Article 21

European Journal of Population 15: 241–278, 1999. © 1999 Kluwer Academic Publishers. Printed in the Netherlands.

241

Economic Growth or Survival? The Problematic Case of Child Mortality in Turkey CEM BEHAR, YOUSSEF COURBAGE and AKILE GÜRSOY Institut national d’études démographiques, 133 Bd Davout 75980, Paris Cedex 14, France Received 14 September 1998; accepted in final form 13 March 1999 The Sons of Edward sleep in Abraham’s bosom (iv.iii.38) William Shakespeare, The Tragedy of King Richard the III But who am I? An infant crying in the night An infant crying for the light: And with no language but a cry Alfred Tennyson, Memoriam, xxxvi.

Behar, C., Courbage, Y. and Gürsoy, A., 1999. Economic Growth or Survival? The Problematic Case of Child Mortality in Turkey. European Journal of Population 15: 241–278. Abstract. Turkey is a country which is demographically unclassifiable because its persistently high infant mortality is out of line with its socio-economic indicators and its low fertility. The rapid modernisation of Turkey over the last three decades, which might have been expected to have had a favourable effect upon infant survival, has not in this respect lived up to expectation. The stresses resulting from economic growth and the high level of female workforce participation have perhaps tended to distract women from child care. Also, neither Ottoman nor republican Turkish traditions have encouraged an enhancement of the status of childhood. Unconventional sources: ethnology, literature, cinema, are deployed here to construct an impression of the cultural environment of the mothers, fathers and families of dead children. Change of attitudes, very slow as far as childhood is concerned, have not yet caught up with the transition in fertility. An infant mortality rate of 53 per 1000, accompanying a total fertility rate scarcely higher than 2, is a combination difficult to find anywhere else. Behar, C., Courbage, Y. et Gürsoy, A., 1999. Croissance Économique ou Survie? Le Problème de la Mortalité Infantile en Turquie. Revue Européenne de Démographie 15: 241–278. Résumé. La Turquie est un pays inclassable en raison de sa mortalité infantile qui a été et continue à être très élevée, alors que les indicateurs socio-économiques et la fécondité auraient présagé une meilleure survie. La modernisation rapide de la Turquie depuis trois décennies, que l’on aurait pu croire plus favorable au sort de l’enfant, n’a pas rempli tous les espoirs. Le stress engendré par la croissance économique et la forte participation féminine a peut-être empêché les femmes de prêter une attention suffisante à l’enfant. Mais les traditions ottomane puis républicaine n’ont pas contribué à une valorisation suffisante de l’enfant.

242

CEM BEHAR ET AL.

Le recours à des sources non conventionnelles: ethnologie, littérature, cinéma ont permis de donner une image de l’atmosphère culturelle dans laquelle beignent les mères, pères et familles d’enfants décédés. Les changements de mentalités, trop lents en ce qui concerne l’enfance, n’ont pas encore rattrapé la transition de la fécondité. Un taux de mortalité infantile de 53 pour mille accompagnant une fécondité à peine supérieure à 2, est une configuration qui se rencontre difficilement ailleurs.

1. Introduction Child mortality in Turkey remains a puzzle. Ever since it became measurable and comparable – since thirty years – it has been very high compared to Turkish adult mortality, or child mortality elsewhere. Demographers, doctors and sociologists have failed to explain why, but have repeatedly asserted that it was Turkey’s main demographic and medical challenge (Bulut et al., 1991; Tunçbilek and Ulusoy, 1988; Behar, 1975). This high level conflicts with other demographic and socio-economic indicators. The existence of specific endemic early childhood diseases, climatic variations from the bitterest cold to the most sweltering heat, effective medical care and between-region development gaps have all been unconvincingly offered as explanations (Shorter and Macura, 1983). “The link between income level and mortality during economic development may vary, but why Sri Lanka with a per capita GDP one third that of Turkey should have half its infant mortality remains unexplained” say sociologists (Aksit and Aksit, 1989). By international organization definitions, Turkey left the “developing” country status behind more than two decades ago. Average annual GDP growth of almost 6% in 30 years, industrial levels of production and exports, place Turkey in the front ranks of South-East European and Middle Eastern countries. Illiteracy is low among women, very low among men. There is more or less universal primary education, high secondary education enrollment, and decent medical coverage. These are all established facts. The demographic indicators are good and fertility declining so rapidly that the 2-child level will be passed by about the year 2000. Broadly speaking, Turkey seems to have completed its demographic transition (Behar, 1997). Turkey’s high economic growth and rapid modernization, at whatever painful cost, are undeniable facts. But has this record pace of modernization been sufficiently internalized by society? Has it actually worked against the intended purpose, against children? The outcome often perverts or thwarts the aim. The upsurge in Islamism in Turkey during the Eighties and Nineties was undoubtedly a political backlash against unbridled westernization and official secularism, as well as reflecting a global trend towards revivalism in religion. One reason why Turkish infant mortality is so much above the norm may be because modernization has not lived up to expectations. Another may be that children are traditionally paid scant attention. If so, opposing trends would paradoxically have worked to the same end.

CHILD MORTALITY IN TURKEY

243

Research on the determinants of child mortality has focused chiefly on their intermediate health variables, through their physiological or medical effects. Variables like dietary history, duration of breastfeeding, infant morbidity, the incidence and role of respiratory, intestinal or other infections and diseases, have all been subjected to detailed statistical and epidemiological analysis (Martorell and Ho, 1984). Socio-economic variables have been compared with the mortality rate, pregnancy ratio, parity, household size and structure, income, education and occupation (Mosley and Chen, 1984). Despite this, the precise effect of these socio-economic variables on the effective death rate remains unclear. Also, how these variables articulate with cultural attitudes and practices in the broad sense, as well as with policies and broad ideologies, remains largely unexplained to the extent that these determinants of child mortality, advanced as authoritative explanatory variables, are a kind of black box which has not yet yielded up its secrets (Gürsoy, 1992). The intangible social environment, that which is not immediately measurable, may be a more promising field of investigation to explain this mortality (Caldwell, 1986). Infant and child mortality rates have long been regarded as simple statistical indicators, as consequences of a country’s development level and the effectiveness of its health policy (Rozensweig and Stark, 1997). But child mortality is also and above all a silent and significant reflection of the life styles of the community under study, the relative values attached to gender and age, and the cultural norms governing the balance of power between age groups and generations. Often, it reflects the parents’ lack of parenting skills and ignorance, the mother or family’s ignoring of instructions in tending a sick child, and even, in some cases, an involuntary but nonetheless fatal neglect (Scrimshaw, 1978). But it could in reality be a matter of shifting gender- and age-specific hierarchies, powers and balances. Scrimshaw shows in the case of Latin America that there is a degree of “avoidable infant mortality” attributable to “selective neglect” and that the “degree of unwantedness” of children is a factor in infant mortality (Scrimshaw, 1978).These attitudes, which are only very remotely and indirectly reflected by the official figures, are enshrined in a tangle of what may be nebulous and diffuse values, complex norms and breaches of those norms. With child mortality, too, we enter the realm of the unsaid, the taboo. Surrounded by such silence, how are we able to identify the values, cultural attitudes and practices which may be at the root of this abnormally high mortality? Silence in Turkey starts with the source data. Vital records are very incomplete. Registration data, even in cities and large towns, cannot be used as they stand to calculate mortality, still less infant mortality (Bulut et al., 1991). We therefore chose a number of usual alternative sources – censuses, surveys, etc. – and nontraditional, not to say unusual, sources for demographers, falling neither within the field of demography nor that of other social sciences strictly speaking. The idea was to delimit a cultural atmosphere, a mesh of diffuse reactions, approaches and attitudes about children and death. Where complex human problems are involved,

244

CEM BEHAR ET AL.

where more obvious to turn than to literature? Thus, echoing the experience of two physicists “. . . the field of what is not understood, even in physics, is so vast . . . that there is no reason to try and take a hard science approach to complex human problems. It is wholly legitimate to turn to intuition or literature for a kind of nonscientific understanding of aspects of the human experience which elude more rigorous analysis” (Sokal and Bricmont, 1997). We therefore turned to literary sources as well as compiling ethnographic records of rural life. To do this, we examined films, scenarios and modern dramatic productions for themes and even direct or indirect references to infant and child death. We lay no claim to have unearthed the solution – that is out of the question as the current state of knowledge stands. Rather, we have striven to situate Turkey’s child mortality within a general cultural context. Neither quantitative, nor qualitative data will give the reader a full and meaningful perspective of the social problem. These qualitative sources must obviously be approached with caution, and used in conjunction with statistical evidence. 2. The macro-demographic evidence 2.1. M ARKED

DECLINE

In the absence of efficiently-kept vital records, 7 surveys and 3 censuses (the 1990 has not yet been analyzed for mortality). Over the period 1961 to 1993 show the progress achieved in reducing child mortality. The results are grounds for optimism. Turkish infant mortality has declined in a generation by three-quarters from almost 200 per 1000 to 53 per 1000 at the beginning of this decade. Child mortality of approximately 60 per 1000 around the Seventies has fallen even further to a seventh of its initial value, and now stands in single figures: 9 per 1000 in 1991–1993 (Table I). 2.2. AGE - SPECIFIC

BIAS

Descriptively, the atypical picture of Turkish mortality – high child mortality, moderate adult mortality – is long-established. Infant mortality gave model life table values lower than those obtained at other ages. Mortality was close to the most atypical of the Coale and Demeny models, the “East” model, but even “further East than the East model” (Behar, 1975; Bulut et al., 1991). High infant mortality among deaths in the first 5 years of life, and that mortality itself over-represented in mortality at all ages, are the salient features in Turkey, where higher order births seem systematically to receive less favourable treatment than lower order births. In the same way, in Istanbul, there is evidence of very high child mortality in the first week of life compared to total mortality in the first 12 months (Bulut et al., 1991), plus a high stillbirth rate: 20 per 1000 live births in 1988–1992. Questions arise as to the possible significant rate of false stillbirth deaths during the neonatal period. Foetal mortality (determined by declared miscarriages) – 166 per 1000 live

245

CHILD MORTALITY IN TURKEY

Table I. Infant and child mortality rates (per 1000) – 1961–1991 Age Year

0

1961 1966–67 1966–67 1968 1973 1975 1978 1983 1986 1986 1989 1991

190.0 153.0 156.0 149.2 157.6 134.0 126.4 108.7 82.2 85.0 62.3 52.6

1–4

0–4

59.4 64.3

199.7 211.8

46.6 37.2 16.7 24.0

167.1 141.9 97.5 107.0

8.8

60.9

Sources: F. Shorter, The Decline of Infant and Child Mortality: Estimates from the Turkish Census, The Population Council Cairo, 1989. Bernard Berelson, ‘Turkey: National Survey on Population’, Studies in Family Planning, December 1964 (for 1961). Ministry of Health, Demographic and Health Survey 1993, Ankara 1994 (recent years).

births – is also very high. Overall, perinatal (late foetal, early live-born), infant and child mortality raise questions about women’s attitudes to their own and their children’s health, medical practice, interactions between health care providers and users (Bulut et al., 1991). 2.3. S EX - SPECIFIC

BIAS : EXCESS FEMALE MORTALITY

The natural source of excess female mortality is son preference. There is avowed son preference in Turkey. Women’s stated additional wanted son preference was 0.82, against 0.64 for daughters – 28% more – in 1988 (Unalan, 1993). But son preference is less intense than the mortality results might suggest (Table II). The sex-differential child mortality standards were borrowed from Hill and Upchurch, based on the historical experience of North Western Europe which is considered to be among the least sexist regions of the world (Hill and Upchurch, 1995). Without exogenous intervention, female mortality is naturally lower than male mortality because of girls’ greater constitutional robustness. In Turkey, on the other hand, female mortality has always been abnormally high. Paradoxically, when general mortality was high, as it was 30 years ago in 1966–1968, female mortality was in line with Northern European standards. More recently, female disadvantage has

246

CEM BEHAR ET AL.

