Author's note: This article is part of a wider study on the Eastern Question in Turkey supported by the Ford Foundation through a MERC Award. I thank the foundation for its support, and four anonymous referees of this journal for helpful comments.

¹ The half is supposed to be the gypsies.

² On the millet system in the Ottoman Empire, see Karpat, Kemal, “The Ethnicity Problem in a Multi-Ethnic Anational Islamic State: Continuity and Recasting of Ethnic Identity in the Ottoman State,” in Ethnic Groups and the State, ed. Brass, Paul (Totowa: N.J.: Barnes and Noble Books, 1985), 102–5Google Scholar.

³ Most of the ethnic groups are not minorities by law. By the Lausanne Treaty of 1923Google Scholar, the Republic of Turkey recognizes only non-Muslim minorities and no ethnically based ones.

⁴ Actually, 51 of the subgroups of Turcomen, namely Tahtaci, Alevi, Karaçay, and Kumuk, are included; see Andrews, Peter Alford, “Catalogue of Ethnic Groups,” in Ethnic Groups in the Republic of Turkey, ed. and comp. Andrews, Peter Alford (Wiesbaden: Dr. Ludwig Reichert Verlag, 1989), 53–178Google Scholar. This is a ground-breaking work and the most comprehensive encyclopedic compilation to date on ethnic groups in Turkey.

⁵ Estimates of the number of Kurds in Turkey in the early 1990s range from 7 million to 15 million; see Özsoy, Ali Erman, Koç, İsmet, and Toros, Aykut, “Türkiye'nin Etnik Yapisinin Ana Dil Sorulaina Göre Analizi,” Turkish Journal of Population Studies 14 (1992): 101–14Google Scholar; Abramowitz, Mortan, “Dateline Ankara: Turkey After Özal,” Foreign Policy 91 (6 06 1993): 174Google Scholar; Dickey, Christopher, “A Game Kurdish Roulette,” Newsweek, 12 07 1993, 33Google Scholar; Gunter, Michael M., The Kurds in Turkey: A Politica Dilemma (Boulder, Colo.: Westview Press, 1990), 6–8Google Scholar. Bruinessen's, Van estimate for 1975 is 7.5 million, which, at the national growth rate of 2.21 percent per annum between 1975 and 1990, should have reached 10.45 million in 1990Google Scholar; Bruinessen, Martin van, Aǧa Şeyh ve Devlet trans. Asian, Remziye (first published in Dutch in 1978) (Ankara: Özge Yayinlan, 1992), 35Google Scholar. Some Kurdish leaders and sympathizers quote figures as high as 20 million for 1990Google Scholar; Burkay, Kemal, Geçmişten Bugűne Kürtler ve Kürdistan: Coğrafya-Tarih, Edebiyat (İstanbul: Deng Yayinlan, 1992) 22–25Google Scholar; interview with Server Tanilli by Sever, Metin in Kürt Sorunu: Aydmlarimiz Ne Düşünüyor, ed. Sever, Metin (İstanbul: Cem Yayinevi, 1992), 262Google Scholar, Bayrak, Mehmet, Kürtler ve Ulusal-Demokratik Mücadeleleri, (Ankara: Özge Yayrnlan, 1993), 585Google Scholar.

⁶ Kendal, gives the number of Kurds in Turkey as 8.5 million for 1970Google Scholar, McDowall, as 9.6 million for the late 1980s (apparently for 1985)Google Scholar, and Izady, as 13.65 million for 1990, accounting for 23.6 percent, 19 percent, and 24 percent of the total population in the respective yearsGoogle Scholar. It is not clear how they arrived at these figures, as they give neither the sources of data nor the method of estimation. But an examination of the census material suggests how they might possibly have arrived at these figures. It seems that they just added up the census populations of the 17 provinces in the east and the southeast (for these provinces, see the note under Table 3 in this article), disregarding the presence of other ethnic groups in the area, especially in the eastern provinces; to this is added another half-million to 1 million to account for the Kurds in the western provinces. Izady's sweep of the population history of the Kurds is so great that he makes estimates of their number since 4000 B.C. However, data and evidence for such estimates are so scant that his figures can hardly be taken seriously. See Kendal, (Nezan), “Kurdistan in Turkey,” in People Without a Country. The Kurds and Kurdistan, ed.Chaliand, G., trans. Pallis, M. (London: Zed Press, 1980), 47–48, 102–3Google Scholar; McDowall, David, “The Kurdish Question: A Historical Review,” in The Kurds: A Contemporary Overview, ed. Kreyenbroek, Philip G. and Sperl, Stefan (London and New York: Routledge, 1992), 32Google Scholar; Izady, Mehrdad, The Kurds: A Concise Handbook (Washington, D.C.: Crane Russak, 1992), 113–19Google Scholar.

