating beyond the Was valley social universe is lower than in the Tari region, Lehmann (2002, pp. 55, 56) reporting that more than 40 per cent of males aged 20–39 years were absent at any time during the 1980s, although the numbers of young men migrating elsewhere for extended periods declined noticeably between the 1970s and 1980s. 29. We should interpret this Southern Highlands statistic with caution, as the changes in population reported between 1980 and 1990 for the three Nipa District census divisions of Nembi Plateau (104 per cent), Nembi Valley (76 per cent) and Wage (92 per cent) are improbably high, suggesting shortcomings with the census data. The average annual per centage rate of increase for the Highland Chimbu population was 0.24–2.28 per cent (Brown & Winefield 1965, p. 181), for the Baiyer river was 1.9 per cent (Becroft et al. 1969, p. 50) and in the Wahgi valley was 2.5 per cent (Burton 1988, p. 25). 30. There is some debate over when that occurred (Brookfield & White Citation1968; Feil Citation1987; Ballard et al. Citation2005)

Ethnology · Geography · New guinea · Socioeconomics · Sociology · Demographic Trends and Gender Preferences · Demography · Global Maternal and Child Health · Intergenerational Family Dynamics and Caregiving