Table II. Trends of female excess mortality 1966–1993 Neonatal 1966–68

Post neonatal

Infant (0)

Child (1–4)

Under 5

B G G/B

163.7 145.2 0.887 0.823

59.1 54.6 0.924 0.941

213.1 191.9 0.901 0.867

B G G/B

140.0 138.0 0.986 0.817

33.7 51.0 1.513 0.930

169.0 182.0 1.077 0.857

B G G/B

72.5 62.7 0.865 0.786

16.2 16.8 1.037 0.876

87.4 78.6 0.899 0.810

70.5 66.0 0.936 0.786

12.4 13.6 1.097 0.876

82.0 78.7 0.960 0.810

Standard 1976 (Yozgat) Standard 1989

Standard 1993

Standard

B G G/B

40.7 34.0 0.835

29.7 32.0 1.077

Sources: Various surveys from 1968 a` 1993. Note: B = boy; G = girl. For Yozgat (small town), (Shorter and Macura, 1982). For Istanbul see Hill and Upchurch (1995).

increased: an excess female mortality of 15% in the first year of life and of 22% in the 1–4 year age group. In all, some 10000 young girls died every year from 1983–1993 above the already high male norm. Excess female mortality appears in early life and rises in the postneonatal period, when female probability is 8% above male probability (normally 21% less according to the standard). 3. Excess mortality in Turkey compared to regional norms 3.1. H IGHER

LEVELS , SLOWER RATES OF DECLINE

Table III paints a far less optimistic picture than Table I. Turkey’s performance is gauged against those of neighbouring countries, according to the infant mortality rate at the start of the decade and the trend between 1960–1964 and 1990–1994. What counts as a “neighbouring country” is necessarily an arbitrary decision. We have used those with shared borders – Bulgaria, Greece, Cyprus (to some extent, since the partition of the island in 1974), Syria, Iraq, Iran, Armenia and Georgia – as well as others which have had strong political and cultural interactions with

247

CHILD MORTALITY IN TURKEY

Table III. Infant mortality (per 1000): Turkey and neighbouring countries Population Infant mortality Index (1000) early 90’s early 60’s (1990/1960)∗100 TURKEY

60838

53

190

27.9

BALKANS/CYPRUS Bulgaria Greece Cyprus Albania Macedonia Serbia/Montenegro Bosnia Romania

61795 8509 10454 745 3383 2156 10251 3569 22728

20 16 9 9 32 27 21 15 24

64 36 50 29 99 117 80 94 60

30.8 44.4 18.0 31.0 32.3 23.1 26.3 16.0 40.0

ARAB MIDDLE EAST Syria Iraq Lebanon Jordan Palestine Arabia Gulf Emirates

60470 14203 20095 3009 3998 2194 13550 3421

35 35 44 28 34 28 29 14

129 125 130 62 125 125 160 82

27.0 28.0 33.8 45.2 27.2 22.4 18.1 17.1

IRAN

59162

32

163

20.0

TRANSCAUCASIA/ CENTRAL ASIA 64727 Armenia 3632 Georgia 5450 Azerbaijan 7531 Kirghistan 4460 Uzbekistan 22762 Turkmenistan 4075 Kazakhstan 16817

37 25 23 33 39 43 57 34

68 34 48 50 75 79 95 66

55.1 73.5 47.9 66.0 52.0 54.4 60.0 51.5

Sources: Turkey, Table I. Balkans, Transcaucasia and Central Asia, United Nations, World Population Prospects As Assessed in 1996, New York, 1997, Arab Middle East, Y. Courbage, ‘Political and economic issues of fertility transitions in the Arab World – Answers an open questions’, IUSSP Population and Environment, Vol. 20(4), March 1999. Iran, M. Ladier-Fouladi, ‘Fertility transition in Iran’, Population – An English Selection, Vol. 9, 1997.

248

CEM BEHAR ET AL.

Turkey through being in the Ottoman Empire – Albania, Macedonia, Serbia and Montenegro, Bosnia, Romania, Lebanon, Palestine, Jordan, Saudi Arabia, United Arab Emirates. The list of former Ottoman Empire Arab countries should also have included: Yemen, Egypt, Libya, Tunisia, or Algeria where the Ottoman influence was diffuse. Their inclusion (with the possible exception of Yemen) would not have affected the analysis, however. Transcaucasia and Central Asia – the “cradle of the Turks” – are also included for their close linguistic ties with Turkey. Apart from tiny Turkmenistan (population 4 million), Turkey’s infant mortality is higher than that of its neighbours (INED, 1997). Before the Gulf War (1991), Iraq’s infant mortality of 44 per 1000 was lower than Turkey’s (Al Saadi, 1996). Defeat and the resulting embargo rapidly tripled its infant mortality to 127 per 1000 (United Nations, 1997). In the Balkans (and Cyprus), the average infant mortality of 20 per 1000 is approximately a third of that of Turkey, falling to less than 10 per 1000 in Greece. The Arab Middle East (35 per 1000), and Iran (32 per 1000) both significantly out-perform Turkey. Infant mortality in Transcaucasia and Central Asia (37 per 1000), though far from the levels achieved in the Balkans, is a third lower than in Turkey. Turkey therefore stands out as a pocket of high infant mortality in the midst of countries where it is no longer more than residual. Even so, Turkey has progressed significantly, because infant mortality has fallen from 190 to 53 per 1000 in a generation. How did its neighbours fare over the same time? The former USSR, Bulgaria and Romania performed poorly. But the Arab Middle East – and especially Iran – outperformed Turkey, with a comparatively faster decline in infant mortality between 1960 and 1990, although from a lower baseline. Generally, the opposite is observed: the higher the benchmark level, the more acute the relative decline. Turkey’s high infant mortality is not a recent phenomenon, therefore, but a long-term one.

3.2. T URKEY,

A LESS - FAVOURED COUNTRY ?

And yet demographically, economically and culturally-speaking, Turkey is not so badly off. Let us briefly consider the indicators (World Bank, 1997) most closely correlated to child mortality. Demographic factors: Urban residence, although in Turkey carrying an undeniable feature of urban ruralization, offers a higher probability of child survival; this is shown by all social surveys. With an urban population of 67%, Turkey is well placed in this respect, better so than the Balkans (58%), Iran (59%), Transcaucasia and Central Asia (51%), and almost equal with the Arab Middle East (71%), which is far less favoured by another factor highly correlated to infant mortality – fertility – which stands at 4.60 children per woman, two more than Turkey’s 2.65, which is also lower than Iran (3.26), as well as Transcaucasia and Central Asia (3.09). In the Balkans (excluding Albania), both mortality and fertility are lower than in Turkey.

CHILD MORTALITY IN TURKEY

249

Economic factors: Turkey’s per capita gross domestic product was estimated (1996) at $7,843 in purchasing power parity. Few other comparable countries are so prosperous, apart from Greece ($12,195) and Saudi Arabia ($10,684). Turkey has a higher standard of living than the Balkans (excluding Greece), the Arab world, Transcaucasia or Turkish-speaking Asia. Even oil-rich Iran has a lower standard of living than Turkey. Educational standards: Illiteracy has not yet been eradicated from Turkey, where 8.3% of males aged 10 and above, and especially 27.6% of women, lack basic literacy skills. So, the lower mortality in the Balkans, Transcaucasia and Central Asia can partly be attributed to the elimination of illiteracy (Greece still has 5% female illiteracy). The driving force is universal school enrollment, especially in secondary education which is more widespread than in Turkey where the gross secondary enrollment ratio is 61%. Imperfect as it may be, however, the Turkish educational system quantitatively outperforms that of the Arab Middle East, and to some extent Iran. Arab (18.9%) and Iranian (21.6%) male illiteracy is double that of Turkey. Turkish women are also more literate than their Arab (37.7%) and Iranian (34.2%) counterparts. The secondary education intake is comparatively higher in Turkey, where the gross enrollment ratio of 61% is higher than in the Arab Middle East (45%), and a little less than in Iran (66%). Healthcare provision: High child mortality in Turkey cannot be blamed on inadequate medical provision. While in terms of medical density (number of persons per doctor), Turkey may be less well-off than its neighbours in Europe, Transcaucasia and Central Asia, with one doctor per 976 people, it has comparatively better medical provision than the Arab Middle East (1,128), and especially Iran where medical density is third that of Turkey. As regards access to healthcare, drinking water and drainage, Turkey is close to European countries and better served than the Arab Middle East and especially Iran, where only 73% of the population has healthcare coverage (proportion for whom treatment for common diseases and injuries are available within one hour’s travel or walk), compared to 100% in Turkey. 78% of births are medically-attended. This makes Turkey’s record on immunization all the more disconcerting. Generally, the healthcare coverage rate is matched by an approximately equal immunization coverage rate. In Iran, it is even substantially higher (95–97% children vaccinated for 78% of population covered by healthcare provision). In Turkey 25% of children are not immunized against measles, and 14% are not immunized against diphtheria/tetanus/whooping-cough (pertussis (DTP)). The 1993 DHS survey data are even lower than those given here, drawn from international sources (DTP): 22% of children were not immunized against measles, 23% for DTP. Also, the proportion of parents that were unable to produce the child’s immunization card is particularly high in Turkey (47%), suggesting that the number of immunized children may have been over-estimated. This lack of coverage is the more serious since nearly a third of children have not received all the WHO-recommended

250

CEM BEHAR ET AL.

immunizations. Of those who have been immunized, another third were vaccinated only after the age of 12 months, so receiving reduced protection. The dropout rate for combined vaccines (DT and polio) is high. 95% of children aged from 12 to 23 months received the first dose of DTP vaccine, but only 78% the third. The comparable figures for polio are 94% and 77%. In Jordan, for example the figures are 97% and 95% for the former, 96% and 95% for the latter (Sommerfelt and Piani, 1997). In short, the indicators should put Turkey in an enviable child survival position in the region, especially compared to the Southern countries, the Arab world and Iran, even more than with the other European and culturally Europeanized countries formerly part of the Soviet Union. The only exception is Turkey’s still imperfect immunization coverage rate, which could account for its high mortality. On the face of it, parental demand for immunization would appear to fall short of what the medical system can supply. One figure illustrates this imbalance well: 16% of Turkish women with secondary or higher education had not had their child immunized – more than one child in 6 in 1993. By accident or by design? But immunization is so direct a determinant of child mortality that to explain the high level of the one by the low level of the other seems tautological. Important as it may be, however, this immunization deficit compared to the Arab world and Iran – approximately 20% for measles and 11% for DTP – cannot account for a 50 to 60% higher infant mortality (Desgrees and Pison, 1995).

3.3. S YRIA

AS A CONTROL CASE

Model life tables were used to describe the characteristics of child mortality in Turkey. The choice of standard for differentials is less obvious. In the previous section, we gave overall indicators for 22 countries. But the determinants of mortality from those countries cannot be used for want of comparable data. The European, Transcaucasian and Central Asian countries must be discounted because the DHStype survey are the exception (Barbieri et al., 1996; Ministry of Health Uzbekistan, 1996; Ministry of Health, Kazakstan, 1996). The demographic size effect means that the sparsely-populated countries of the Arab Middle East – Lebanon, Jordan, Palestine, the Gulf States – can be eliminated. Saudi Arabia has certain mortalityspecific characteristics, extremely high investments on children, particularly on health, which enables a much lower mortality than expected. Pre-Gulf War Iraq, with a population of nearly 20 million, would have been a suitable case for comparison with Turkey, but its PAPCHILD survey has not been published. Iran, as populous as Turkey, would have been even more relevant, but there is no national survey of comparable scale, only regional (Agha et al., 1996). In the final analysis, Syria, notwithstanding its size – one-fourth of the Turkish population – seemed the best bet: four centuries of shared history with Turkey (1516–1918), geographical proximity, similar climates, religious composition and even language – Arabic is spoken in Southern Turkey, Turkish in Northern Syria –, family migration and

251

CHILD MORTALITY IN TURKEY

Table IV. Infant and child mortality in Turkey and Syria (per 1000) Period

Neonatal

Post neonatal

Infant (0)

Child (0–4)

Under 5

Turkey Syria Turkey Syria Turkey Syria Turkey Syria Turkey Syria 1978–82 37.5 1983–87 44.6 1988–93 29.2

15.7 16.7 18.0

54.5 36.9 23.4

14.6 13.8 16.6

92.0 81.5 52.6

30.3 30.5 34.6

23.7 16.8 8.8

8.8 7.6 7.4

113.5 96.9 60.9

38.8 37.9 41.7

Sources: Turkey, DHS 1993. Syria, PAPCHILD/Central Bureau of Statistics, Mas’h al -joumhouriat al arabiya al souriya. Op. cit.

frequent intermarriage. Finally, the 1993 PAPCHILD survey (PAPCHILD/Central Statistical Office, 1995) was carried out in the same year as Turkey’s DHS survey, and the data is of good quality. For Syria to be used as a control country for Turkey may however, seem odd given the tendency in Turkey to look to the West (or occasionally the East) rather than to the South. Despite Turkey’s accelerated mortality decline in the 0–5 year age group – halved in the space of 10 years – compared to Syria’s stable or slightly increased rate, the level in Turkey remains appreciably higher: 46% more in the 0–5 year age group. In the post-neonatal period – 1 month to 1 year – excess mortality in Turkey stands at 41%, and 19% for child mortality. A variable immunization effect between the two countries is discernible here, but not precisely quantifiable. In Syria, 87% of children under the age of 5 have a health record (including the immunization schedule), compared to just 42% in Turkey. Without pushing the comparison too far, we would simply note that while 64.7% of Turkish children aged 12–23 months were fully immunized, the same was true for 73% of Syrian children. But it is during the neonatal period (the first 28 days of life), when immunization has virtually no influence on mortality, that Turkey’s disadvantage is most marked: 62% excess mortality (Table IV). Table V shows the various socio-economic and demographic factors of child mortality in Turkey and Syria. Although both surveys were conducted in the same year – 1993 – the observations may refer to different base periods. This is not a particularly damaging flaw, insofar as the decline in Syria’s mortality has leveled-off over the past 10 years. The outstanding feature is that, with very few exceptions, the child mortality rate among Turkey’s better-off socioeconomic and population groups is in every case worse than that of the less affluent Syrian groups (Table VI). Infant and child mortality is 25% higher among Turkey’s urban population (51 per 1000) than Syria’s rural one (41 per 1000), the only exception appearing is in the postneonatal period (14 compared to 17 per 1000). More perplexing still is the premise often advanced that the mother’s educational level is critical to child survival, when the educated Turkish woman runs a 51 per 1000 risk of seeing her

252

CEM BEHAR ET AL.

Table V. Child mortality differentials – Turkey and Syria Neonatal

Post neonatal

Infant (0)

Child (1–4)

Under 5

Turkey Syria Turkey Syria Turkey Syria Turkey Syria Turkey Syria Residence Urban Rural

29.9 28.1

17.2 17.5

14.1 37.4

13.1 17.1

44.0 65.4

30.3 34.6

6.8 11.8

8.8 6.2

50.5 76.4

38.8 40.6

Education No/Inc. Primary 31.4 Prim. complete 27.9 and above

17.9 13.9

36.5 15.7

15.8 9.5

68.0 43.6

33.7 23.4

12.6 6.1

7.5 8.6

79.7 49.7

40.9 31.8

Region West South Centre North East

29.7 34.6 29.4 16.2 29.9

13.0 20.8 28.5 28.0 30.1

42.7 55.4 57.9 44.2 60.0

5.6 7.8 12.0 5.6 11.0

48.0 62.8 69.2 49.5 70.4

Prenatal care Without With

27.9 29.6

38.9 14.3

66.8 43.9

10.7 6.1

76.8 49.7

Mothers’s age –20 20–29 30–39 40–49

52.0 27.7 55.8 41.8

17.4 13.7 19.1 29.1

40.8 27.3 32.1 60.2

15.0 15.9 12.9 15.1

92.8 55.0 87.9 101.9

32.4 29.6 31.9 44.3

11.9 13.5 12.7

8.6 5.9 5.7 15.9

103.5 67.8 99.5 101.9

40.7 35.4 37.5 59.5

Birth rank 1 2–3 4–6∗ 7+

37.4 26.2 41.1 75.4

18.0 14.0 19.1

26.6 24.4 39.3 49.7

11.0 15.4 16.4

64.0 50.6 80.3 125.1

29.1 29.4 35.5

8.8 11.0 20.0 16.5

5.5 6.3 8.9

72.2 61.0 98.7 139.5

34.5 35.5 44.1

Birth intervals –2 years 2–3 4 and +

63.3 23.2 20.4

34.0 11.7 12.4

50.1 27.3 15.0

24.1 10.4 6.9

113.4 50.4 35.4

48.1 22.2 19.3

24.5 11.5 3.9

8.1 7.6 8.8

135.1 61.3 39.1

55.8 29.6 28.0

Sources: Ministry of Health, Turkey Demographic and Health Survey, op. cit. PAPCHILD/Central Bureau of Statistics, Mas’h al -joumhouriat al arabiya el souriya . . . op. cit. ∗ 4 and more in Syria.

253

CHILD MORTALITY IN TURKEY

Table VI. Cases where upper level child mortality in Turkey is higher than that of lower levels in Syria

Residence Education Mother’s age Birth rank Birth intervals

Neonatal

Post neonatal

Infant (0)

Child (1–4)

Under 5

Yes Yes Yes Yes No

No Same Yes Yes No

Yes Yes Yes Yes No

Yes Yes Yes Same No

Yes Yes Yes Yes No

child die before the age of 5–25% higher than the illiterate or semi-literate Syrian woman, for whom the death probability is 41 per 1000. Even where born at optimum maternal age, i.e. between 20–29 years, Turkish children have a higher death probability (68 per 1000) than Syrians born in less favorable circumstances, to very young (41 per 1000) or very old (60 per 1000) mothers. Likewise, in birth order terms, 2nd or 3rd order children are perhaps less vulnerable in Turkey (61 per 1000) than lower, or higher order births, but have a higher probability of dying than 1st order (35 per 1000) or 4th and higher order (44 per 1000) Syrians. Only birth intervals are the exception to the rule: a 4 year or more child spacing produces a fall in mortality before the age of 5 in Turkey with 39 per 1000; in Syria, a child spacing of under 2 years – high risk judged by inter-Turkish differences – produces a higher mortality of 56 per 1000 in Syria. Comparison with Syria was not possible on other criteria. There is considerable inequity of child mortality across regions in Turkey: 70 per 1000 in what is generally regarded as the under-developed East, but not vastly less in the West which includes the most prosperous areas of the country, Istanbul, the Aegean coast and European Turkey – 48 per 1000 or 32% less. It must be emphatically stressed that Syria’s lower child mortality is not due to a higher level of medical care during pregnancy and childbirth (Boukhaima, 1997). At the same survey date (1993) 63% of Turkish women and 54% of Syrian women benefited from antenatal care. 64% of the latter opted for home births against 40% of Turkish women. Traditional birthing attendants were present at 20% of births in Syria against 13% in Turkey. Antitetanus vaccine was administered to just 19% of Syrian women against 43% of Turkish women. This comparison between neighbouring countries, the use of a control case to emphasize differentials, suggests that child mortality in Turkey is not really governed by the standard determinants. For an urban-born educated Turkish woman of low parity at intermediate age, benefiting from antenatal care and medical attendance at birth, the loss of a child is nothing out of the ordinary.

254

CEM BEHAR ET AL.

Table VII. Female disadvantage indices∗ in Turkey and some neighbouring countries Infant (0)

Child (1–4)

Under 5

Turkey

0.150

0.221

0.150

ARAB MIDDLE-EAST Syria Leban. Jordan Palestine (West Bank and Gaza) Saudi Arabia Gulf emirates

0.101 0.079 0.102 0.247 0.055 0.108 0.019

–0.070 0.071 0.079 0.081 –0.152 –0.342 0.162

0.062 0.075 0.088 0.217 0.005 0.006 0.068

∗ Excess of actual ratio of female to male mortality over the expected ratio of

female to male mortality in the standard population of Hill and Upchurch (1995). Source: DHS and PAPCHILD surveys most recent estimates.

3.4. E XCESS

FEMALE MORTALITY, A TURKISH PHENOMENON ?

Table VII gives the measures of female disadvantage in Turkey and the Arab Middle East. Discrimination against women in the form of poorer nutrition and an unwillingness to secure medical attention is as common in Turkey as in the rest of the Muslim world and Middle East. Truth or stereotype? We recalculated the measures of female disadvantage for infant and child mortality and in the 0–4 year age group based on recent survey data for the Arab Middle East (except Iraq). The results speak volumes. Except for Jordan, where female disadvantage is oddly more pronounced than in Turkey, the Arab Middle East comes out better, although female disadvantage clearly persists, but can be estimated overall at 6%, against 15% in Turkey, for girls under the age of 5. So excess female mortality cannot be attributed to the shared religion or culture, but rather to society-specific features. Might it be a family formation strategy shaped to ideal norms, which affects both girls and boys, but the former far more than the latter? 4. Socio-economic or cultural determinants? Two recent sets of baseline data enabled us to analyze infant mortality and question the validity of the real influence of certain variables on child mortality. One is an anthropological research based on in-depth interviewing conducted in 1988 on a sample of a thousand households in an Istanbul inner suburb populated by relatively recent rural incomers (Gürsoy, 1992). This survey had the enormous advantage of using open-ended questions purpose-designed to identify the various family and cultural factors of child mortality. The other is that of the 1993 national

CHILD MORTALITY IN TURKEY

255

DHS survey (Ministry of Health Turkey, 1994) on a representative sample of 6,519 ever-married women. This survey was not primarily referable to child mortality, but the retrospective birth history data allow it to be estimated with fair accuracy. The detailed information on economic and social household characteristics enabled us to calculate regression coefficients of mortality factors for the purposes of this study. These two surveys together clearly demonstrate the importance of nonconventional factors on Turkish infant mortality, which seems to be highly atypical, as shown above. The mortality index by which the variables are correlated is the ratio of the number of child deaths per woman to the average number of child deaths for all the women surveyed: observed deaths / expected deaths, known also as the ‘Preston’s’ index (Farah and Preston, 1982). The index ranges from 0 for women with no deaths to 10.4 in the extreme case. It is exactly 1 for women with a number of deaths of children equal to the sample average. The coefficients in fact, were calculated on 14 variables including the mother’s rural or urban origin, the number of years’ urban residence, household consumer durables, household daily bread consumption, a composite religious observance index, an index of the general health of all members of the household, the woman’s educational standard, the educational standard of both the wife’s parents, etc. But only 4 variables have significant regression coefficients: alcohol, tobacco intake in the household (+0.13), female negative attitude towards abortion (+0.13), patrilocal and extended households (+0.20), husband’s level of education (–0.24). Significantly, variables related either to the husband’s behaviour and educational standard, or to household size and structure have more influence on child mortality than the more traditional directly mother-related variables – urban or rural origin, woman’s educational level, health, etc. This characteristic is often attributed to women’s disadvantaged status and their situation of dependency within extended patrilocal families. The questionnaire and interviews conducted for the Gockent-Istanbul survey had the invaluable advantage of not only describing the household structure at the precise interview time – hence immediate – but also reconstituting household size and structure at the time of each birth to the woman, whether the child subsequently died or was still living. The index was 0.631 for nuclear families and 1.336 for extended patrilocal families. (Note that the DHS survey provided information on households at the time of the survey only). This difference observed also in some African countries (Caldwell, 1979) was attributed to the fact that where presumably more traditionalist parents-in-law were living with the family, the mother’s freedom and ability to take decisions as to her own child’s health and well-being would be restricted (Caldwell, 1979). The high negative correlation of child mortality with the father’s educational level would reflect this same patriarchal skew. Close to Arab and Iranian kinship patterns, Turkey has a high consanguineous marriage rate, especially in rural society: 22.8% of women in the 1993 DHS sur-

256

CEM BEHAR ET AL.

vey self-reported as having married a relative (a cross first cousin in about half the cases). Gokalp (1994) refers to the still-dominant Turco-Mongolian exogamy (that of the Turks before their arrival in Anatolia), and the more marginal so-called “Arab” endogamy (which became more common after the conversion of the Turks to Islam). The Turco-Mongolian system is rigidly inegalitarian, whereas equality of the partners remains an avowed ideal in “Arab” marriages. The author ranks elder-younger hierarchy as the most glaring inequality of the Turco-Mongolian system. In a private interview, this author told us that in the exogamic system which remains predominant in Turkey, the husband still perceives both the wife and children as outsiders, and that the family strategy in this type of marriage is to invest in one child only, even if that entails neglecting the others (Gokalp, 1994). Responses like “respect for the family order” were prominent among answers to the question on the reasons for this type of marriage preference, the subtext being: making sure that the daughter-in-law respects her parents-in-law, who are also her blood relatives (Gokalp, 1994), and her in-laws in general. It is likely that the extended patrilocal families in the 1986 survey contained a far higher proportion of consanguineous couples than the nuclear families. Effects of consanguinity on child mortality are mixed or ambiguous, however, in Turkey (Tuncbilek and Ulusoy, 1988) as elsewhere (Boisvert and Meyer, 1994). The socio-economic variables, like living environment, occupation, income, mother’s educational standard etc., have far less influence on child mortality than other factors (Table VIII) which are harder to delimit and quantify. The same applies to women’s attitudes to abortion, as determined through open-ended questions. All other things being equal, the higher the disapproval rating of abortion, the higher the index. Women who stated that pregnancies absolutely had to be carried to term were found to experience the highest relative number of child deaths, whereas those who were more pro-abortion in principle, seeing it as the woman’s right to choose, experienced fewer deaths of children. The most proabortion women demonstrate the type of women who are more ready to take control of their own reproductive life course. We are unaware of any research on the link between women’s (or the family’s, or men’s, or the community’s) attitude to contraception and abortion and later child mortality rates. The Turkish case is a challenge to the reproduction control process, which is not limited to fertility regulation alone. Lastly, the coefficients show an increased mortality risk for children of heavy drinkers and smokers. The reason is less important in this particular case. It may simply be a matter of an unhealthy environment, the diversion of household income into these purchases or alcohol-induced violence. What is striking is the link between child mortality and conditions created by people other than the mother or the child itself: in the Göckent study, cigarette and alcohol consumption in the household was measured only for the men of the household, since no woman smoked or drank significantly. The analyses we carried out based on the 1993 DHS survey reveal striking similarities with the above anthropological field research

CHILD MORTALITY IN TURKEY

257

carried out in the 1988 in Istanbul: IUD use, previous use of contraception, husbands level of education. The statistically most significant of all these variables is the last. As in the Istanbul-Göçkent in-depth survey, the husband’s educational standard and occupation on the one hand, and the woman’s attitude to birth control on the other were among the most significant factors in child mortality. The 1993 DHS survey did not include questions from which the pattern of household consumption (of alcohol, tobacco etc.) could be deduced. Women were classed by duration of marriage, then assigned a mortality index of actual deaths divided by expected deaths for the age group. As it was a large-scale national survey, we also carried out simple regression analyses on nearly 300 traditional variables (demographic characteristics, type of accommodation, household income, occupation, education etc.). These analyses invariably gave the expected results, although with fairly low coefficients. Unsurprisingly, the woman’s educational standard was not statistically significant in the multiple regressions here either. Contrary to common belief, higher educational levels among Turkish women are no automatic guarantee of child survival. The two variables are not directly linked, in fact. On the contrary, situations could quite easily be imagined in the Turkish cultural context where too wide a gap in either direction between the wife’s educational standard and that of her immediate family circle could marginalise and deprive her of some or all of her network of social relations. Women’s approach and attitudes to contraception prove to be crucial (the DHS survey had no questions on attitudes on abortion) here too. In Turkey where, in 1993, 66.2% of married couples regularly used some means of contraception (Ministry of Health Turkey, 1994). An analysis of the 1993 DHS data shows that a woman’s more or less regular previous use of family planning method had a significant reducing influence on her child mortality. So, the children of a woman prone or willing to use IUD will have a lower probability of dying in infancy. Conversely, low or no IUD use are associated with higher mortality. It therefore seems clearly established that women’s control of the reproduction process (both in terms of birth timing and net completed fertility) is not limited solely to contraception. A few additional words must be said about a comparison between the two surveys. In Turkey as a whole (1993 survey), prosperity correlates negatively with child mortality. But in Göçkent, which was a relatively homogeneous, poor community, capital accumulation was found to correlate positively with child mortality. In this group, ownership of consumer durables (16 items, including TV, radio, stereo system, etc.) or property either in Istanbul or the region of origin, was detrimental to child survival. This tends to suggest that an economic step forward may be a step back in terms of child survival for poorer communities. Such dilemmas reflect global trends that pose similar issues at micro and macro levels (This was also the subject of Hülya Koçiyigit’s film Kurban). The dilemma presented by economic growth versus human costs has also remarkably been explored in other contexts. An essay pointed out that ‘the so-called Brazilian Economic Mir-

258

CEM BEHAR ET AL.

acle, (which is) a policy of accumulation . . . has increased . . . the Gross National Product (but the) childhood mortality rate has been steadily rising throughout the nation since the late 1960’s’. The author, N. Scheper-Hughes, mentions political leaders advocating free distribution of children’s coffins in the district as election propaganda (Scheper-Hughes, 1985). Both the 1993 DHS survey, and the anthropological study of 1988 confirm the existence and reveal the unsuspected importance of unexpected local cultural circumstances, in the broad sense of the term, on child mortality. Traditional socioeconomic determinants like housing, income, occupation, educational standard, health and their attendant factors obviously recur in the statistical analysis, but are not really explanatory. The way in which these unexpected variables act on child mortality is clearly quite complex. Among other things, it involves a set of attitudes and patterns of conduct by women towards contraception and abortion, a specific type of male-centered family structure, male education, relatively heavy smoking habits and alcohol consumption on the part of the men in the household, all far more to do with a diffuse everyday cultural atmosphere than conventionally used, clearly defined, comparable and standardized socio-economic variables, or with variables directly attributable to the mother herself. These “determinants” are difficult if not impossible to quantify and delimit statistically. We were therefore compelled to turn towards different types of source. 5. Child mortality as a fertility regulating method? High child mortality is certainly an untoward consequence of high fertility. But in certain circumstances it can become a fertility regulating method. The case of Turkey illustrates the specific relation between these two phenomena. 5.1. M ODERATE

FERTILITY

Judged by developing world standards, Turkish fertility has never been very high. All censuses and surveys evidence clearly indicates that cohort fertility was always relatively moderate (Table VIII). Mustafa Kemal Atatürk “the father of the Turks”, set a lead followed by many politicians, senior civil servants and executives, having no other children than . . . the Turkish nation. Without taking population control to quite those lengths, all female generations born before the Second, and even before the First World Wars averaged an altogether reasonable fertility of 5.5 children per woman. This cohort result is more or less borne out by period rates (Shorter and Macura, 1997). This number – more or less that of old Europe (Coale and Watkins, 1986) – is significantly lower than the long-time norm for the South of Turkey. In Syria, for example, it has seldom fallen below 7.50 to 7.70 children per woman, 36 to 40% higher. All this suggests that relatively moderate Turkish fertility is a long term fact. It has certainly been so in Istanbul since the last quarter of the 19th century, where fertility of females born in the latter half of the 19th century did not

259

CHILD MORTALITY IN TURKEY

Table VIII. Completed parities by female birth cohorts (45–49 years) Cohort

Year

Parity

1913–18 1920–25 1925–30 1928–33 1930–35 1933–38 1938–43 1939–43 1943–48

1963 1970 1975 1978 1980 1983 1988 1989 1993

5.80 5.23 5.35 6.26 5.74 5.27 6.02 5.22 4.90

Sources: Censuses: 1970, 1975, 1980, 1985 Surveys: 1963, 1978, 1988,1989, 1993

exceed 4.4 children (Duben and Behar, 1991). Elsewhere in the Turkish Ottoman Empire there are also presumptions of moderate fertility, especially in the urban West of the country (Courbage and Fargues, 1997).

5.2. . . . B UT PERCEIVED

AS EXCESSIVE

This moderate average parity stemmed from a moderate, and especially marked lower, number of children wanted. Without stretching the comparison between two fairly heterogeneous measures too far, it transpires from the various surveys that the average desired family size is half the average number of live-born children. In other words, the pronatalist policy of the earlier republican governments and their successors up to 1965, elicited only a tenuous public response. Thereafter, government policy changed to come more in line with public attitudes and to accommodate the general lack of public support for population growth. In 1993, the refined Bongaarts indicator of the Wanted Total Fertility Rate (Bongaarts, 1990) gives an index of 1.87, or just 1.71 taking account of both birth rate and timing, which are not only lower than the already low effective period fertility: 2.65, but also below replacement level: 2.10 (Table IX). Never a high fertility society, Turkey perceived itself and still does, as over-fertile – with one child too many according to the gap between effective fertility and optimal fertility as regards child number and birth spacing (Unalan, 1993). Without quantifying the unmet needs in contraception, an immense need for fertility regulation can be hypothesized. In this respect, Syria presents a mirror image: a wanted fertility of 5.0 children, markedly above the effective fertility of 4.2. An excess of children on the far bank of the Euphrates, and a deficit on the near bank – can child mortality remain unaffected by these imbalances?

260

CEM BEHAR ET AL.

Table IX. Mean ideal number of children 1963–1993 Year

Mean number

1963 1968 1973 1978 1983 1988 1989 1993

3.19 (males: 3,86) 3.20 2.60 3.03 2.70 2.41 2.57 (males: 2.67) 1.87 (wanted total fertility rate, 1.71 planned total fertility rate)

Sources: Surveys from 1963 to 1993 and T. Unalan, Reproductive Expectations and Fertility Trends in Turkey, Demographic and Health Survey, Ankara, 1996.

5.3. T HE USUAL

REGULATING METHODS ARE NOT WORKING

Apart from breastfeeding of sufficient intensity and duration – not prevalent in Turkey where the average duration of breastfeeding is just 11.9 months, the ageold method of late marriage to regulate fertility, never found favour with Turkish society. The indicators reveal the marriage rate to be highly resistant to structural change and the vagaries of the economic or political situation. Turkey’s unrestrained westernization was checked in this respect. Between 1935 and 1975, the Bongaarts marriage index Cm shows no propensity to decline and has remained fairly stable since then, contrary to all expectations, at an average level of 0.75–0.79. Turks continue to marry very young – an average 19.4 years old for the generations born in 1943–1968. The singulate mean age at marriage for females (SMAM) has risen very little: from 21.00 years in 1978 to 22.39 years in 1993. The comparison with Turkey’s neighbors, especially the Arab countries where the measure is as high as 24 (Syria) or 27 years (Lebanon), is illuminating. Turkey, comparable in this connexion to Indonesia (Courbage, 1997) thus present a scenario of high, early nuptiality combined with a marked decline in fertility. Contraception could attenuate the failure of traditional methods – breastfeeding and postponement of marriage – of fertility control. Current contraceptive use – 66% in 1993 – tends to suggest that Turkey is on the path to fertility regulation. But contraceptive prevalence has leveled off since 1988 at 66%, and as the Islamic Republic of Iran shows, the limit is far from having been reached (80% of users in Shiraz, a city of 1 million inhabitants, compared to 71% in the highly urbanized western region which includes Istanbul in Turkey). More serious still is the extraordinary popularity of the so-called male methods (Table X). Male predominance in whether and how fertility should be regulated is a salient feature of reproduction in Turkey (Behar, 1995). Withdrawal was and remains by far the most popular method, protecting 43% of user couples, and withdrawal and

261

CHILD MORTALITY IN TURKEY

Table X. Contraceptive use (%) for married women and failure rates by method, 1993 Using

Failure rate

Pill IUD Vaginal methods Condom Female sterilization

4.9 18.8 1.3 6.6 2.9

9.7 1.0 24.6 11.7

MODERN

34.5

Withdrawal Periodic abstinence and other

26.2 1.9

TRADITIONAL

28.1

USERS

62.6

NON USERS

37.4

18.8

Source: Ministry of Health, Contraception, Abortion and Maternal Health Services in Turkey, Demographic and Health Survey, 1996.

condom combined, 52%. However, all the indications are that male monopolizing of reproductive control is not effective. The failure rates are 19% and 12%, respectively: one woman in five falling pregnant through use of the former method, and one in eight from the latter. But failure is far from being a deterrent. Two-thirds of women protected by withdrawal re-use the same method; the method itself is gaining in popularity, indeed, since very recently a fifth of those who used a modern method that had failed are now using withdrawal. One possible reason why first the Ottoman penal code of 1858, and then the Republican code of 1934, thought fit to prohibit abortion may be because of its extraordinary popularity. Both codes were adopted in times where there was serious concern and attempt to adopt western values and life styles. Abortion was not popularly regarded as a moral or religious sin (Gürsoy, 1996). In fact, the accounts of travelers and contemporary testimony of doctors can compensate for the lack of statistics. 19th century travellers ascribed the low fertility of Turkish Muslim – as opposed to Christian – women in the Ottoman Empire to their high propensity for abortion. The low birthrate can be attributed to the injurious methods used by lower class Turkish women to avoid multiple births. The British Consul also postulated abortion and the alarming prevalence of an unnatural practice as significant reasons for low population growth in the Turkish Muslim population under the Ottoman Empire (Senior, 1856; Duben and Behar, 1991) It was not

262

CEM BEHAR ET AL.

Table XI. Estimate of unwanted and non-aborted pregnancies % for 100 pregnancies I Total pregnancy rate of which foetal deaths of which still births of which induced abortions Total fertility rate Pregnancies yielding a desired child Pregnancies yielding an undesired child Total number of unwanted pregnancies % of pregnancies resulting in an undesired child on total number of unwanted pregnancies

3.75 0.19 0.06 0.85 2.65 1.71 0.94 1.79 52.5

II 3.75 0.19 0.06 0.85 2.65 1.87 0.78 1.63 47.9

I, Rates from DHS 93 survey, with correction, Planned total fertility rate. II, Rates from DHS 93 survey, with correction, Wanted total fertility rate. Sources: Minstry of Health Turkey, 1994 and Unalan,1993.

until half a century after the establishment of the Republic that the anti-abortion law was repealed entirely in response to public demand. The first Abortion Act passed in 1965 permitted termination of pregnancy only on health and “social” grounds. It was superseded by a very liberal Act in 1983, allowing termination of pregnancy on demand by the woman up to the 10th week of pregnancy. In practice, this time limit is flouted, since a quarter of abortions are performed after the 3rd month of pregnancy. Public hospitals perform 27% of abortions (private clinics 67%). It is significant in this respect that at no time did the Prime Minister of the short-lived Islamist government, Necmettin Erbakan (1996–1997), himself a committed populationist and avowed pronatalist, consider a return to the status quo ante in this matter. Abortion is now still widespread. In the two years prior to the 1993 survey, there were 18 abortions per 100 pregnancies or 25 per 100 live births – not too far removed from Western rates (the United States, 34 per 100). Appearances notwithstanding, however, abortion is held in check. In Istanbul, for instance, one woman in eight encountered family or other obstacles to abortion (Gürsoy, 1995). That abortion alone cannot satisfy the unmet need for family size control can be seen from Table XI. This reconstitutes the progress of 3.75 pregnancies from the moment they become observable (corresponding to the TFR for the period 1990– 95). They break down into 0.37 foetal deaths, 0.06 stillbirths and 0.67 induced abortions, corresponding to the total fertility rate of 2.65 children. Of these live births, 1.71 were wanted and 0.94 not. In other words, unwanted live births outnumber births averted by abortion. About 50% of unwanted pregnancies resulted in an unwanted live birth.

CHILD MORTALITY IN TURKEY

263

The very approximate nature of these estimates, however, led us to reparameterize them, mainly by reclassing half the foetal deaths as induced abortions. This is hypothesized by the authors of the report on the DHS 93 survey, who suggest that a substantial proportion of spontaneous abortions are in fact induced abortions. Ambiguity is compounded by the common use of the same word düsük for both. The word kürtaj (from the French curetage – curettage) is more recent and not in common use, except among the medical community. The Arab expression iskat-i cenin (“discarding the embryo”) strictly refers to an induced abortion but was dropped from the vocabulary after the campaign to cleanse the Turkish language (Gürsoy, 1996). We can also use the Wanted Total Fertility Rate rather than Planned Total Fertility Rate. This indicator considers as unwanted all children above ideal family size and all ill-timed, early births. The former measure is less restrictive and includes only children in excess of ideal family size, even if ill-timed. These calculations show that 50% of unwanted pregnancies result in unwanted children, the other 50% of unwanted pregnancies being aborted. Ineffective contraception, imperfect abortion compounded by the opposing effects of tradition which favours early marriage, and modernization, which favours early weaning, produce the astonishing result that on average, one in two children are unwanted births to parents who could not or would not seek abortion.

6. Planned neglect? At a pinch, a distinction could be made between “acceptable” and “unacceptable” deaths of the elderly and those in failing health. But informed sources in the Istanbul hospital community suggest that the idea is prevalent and applied to children in certain communities. How widespread is this in the region? Which children’s deaths would be “acceptable”? No firm answer can be given, since the subject is, by definition, not discussed. It can be hypothesized that unwanted children are most likely to be the victims of unintentional or even willful neglect by their parents because they are surplus to the planned family life cycle: children with birth defects, the wrong sex, generally female (Das Gupta, 1987; Muhuri and Preston, 1991) surplus and hence higher order, or ill-spaced. The idea that child mortality may be a response rather than a stimulus to high fertility is seldom mooted (Scrimshaw, 1978) because we take the value of children for granted, forgetting that in perhaps a few societies, desire for children is not extended. Some children are less valued than others, receive less medical attention, poorer nutrition and perhaps more importantly, parental affection; this vulnerability increases their death probability. A recent research in neuroscience illuminates the biochemical consequences of neglect. Children who are deprived of regular reassuring physical contact were found to have levels of normal stress hormones far higher than normal. These can lead to impaired growth and development of both the brain and the body. The study concluded that too many infants continue to be denied the affection that is vital to their development and that their lives are stunted

264

CEM BEHAR ET AL.

Table XII. Percentage of those who consider an only child as undesirable because of mortality risks – Turkey and other VOC survey countries

Females Males

Turkey

Philippines

Indonesia

Thailand

Taiwan

41 27

17 16

12 12

33 31

8 12

Source: Bulatao, 1979.

by neglect (Myntti, 1993). In this respect, Europe’s own poor – often glossed-over – historical record on childhood offers some perspective on the universal ideal of the special status of children: Inability to control births increased the number of unwanted children. And the hope that death would lighten the burden found fertile ground in the already significant levels of infant mortality (Ferry, 1996). There is, therefore, an easily “avoidable” (to use Scrimshaw’s phrase) excess infant and child mortality ranging from parental ignorance to willful neglect. Is it quantifiable? Probably not to any degree of accuracy, but having regard to Turkey’s development and cultural level indicators, and excess female mortality, it is probably fairly safe to say that its infant mortality is double the norm (Tables 3 and 4). Taking illiteracy for example (other indicators give similar orders of magnitude), we find that with just 27.6% female illiteracy, Turkey’s expected infant mortality should be 28.7 and not almost double that at 53 per 1000. The Value of Children surveys clearly show that Turkish parents in the Seventies were far from unaware of the high probability of child death. To the question “why do you want not less than N children?”, about 15% of parents, men and women together, replied ‘fear of losing a child’ (Kagittcibasi, 1982). The idea of only children was rejected precisely because of the death probability (Table XII). Turkish parents were aware of the risks which high infant mortality posed for their children. Mutually self-reinforcing ignorance and social habitus can thus be discounted as explanations for high mortality. That leaves neglect. There are grounds for the presumption. Under-immunization in such a highly urbanized country with nationwide medical coverage places the blame with the family rather than State provision. But non-immunization is not an automatic death warrant. Excess female mortality offers a first source of easily avoidable deaths. Without aspiring to Scandinavian levels – which may not be achievable – simply equaling those of the Arab Middle East would reduce child mortality by three points to below the 60 per 1000 level. But all the indications are that it is not just girls who are discriminated against – far from it. High child mortality, high birth orders and short birth intervals are usually attributed to biological rather than behavioural factors. It is a fine distinction to draw. High parity children are often born to older mothers, worn out by successive births, increasing the death probability of the current child (Scrimshaw, 1976).

CHILD MORTALITY IN TURKEY

265

A comparison can establish what can be attributed to biological factors in Turkish child mortality, rather than behaviour – selective parental neglect of selected children. In highly pronatalist Syria where there are no a priori grounds for suspecting that mortality is used as a fertility regulating method, we looked at the birth orderor birth interval-specific effect on child mortality. The biological effect is clear: moving from orders 2–3 to 4 and above increases infant mortality by 21% and child mortality by 41%. In Turkey, the increases are almost double: 45% and 73%. The intensity of the relative vulnerability of higher-order children is much greater. The variations are even clearer with birth intervals: the reduction from an interval of 4 years or more to one of less than 2 years produces a threefold rise in Turkish infant mortality and a sixfold rise in child mortality. So sharp are the contrasts between the two neighbors, that it is tempting to ascribe the increased mortality of ill-spaced or excess children in Turkey partly to biology but partly also to extremely strict planning behaviour. In short, the auspices at birth are not good for a significant number of children. The home environment may not be systematically conditioned to accept all new arrivals willingly, particularly where children are valued for necessarily uncertain economic or utilitarian reasons and far less for the gratification they bring in terms of love and affection.

7. Children: Objects of affection or useful objects? The Value of Children surveys carried out in various countries – The Philippines, Thailand, Korea, Indonesia, Taiwan, the United States, West Germany and Turkey – in the late 1970s brought to light a specific peculiarity of the Turkish attitude to childhood (Bulatao, 1979). The surveys were based on the Caldwell model of the reversal of the net intergenerational wealth flow as a country becomes modernized. Their fundamental hypothesis was that the costs and benefits of children were reversed during demographic transition. These surveys sought to analyze the socio-psychological mechanisms of the more or less pronatalist attitudes in these countries. They included a standard questionnaire and were conducted in approximately the same way in all the countries concerned. It transpired that Turkey was an exception within a set of countries whose results fit the basic premise fairly well. Turkey’s results did not conform to the assumption that the material benefits expected of children by their parents would diminish as fertility declined or the standard of living rose. A mis-match was observed between the (fairly small) desired family size and the already declining effective fertility on the one hand, and the still very high short- or long-term material and economic advantages expected by parents of their children. Hence the percentage of positive replies given by parents to the question on the child’s expected economic contribution (Table XIII).

266

CEM BEHAR ET AL.

Table XIII. Indicators of the economic significance of children VOC countries 1975–1976 Indicator

Turkey

Philippines

Indonesia

Thailand

Korea

65 93 77 95 87

83 86 67 88 84

81 82 60 83 81

73 89 71 92 86

82 85 71 87 85

Expect help from their son (in %) In the household In old age By their salary In case of emergency Help for schooling of siblings

Source: Bulatao, 1979. Note: VOC countries are Value of Children Survey countries.

More relevantly, Turkey should have been compared with the countries of the Arab Middle East and Iran, or those of the Balkans and Transcaucasia, already mentioned by way of comparison, because the other four countries surveyed fall within a very different culture area. But these countries did not ask for such surveys to be conducted. Turkey yields the highest rate of positive replies to all the questions to parents on the material satisfaction expected from their children. What these surveys called “the economic role of children” proved to be considerably higher among Turkish respondents than everywhere else. Of all these countries, the instrumentalisation of children seems to have been most deliberate and taken furthest in Turkey. From Indonesia (then the country with the highest fertility) to Germany, the progressive alignment of attitudes tends to confirm the underlying hypothesis on the cost and value of children, except in Turkey which stands out as a spike in the graphs; the utility value of the child is higher then than that prevailing in higher fertility countries. In “Ariesan” terms (from the French historian P. Ariès) (Ariès, 1960) the revolution of childhood had not yet taken place in the Turkey of the 1970s, contrary to what might be expected from the general economic and demographic indices. Children seem to have been wanted less for themselves than for the present and future welfare of their parents. The parents, therefore, and not the child, were the economic priority. “The degree of material and financial benefit expected [of the child] declines uniformly between high and average fertility countries, apart from in Turkey”. The instrumental value placed on the child can be directly seen in certain results. The replies to the open-ended question put to parents as to the reasons for wanting an additional child were as shown (Table XIV). More than a third of parents said their primary concern was the economic benefit that an additional child would bring to the family. Over half (55%) ranked the child’s economic benefit first or second. The social, normative or emotional value of the child also mentioned in the answers occurs with much less frequency than its short and long-

267

CHILD MORTALITY IN TURKEY

Table XIV. Main reason for having one more child (in%) Reason

Males

Females

Affection Financial help for the household To have another son Companion for their siblings Help in household work Support in old age

4.9 12.9

5.7 12.8

12.1 3.2 2.9 14.3

15.5 3.6 1.2 22.0

1.8 7.4 9.4 3.6 6.3 4.6

2.4 4.0 17.7 5.7 22.2 3.8

Help in household work Pleasure of raising a child Consolidate couple To have another girl Family name continuation For the sake of pleasure Source: Kagitcibasi, 1982.

term economic value. In Turkey, therefore, children seem to have been most valued for their material contribution to family life, both in their own youth and in their parents’ old age. This reflects “a utilitarian attitude” which makes heavy demands on children to assist and support their parents. Correlating these attitudes with the economic activity of children and young people in Turkey compared to those of neighbouring or similar countries (ILO, 1995), we find that the utilitarian attitude of parents towards their children in Turkey has its real-life counterpart in a higher economic activity rate among young people compared with neighbouring countries.

8. Tradition and culture: Childhood and death in ethnographic and other sources 8.1. VALIDITY,

SIGNIFICANCE AND MEANING

Obviously, we do not seek to claim that such materials can replace the usual mortality data. However, statistical data are not infrequently lacking, or not sufficient to ground a conclusively unambiguous causal link. Qualitative sources and materials (different ethnographic and folklore materials, literary and other works, radio and television broadcasts, films, etc.) may be used to help explore the meanings behind the exceptional child mortality rates. These non-demographic sources could be regarded as a sort of reflection of the collective unconscious and certain cultural constants. They could serve as a kind of “pointer to the way people think” or

268

CEM BEHAR ET AL.

“index of general cultural attitudes”. As a best-case scenario, they can provide circumstantial evidence; at the worst, they offer simple ways forward to advance or support a particular hypothesis on mortality. Such sources can obviously only be used on a case-by-case basis and with caution, without trying to work them up into norms, direct and unequivocal causal links or generalizations, nor, obviously, to read into them more than they can mean. Because they can claim neither validity in the statistical sense, nor completeness in time or space. Also, as “non-quantitative” materials, they are by their nature inevitably selective at source, more or less arbitrarily filtered, variable in time. We are aware that, as in research that aims for quantitative analysis, qualitative research such as ethnographic description has issues related to bias and scientific objectivity. When we employ literary and artistic works and expressions to probe into child mortality we are entering the realm of symbolic representation. The frequent recurrence of a theme, attitude or specific fact (or non-recurrence where it might be expected) may be an indicator of its relative importance and place within a social value system. By contrast, it may also be that a single folk custom concerning childhood or death, or a single truly striking image or literary theme may with hindsight assume a symbolic value of general and lasting significance. It is these two aspects (diffusion and singularity) that we shall strive to highlight in our examination of materials on the values, attitudes and behaviours with respect to child mortality. The usable sources are so plentiful that we had too much to choose from. We therefore turned first to the abundant ethnographic and folk literature on Turkey, focusing on how it reflected the relation between children and death. We also used a number of detailed surveys of the habits and customs of rural Anatolian communities who make up nearly 45% of the population. We also considered attitudes and values relating to childhood as relayed through Turkish film, as the chief expression of urban popular culture, especially from 1950–1980. Lastly, we dipped into the literature, especially contemporary drama as the primary vehicle for modernist thought.

8.2. E THNOGRAPHIC

SOURCES : SILENCE OF THE LAMBS ?

The most significant systematic collection of ethnographic material on attitudes towards death, burial practices and funerary rites in Anatolia is that of Sedat Veyis Örnek (Örnek, 1971), which lists and classifies in minute detail all Turkish practices and beliefs concerning, inter alia, premonitions of death, the preparation of the body, burial, the burial ceremony. It deals comprehensively with coffins, cemeteries and tombstones, mourning and post-funeral etiquette, weeping, funeral vigils, funeral wakes, condolences, oral literature on death etc. It is astonishingly thorough and detailed. It covers virtually all Asian Turkey. Interestingly, neither the researcher, nor the hundreds of local sources he consulted knew of any sources of material culture related to infant and child death and

CHILD MORTALITY IN TURKEY

269

funeral practices. Social habituation to death in early childhood? Most probably. However, he has been able to identify numerous agits (standard lamentations), mostly chanted rhythmically, that strikingly reflect the mothers pain and protest over the death of her child. Örnek gives a record of about 60 agits in his book on child folklore where he is able to cite about 90 ninni (sleep chants). The history of childhood is obviously not so straightforward. The emotional impact is obviously there, but there is a fixed hierarchy of expression and externalization. The type of mourning is dictated by the age, sex and status of the deceased. Women are wept for less than men, children less than young people and adults, and common people less than individuals of greater renown or higher social status (Örnek, 1977). Thus, the death of infants and children seem to bring an outcry mostly from mothers and rather than finding material funerary customs and objects related to child death, we find standard emotional expressions in the form of ritualized laments. Also, because death in childhood is so common, excessive outpourings of public grief are frowned upon. There are many popular prohibitions to this effect (Örnek, 1979; Tan, 1993). For example, for a mother to publicly express her grief at the loss of her baby or infant is considered in some areas as blatant disrespect for her own parents and parents-in-law. Even less so should she consider herself as in mourning. Public grieving is quite simply sinful. Excessive outpourings of grief would also prevent the child’s (pure and innocent) soul from interceding in favour of its parents in the hereafter, a sort of instrumentalisation of the child after death (Delaney, 1995). Unlike rural society in some South American countries, there is in Turkey no joyful celebration of an infant’s death, or public rejoicing at the death of an infant (apotheosized into a cherubim or angelito) organized by the father in place of a funeral ceremony (Scrimshaw, 1976). Even so, everything is designed to encourage habituation and minimize the emotional, social and family impact of death in childhood. Another sign of social habituation to the death of a child is the bestowing of an umbilical or temporary name at birth. Popular belief demands that should a child not survive, it cannot leave this world unnamed. The child’s true name, that with which it will be officially registered and will bear thereafter, is bestowed only some (seven?) days after birth (Nicolas, 1972). The temporary name is usually a typically Muslim name, because everyone will be called by name at the last trump. Temporary naming is a fairly widespread practice, and not confined to the countryside. It wards off the (omnipresent and very high) danger that the child may depart this world unnamed, uninvested by a semblance of social and religious identity. Since nothing can be done to ensure the child’s physical survival, its soul is immediately tended to. But what of stillborns? There is only silence. Some acts proscribed to pregnant women are clearly motivated by similar considerations, since they intimately link pregnancy and death, as well as birth and death. In some parts of Anatolia, pregnant women are exempt from certain social customs connected with a death in the family or immediate circle. They are dispensed from making condolence calls. They do not have to watch the dead or dying.

270

CEM BEHAR ET AL.

They do not have to receive the recently bereaved (Nicolas, 1972). Even less must they be allowed to view a corpse. Likewise, under no circumstances must they have any contact with the water used to wash a corpse. The future new-born baby is thus symbolically protected during pregnancy from the danger of death which awaits it after birth. Frequent child death has also given rise in rural communities to certain preservation rites intended to secure a long life for the newborn child. They are mainly practiced by families previously bereaved of an infant (Acipayamili, 1974). They are chiefly magic rites, found in similar forms in other cultures, designed to shield the newborn child from natural and supernatural dangers. The fictitious “sale” of the child to a third person, breastfeeding at least once by a mother with living infants, and the bestowing of particularly protective personal names, are the most current preservation rites. The making of a special item of patchwork clothing made of seven or sometimes forty (both fateful figures) pieces of cloth is very prevalent. Each piece must be borrowed from a family which is not in mourning and has at least one living child. The child must then wear these magic clothes for a specified (fateful, magic) number of days or years (Tansug, 1993). These many and varied rites and practices do not seem to be sex-differential. A practice reported in the Bergama region, on the Aegean coast close to Izmir, suggests a form of infanticide by omission – presumably more imaginary (but conceivable, however, which is what counts) than real. According to the reports (given at third-hand by the ethnologist who had never observed it with his own eyes nor had it reported by anyone who had), if a child was perceived to be stricken with a birth defect, the umbilical cord would not be immediately cut with a wire, as is customary in rural society, but left to cause a fatal hemorrhage in the new-born child (Nicolas, 1972). Such a procedure might not be entirely without “effectiveness” medically speaking, especially when it is used with sickly or premature new-borns, where a delay in cutting the umbilical cord may cause fatal anaemia. The dangers that this type of practice may have for the child are known in some Central American countries (Scrimshaw, 1976). Infanticide is a universal social phenomenon, also of contemporary significance for industrialized countries. Interestingly, most judicial decisions are based on the age of the infant and child. It may well not have been widespread in Turkey, have died out centuries ago, have survived not in fact but only as anonymous folklore. It may be restricted to a small geographical area. But the fact that there is or was a practice of not cutting the umbilical cord could indicate that a Spartan-type eugenist infanticide was, in that specific area, at the very least thinkable. The fact that the hadith emphatically refer to a prohibition on female infanticide also points to the cultural reality that this practice was part of the collective sub-conscious of the cultural and geographical region.

CHILD MORTALITY IN TURKEY

8.3. C HILDREN ,

271

FAMILIES AND THEIR DEATH IN FILMS

The Turkish film industry is prolific – the third or fourth largest in the world – in terms of output (about 300 at its peak in 1972). It is therefore nigh-impossible to view even a selection and give a general assessment of the image of childhood conveyed by the popular entertainment par excellence which is the cinema in Turkey. Detective/police films, westerns, costume and real-life drama, melodrama, light comedies, science fiction, local remakes of American films – the Turkish cinema has it all. It has obviously gone through several phases and waves, and under very recently catered often for a predominantly male audience. Our particular focus is on films featuring family life. 8.3.1. All in the family Almost all Turkish films are familist. Family values are sacred. The family as an institution is always valued, always comes first. The family order often epitomizes the social order. The family institution is the repository of many religious and moral values, filmically reinforced. Family unity and solidarity must be preserved at all costs. If a family is divided, its remaining members must cleave to one another. Any sacrifice for unity’s sake, including death, is unfailingly approved and valued (Abisel, 1994). It is almost impossible to find a film critical of family institutions. Criticism is reserved exclusively for rich, bourgeois, “degenerate” families (soulless and immoral). Wealth, success, fortune and degeneration of family values are linked together in a somewhat naive tunnel vision. The ideal family is lower middle or working class, small, self-contained, self-reliant and honest (Abisel, 1994). Once in the big city, some family members’ aspirations for personal success and happiness begin to create problems. Their ambitions tend to clash with traditional family solidarity, and are not highly valued. Many films deal with this type of family conflict and have it as their dramatic core. The constraints of city life thwart personal ambition and threaten the survival of the family unit. Money and a different life are the downfall of those that prize them too highly. Personal ambitions often destroy the family spirit, and personal or collective honor, lead to the moral degeneration of the individual and the family, failure, and return to the native village, or death. The family must withdraw into itself to protect and keep intact its traditional values (honour, morality, mutual assistance, solidarity, etc.) even though they are no longer effective. Film-makers have played countless variations on this dramatic theme. The family unit is greater than the sum of its individual members. 8.3.2. Family structures In Turkish films, families which migrate to city are typically or often nuclear families (Guchan, 1994). This is cinematographic telescoping: whether nucleation is willed or suffered is not completely clear. Where a rural migrant family includes a grandmother, it is always because there is no male head (dead, in prison, working

272

CEM BEHAR ET AL.

in Germany, etc.). The extended family is there for pure form’s sake, because traditional intra-family relations (division of labour, solidarity, abnegation, division of authority etc.) are different in the city. The womenfolk gather together and share the same space out of pure necessity. Nevertheless, Turkish films show in astonishing variety the interplays of migration, changes and continuities in family life and the outcome these social phenomena have on child survival and well being. The film Kurban, for example focuses on the eventual (avoidable) death of a child with heart disease. When the extended family migrate to the city, the grandparents and the uncle of the child are much more interested in investing in opening a market than in investing in his medical needs. The child in the film symbolically dies on the day of the sacrificial religious festivity. As the child has a heart seizure, he releases the sheep intended for sacrifice, and dies instead. Striving for economic gains have triumphed over attention to child survival. Against the will and the outcry of the mother, the dynamics of the urban extended family dictates that collective economic gains of the family supersede individual health investment for the child. Death of children in Turkish movies almost always seems to carry the symbolism of contested social values and aspirations. Death by rabies of the son of the landlord in the film Kuduz (Tarik Akan) – based on a true story in a remote Turkish village in the 1940’s – symbolizes the dangers of war for Turkey during the second world war. The child already suffers neglect and abuse, as a result of a split marriage where he is not allowed to see his mother. The film focuses on the lives of different children in the village, to show how solidarity, communal values as opposed to individual avarice and selfishness of parents (particularly fathers) lead to child happiness and survival or death. 8.3.3. Now you see them, now you don’t Children are often present, even and especially by their absence. Very young children feature very little in films. New-born babies and infants always fulfill a “functional” purpose to symbolize a happy marriage, accomplished motherhood, or as the object of female rivalry. In these films, childlessness is a tragedy for a married woman. The child is the supreme being for the mother (Abisel, 1994). The lack of a child is pregnant with danger for the woman and a harbinger of marital breakdown. Birth binds the husband to the home. Here again, film-makers have exploited every possible variation on the theme. Countless films have been made about the misfortunes of childless married women. Plays on the theme have been adapted for the screen. Pregnancy often marks the end of a woman’s family problems. Likewise, the prospect of the child’s death, serious illness or even abduction places the newlyweds’ relationship under great strain. The classic film Kirik Hayatlar (Broken Lives) by Halit Refig, based on the literary work of Halit Ziya Usakligil, is centered around family relations and the death of a child. This loss affects the whole substance of the inter-family dynamics. Interestingly, the director says that despite the political and economic changes in Turkey which happened

CHILD MORTALITY IN TURKEY

273

since 1900 (when the novel was written) and 1965 (the film), there has been no change in the economy-related drama of individual people of middle class. The “family” films most popular with the Turkish public have been those in which a child is the main character. These “children” are the deus ex machina that prevent family breakup, reunite a dispersed family, or create a new family for themselves. They are melodramatic “weepies”, churned out by the dozen if not hundred. The bright and mischievous “child” (girl or boy) is the family provider, looks after the mother and siblings, and is virtually a surrogate head of household. They outwit the adult world, are street philosophers and streetwise. Well-worn cliché is piled on pure schmaltz. Tear-jerking happy endings (the family reunited and kept together, reunion with the real father and/or mother) are the invariable rule. It is no surprise that in the prize-winning film Berdel where the mother dies in childbirth, her five girl children of ages up to 15 years old, act sophisticated characters, demonstrating varied reactions to the demand of the father of wanting a son, at whatever cost and thus discriminating against his daughters.

8.4. T HE CHILD

IN MODERN DRAMA : SPOILSPORT OR SCAPEGOAT ?

The themes of childhood and death are often juxtaposed in modern Turkish drama. A product of the late 19th century, it is influenced by the realism, epic drama and naturalism of Strindberg, Ibsen, Chekhov, Brecht and others, and, light comedies aside, it is almost always directed towards social criticism. A study of Turkish 20th century plays underlines this linkage of childhood and death (Sener, 1993). References to childhood abound. As with films, however, plays in which children are directly seen on stage or involved in the action are thin on the ground. Their presence is only ever indirect, when they are invariably the spring or centre of tragedy and pathos of the dramatic action. Children are seldom presented as the product or symbol of a happy union, as having always been unqualifiedly wanted, or as a source of joy for the parents or family. Many plays centre on the problems created by the birth of a child. Childhood is often presented as a period of misunderstanding, exclusion or woe. It is the source of tragedies to come, the barely concealed origin of trials and tribulations. Children tend to be associated with negative values, despair, powerlessness, serious mistakes, destitution and misery, suffering and death. The causes of a troubled childhood are either poverty or uncomprehending parents and adults, generally both together. The dramatic framework of a disturbed childhood affords playwrights a useful backdrop for critiques of urban or rural society. Even a comparatively happy, carefree childhood is contrasted with present adversity as a source of sadness, regret and nostalgia. Direct exploitation of children – forced labour, domestic servants, etc., – by his/her parents or society as a whole is also foregrounded. Only a very tiny minority of plays give children or adolescents roles as dramatic characters, or portray childhood as a serene and happy time. Death, or some other tragedy, often serves as a rite of passage for the

274

CEM BEHAR ET AL.

child into adulthood. And, however abrupt, this passage to adulthood elicits a sort of feeling of relief, of having at last assumed their place in the “real” life of society. In short, childhood is a painful and marginal period, but one which has to be got through, preferably as quickly as possible. Childhood is merely the anteroom to adult paradise, a purgatory before real life, the life after childhood. So there is no point in intellectualizing it, or even portraying it correctly. The child in modern Turkish drama is also sometimes used to symbolize the prototypical sacrificial rite, as shown in ‘The Sacrifice’, a play by Dilmen (1967). The scene is a traditional rural community. A mother kills her two children, a boy of nine and a girl of four, then commits suicide. She does not want her children growing up in a world devoid of trust. The father intends to take another wife. The sacrificial theme is introduced from the outset, with the father telling his children the biblical legend of Abraham. The children are horrified at the thought that the lamb they have fed and tended will be sacrificed. But the mother will free the sacrificial lamb, and the children will be sacrificed in its place. The legend of Abraham is reversed. The theme is adapted to contemporary (especially rural) society. The traditional theme of the unheeded or sacrificed child (the innocent victim of adult discord or misfortune) recurs frequently in various guises. The child is sacrificed (or rather falls victim) to traditions, archaic customs and practices, passions of many kinds, and even the adults’ sense of duty. Sometimes, that sacrifice involves the physical death of the child. And when the child itself is not the ultimate sacrifice, that lot falls to an adult, to peace, or to happiness.

9. Overview Turkey is demographically unclassifiable, and especially so in terms of infant mortality. Be it in terms of geographical or geopolitical situation, range of political structures, or economic development criteria, international organizations have always been hard put to categorize it. In every comparative table used in this study, moreover, Turkey is prominently the odd man out. Why has Turkey’s child mortality rate been and remained so high? Couched in those terms, there is probably no answer, or at least no single answer, and certainly no strictly demographic answer, to the question. As we have seen, family traditions and on occasion political opportunism, mean that children have not always been valued in the Ottoman, then Turkish, tradition. In the court at Istanbul, it was not uncommon to have the youngest scions dispatched to avert dynastic feuding. By this Fratricide Law, ordained for “the Good of the State”, with the approval of a majority of ulamas (council of learned men) the sovereign effectively reinforced his own grasp on power, and the rights of his own offspring, (Veinstein, 1989), a situation more or less parallelled at the Chinese imperial court (Lee, Feng and Campbell, 1994). Later on, the fervently populationist republican leadership failed to set an example by their particularly small families. Turkey’s rapid modernization in the

CHILD MORTALITY IN TURKEY

275

space of three decades, which might have been assumed to be more favourable to the child’s lot, has disappointed expectations. It may be due to the stress created by economic growth. This can be perceived through the exceptional female labour force participation statistics – women account for 35% of the workforce, increasingly in the industrial and service sectors with their work organization constraints. Labour force participation of women of childbearing age was above 48% at peak reproductive ages in the 1990 census, 2 to 3 times that of Turkey’s southern neighbours. This means that women’s work responsibilities may impinge on their childcare responsibilities, particularly as many women workers are not in the formal sector and hence not covered by social protection. But, neither these demographic or social data, nor the conclusions that can be drawn from them, are enough to offer a wholly satisfactory explanation founded on concrete, quantified proofs. Our reference to nonconventional sources in search of specific cultural configurations which connect childhood and death were not intended to replace hard data, but have to some extent, afforded an indication of the cultural atmosphere in which mothers, fathers and families of deceased children are immersed. What can be said for certain, however, is that unduly slowly changing attitudes towards childhood have not kept pace with Turkey’s rapid, and almost completed, fertility transition. Here again, Turkey is an exception. An infant mortality rate of 53 per 1000 with a fertility barely above 2 is a combination encountered nowhere else. In fact, the childhood revolution has not yet taken place, and has therefore neither preceded nor accompanied the very rapid fertility decline. There are not fewer children because they have more emotional value. On the contrary, it would seem that children are acquiring more value precisely because there are increasingly fewer of them. The expected change in attitudes and mentalities towards children has not preceded or accompanied the process of demographic transition. It must follow it. Turkey’s fertility transition will be completed long before its mortality transition, because with a period fertility 20% above the replacement rate, child and infant mortality rates are almost 10 times those of the most developed countries. So there remains a longer way to go. The length of the lag between the two transitions cannot be predicted. In the immediate future, Turkey will enter a post-transitional situation. The leeway for progress remains the very wide regional variation in rates, including mortality, and the extreme heterogeneity of the country, with the most abrasive modernism and most recalcitrant traditionalism living side by side. The gap between mortality and other demographic indicators will tend to close. References Abi¸sel, N., 1994. Türk Sinemasì Üzerine Yazilar (Writings on Turkish Cinema). Ankara, ed. Imge. Aries, Ph., 1960. L’enfant et la vie familiale sous l’ancien régime. Plon, Paris. Al Saadi, A. F., 1996. ‘Infant mortality and the economic embargo in Iraq?’, Population Bulletin of ESCWA 44.

276

CEM BEHAR ET AL.

Acipayamli, O., 1974. Türkiye’de Dogumla ilgili Adet ve inanmalarìn Etnolojik Etüdü (An ethnological study on the customs and beliefs concerning childbirth in Turkey). Ankara, Atatürk Üniversitesi Yayìnlarì. Agha, H., Chasteland J-C., Courbage, Y., Ladier-Fouladi, M. and Mehryar, A., 1997. Famille et fécondité à Shiraz – 1996. Dossiers et Recherches N◦ 60, INED, Paris. Aksit, B. and Aksit, B., 1989. ‘Sociocultural determinants of infant and child mortality in Turkey?’ Social Science and Medicine 28: 571–576. Barbieri, M., Blum, A., Dolgikh, E. and Ergashev, A., 1996. ‘Nuptiality, fertility, use of contraception and family policies in Uzbekistan’. Population Studies 1: 69–88. Behar, C., 1975. ‘Les tables de mortalité de la Turquie’, in L’Analyse démographique et ses applications, Ed. du CNRS, Paris. Behar, C., 1995. ‘The fertility transition in Turkey – Reforms, policies, and family structure’, in Carla Makhlouf Obermeyer (ed.), Family, Gender and Population in the Middle East, Policies in Context. American University in Cairo Press, Cairo. Behar, C., 1997. ‘La Turquie: Une transition démographique presque achevée’, in Jean-Claude Chasteland and Jean-Claude Chesnais (eds.), La Population du Monde – Enjeux et Problèmes. Paris, INED-PUF, 331–343. Boisvert, M. and Meyer, F., 1994. ‘Mortalité infantile et consanguinité dans une population endogame au Québec’. Population 3. Bongaarts, J., 1990. ‘The measurement of wanted fertility’. Population and Development Review 3. Boukhaïma, S., 1997. ‘Recomposition du système de santé’. La Syrie de l’après guerre froide: permanence et changements, Maghreb-Machrek 158: 38–46. Bulatao, R., 1979. On the Nature of the Transition in the Value of Children. East-West Population Institute, Honolulu. Bulut, A., Gokcay, G., Neyzi, O. and Shorter, F., 1991. ‘Perinatal, neonatal and under-5 mortality in Istanbul based on representative samples of burial Records’, in Seminar on Measurement of Maternal and Child Mortality. Population Council, Cairo. Caldwell, J., 1979. ‘Education as a factor of mortality decline: an examination of Nigerian data’. Population Studies 33: 395–413. Caldwell, J., 1986. ‘Routes to low mortality in poor countries’. Population and Development Review 2: 171–220. Coale, A. and Watkins, S., 1986. The Decline of Fertility in Europe. Princeton University Press, Princeton. Courbage, Y., 1997. ‘L’Indonésie: Une transition presque achevée dans le plus grand pays d’islam’, in J-C. Chasteland and J-C. Chesnais (eds), La population du monde – Enjeux et problèmes. Paris, INED. Courbage, Y. and Fargues, Ph., 1997. Christians and Jews under Islam. Tauris, London. Courbage, Y., 1999. ‘Political and economic issues of fertility transition in the Arab world – Answers and open questions’, Population and Environment 4. Das Gupta, M., 1987. ‘Selective discrimination against female children in India’. Population and Development Review 1: 77–100. Delaney, C., 1995. The Seed and the Soil. University of Carolina Press. Desgrees du Lou, A. and Pison, G., 1995. ‘Le rôle des vaccinations dans la baisse de la mortalité des enfants au Sénégal’. Population 3: 591–620. DHS and Hacettepe University Institute of Population Studies, Ankara, 1994. Fertility trends, women’s status and reproductive expectations in Turkey. Dorsay, A., 1989. Sinemamìzìn umut yìllarì (Years of hope of our cinema). Istanbul, ed. Inkìlap. Duben, A. and Behar, C., 1991. Istanbul Households – Marriage, Family and Fertility – 1880–1940. Cambridge University Press, Cambridge. Farah, A. and Preston, S., 1992. ‘Child mortality differentials in Sudan’. Population and Development Review 8: 365–383.

CHILD MORTALITY IN TURKEY

277

Ferry, L., 1996. L’homme-Dieu ou le sens de la vie. Grasset, Paris. Gokalp, A., 1994. ‘Mariage de parents: entre l’échange généralisé et le mariage parallèle – le cas de la Turquie’, in P. Bonte (ed), dans Epouser au plus proche: Inceste, prohibitions et stratégies matrimoniales autour de la Méditerranée.Editions de l’EHESS, Paris, pp. 439–453. Guçhan, G., 1992. Toplumsal Degisme ve Türk Sinemasì (Social change and Turkish cinema). Ankara, ed. Imge. Gürsoy, A., 1992. ‘Infant mortality: a Turkish puzzle’. Health Transition Review 2/2: 131–149. Gürsoy, A., 1996. ‘Abortion in Turkey: A matter of state, family, or individual decision’. Social Sciences Medical Review 42(4): 531–542. Gürsoy, A., 1995. ‘Child mortality and the changing discourse on childhood in Turkey’, Children in the Muslim Middle East. ed. E.W. Fernea, UT Press, 199–221. Hill, K. and Upchurch, D., 1995. ‘Gender differences in child health: Evidence from the demographic and health surveys’. Population and Development Review 21: 127–151. Institut National D’Etudes Démographiques, 1993/1997. ‘Tous les Pays du Monde 1997’. Populations et Sociétés 282 and 326. International Labour Organization, 1995. Yearbook of Labour Statistics 1995. Geneva. Kagitcibasi, C., 1982. The changing value of children in Turkey. East-West Population Institute, Honolulu. Ladier-Fouladi, M., 1997. ‘The fertility transition in Iran’. Population, An English Selection 9: 191– 214. Lee, J., Feng, W. and Campbell, C., 1994. ‘Infant and child mortality among the Qing nobility: Implications for two types of positive check’. Population Studies 48: 395–412. Longone, P., 1971. La mortinatalité. Populations et Sociétés 34. INED, Paris. Martorell, R. and Ho, J., 1984. ‘Malnutrition, morbidity and mortality’, in W. H. Mosley and L. C. Chen (ed), Child Survival: Strategies for Research. the Population Council, New York, 49–68. Ministry of Health and Hacettepe University Institute of Population Studies, 1994. Demographic and Health Survey – Turkey – 1993. Ankara. Mosley, W. H. and Chen, L. C. (ed.), 1984. Child Survival: Strategies for Research. the Population Council, New York, 49–68. Muhuri, P. and Preston, S., 1991. ‘Mortality differentials by sex in Bangladesh’. Population and Development Review 3: 415–434. Myntti, C., 1993. The social determinants of child mortality and mortality in Yemen: Insights from social science theory. Paper presented at the IUUSP 22nd General Population Conference, Montreal, Canada. Nicolas, M., 1972. Croyances et pratiques populaires turques concernant les naissances. Publications Orientalistes de France, Paris. Örnek, S. V., 1971. Anadolu Folklorunda Ölüm (Death in Anatolian folklore). Ankara Üniversitesi, Ankara. Örnek, S. V., 1979. Geleneksel Kültürümüzde Çocuk (The child in our traditional culture). Is Bankasì Yayìnlarì, Ankara. Örnek, S. V., 1977. Turk Halkbilimi (Turkish ethnography). Ankara. Özön, N., 1968. Türk Sinemasì Kronolojisi (Chronology of Turkish cinema). ed. Bilgi, Ankara. Papchild/Central Bureau of Statistics, 1995. Mas’h al -joumhouriat al ‘arabiya es souriya haoula sohhat al um oua al toufl, Al takrir al raissi (Syrian Arab Republic Survey on Mother and Child Health – Main report). Damascus/ Cairo. Pinker, S., 1997. ‘Why they kill their newborns’, The New York Times, November 2, 1997, 50–54. Pritchett, L., 1994. ‘Desired fertility and the impact of population policies’. Population and Development Review 1: 1–55. Schepher-Hughes, N., 1985. ‘Culture, scarcity and maternal thinking: Maternal detachment and infant survival in a brazilian shantytown’. Ethos 13(4): 291–317. Scognamillo, G., 1988. Türk Sinema Tarihi (History of Turkish cinema). ed. Metis, Istanbul.

278

CEM BEHAR ET AL.

Scrimshaw, S., 1978. ‘Infant mortality and behaviour in the regulation of family size’. Population and Development Review 4/3: 383–403. Senser, S., 1993. ‘Türk Tiyatrosunda Çocuk’ (The child in Turkish theatre), dans Toplumsal Tarihte Çocuk (The child in social history). Tarih Vakfì Yurt Yayìnlari, Istanbul, 92–101. Senior, N., 1856. A Journal Kept in Turkey and Greece. London. Shorter, E., 1975. The Making of the Modern Family. Basic Books, New York. Shorter, F., 1989. The Decline of Infant and Child Mortality: Estimates from the Turkish Census. The Population Council, Cairo. Shorter, F. and Macura, M., 1983. Trends in Fertility and Mortality in Turkey: 1935–1975. National Academy Press, Washington. Sokal, A. and Bricmont, J., 1997. Impostures Intellectuelles. Odile Jacob, Paris. Sommerfelt, A. and Piani, A., 1997. Childhood Immunization: 1990–1994. Demographic and Health Survey, Calverton. Tan, M., 1993. ‘Çocukluk: Dün ve Bugün’ (Childhood yesterday and today), in Toplumsal Tarihte Çocuk (The child in social history). Tarih Vakfì Yurt Yayìnlarì, Istanbul, 11–31. Tansug, S., 1993. ‘Anadolu’da Çocuk Giyimi’ (Children garments in Anatolia), in Toplumsal Tarihte Çocuk (Child in social history). Tarih Vakfì Yurt Yayìnlarì, Istanbul, 127–132. Tuncbilek, E. and Ulusoy, M., 1988. ‘The relationship between infant mortality and consanguinous marriages’, in Infant Mortality in Turkey: Basic Factors. Haceteppe University, Ankara. Unalan, T., 1993. ‘Sex preferences in Turkey’. Turkish Journal of Population Studies 15: 23–32. United Nations, 1996. Child Survival, Health and Family Planning Programs and Fertility. New York. United Nations, 1997. World Population Prospects as Assessed in 1996. New York. Veinstein, G., 1989. ‘L’empire dans sa grandeur’, in R. Mantran (ed), Histoire de l’Empire Ottoman. Fayard, Paris, 165–167. Wolpin, K., 1997. ‘Determinants and consequences of the mortality and health of infants and children’, in M. Rozensweig and O. Stark (eds), Handbook of Population and Family Economics, Vol. 1B. North-Holland, Amsterdam, 484–557. World Bank, 1997. 1997 World Development Indicators. Washington.