⁷ The Kurdish rebellions are not new; they have been going on since 1788. See Balli, Rafet “Türkiye,” in Kürt Dosyasi, ed. Balli, Rafet (İstanbul: Cem Yayinevi, 1991), 50, 82Google Scholar; Kabacali, Alpay, Tarihimizde Kürtler ve Ayaklanmalari (İstanbul: Cem Yayinevi, 1991)Google Scholar; Celil, Celile, XIX. Yüzyil Osmanli İmparatorluğu'nda Kürtler, trans, (from Russian) Demir, M. (Ankara: Özge Yayinlari, 1992)Google Scholar; Timuroğlu, Vecihi, Dersim Tarihi (Ankara: Yurt Kitap–Yayin, 1991)Google Scholar; also Bayrak, Kürtler, passim, İmset, İsmet G., PKK: Aynhkçi Şiddetin 20 Yilt (1973–1992) (Ankara: Türkish Daily News Yayinlan, 1993), 373Google Scholar. For a concise account of the Kurdish question in Turkey since the birth of the Turkish Republic, and of the PKK, see Gunter, , The Kurds, 12–19, 57–117Google Scholar.

On the Kurds' demands for a separate or federated state, see the interviews conducted by Balli with Abdullah Öcalan, the leader of the PKK, Mahmut Baksi; and Okçuoğlu, Ahmet Zeki in Balli, Kürt Dosyasi, 236, 144, 192, respectivelyGoogle Scholar. Although the militant groups want to secede from Turkey, support for this idea seems to be very limited among Kurds who are not of a socialistic and militant persuasion; see Akyol, Taha, “Güneydoğu Sorununda Diyarbakrr Örneği;” Milliyet, 5 06 1993, 15Google Scholar.

Some see autonomy for the Kurds as inevitable. For example see, Barkey, Henri J., “Turkey's Kurdish Dilemma,” Survival 35 (4), 66–67Google Scholar. Abramowitz sees an autonomous Kurdish area in the southeast as “improbable,” but reckons that “a serious federalism” may be needed: Abramowitz, “Dateline Ankara,” 175–76Google Scholar.

⁸ Kişlah, Ahmet Taner, “Burkay, Anter ve ‘Çuvahni Seçmek’,” Cumhuriyet, 25 08 1993, 2Google Scholar.

⁹ See Andrews, , “Introduction,” in Ethnic Groups in the Republic of Turkey; Brass, “Ethnic Groups and the State,” in Ethnic Groups and the State, 16–17Google Scholar; idem, “Ethnicity and Nationality Formation,” Ethnicity 111.3 (1976): 225–28Google Scholar.

¹⁰ Andrews, , “Introduction,” Ethnic Groups in the Republic of Turkey, 20–27Google Scholar.

¹¹ An example of an etic view of an ethnic group is provided by the case of Laz in Turkey. See Meeker, M. E., “The Black Sea Turks: Some Aspects of Their Ethnic and Cultural Background,” International Journal of Middle East Studies 2 (1971): 321Google Scholar.

¹² Mumcu, Uğur, Kürt-İslam Ayaklanmasi, 1919–1925, 7th printing (İstanbul: Tekin Yayinevi, 1993), 59Google Scholar; Olson, Robert, Kürt Milliyeçiliğinin Kaynaklari ve Şeyh Said İsyani, trans. Peker, Bülent and Kiraç, Nevzat (Istanbul: Özge Yayinlan, 1992), 39Google Scholar; Burkay, , Kürtler, 447Google Scholar; Dersimi, M. Nuri, Dersim ve Küirt Milli Mücadelesine Dair Hatiratim (Istanbul: Özge Yayinlan, 1992), 66, 123Google Scholar; İmset, interview with Öcalan in November 1991, PKK, 387Google Scholar; Balli, interviews with Öcalan, and Burkay, , in Kürt Dosyasi, 389, 402Google Scholar; Andrews, , “Introduction,” Ethnic Groups in the Republic of Turkey, 35Google Scholar.

¹³ Gökalp, Ziya, Kürt Aşiretleri Hakkmda Sosyolojik Tetkikler (Istanbul: Sosyal Yayinlar, 1992), 24–25Google Scholar. Gökalp, wrote the book, in the form of a report, circa 1922Google Scholar. Minorsky, Vladimir, Kürtler, trans. Yayinlan, Koral (from the French edition of Encyclopedia of Islam (Istanbul: Koral, 1992), 76–80Google Scholar. For the etic view of the Kurds as those speaking Kurdish in the 1930s, see Beşikçi, İsmail, Kürtlerin Mecburi İskam (Ankara: Yurt Kitap, Yayin, 1991), 111–30Google Scholar; Hatipoğlu, Ömer Vehbi, Bir Başka Açidan Kürt Sorunu (Ankara: Mesaj Yayinlan, 1992), 37–40Google Scholar.

¹⁴ Interview with Erez, Yaylm, the president of the Union of Turkish Chambers of Commerce and Bourses at the time of the interview, “Doğu'da Yeniden Yapilanma Lazim,” Nokta, 6–12 06 1993, 21Google Scholar.

¹⁵ interview with Taş, Dr. Kazim, the owner of Ataköy Hospital and Martinez Hotel in Istanbul, “İstanbul'da Bulunmaktan Memnunum Ama…,” Nokta, 6–12 06 1993, 18Google Scholar.

¹⁶ Han, Şeref, Şerefname: Kürt Tarihi, trans, (from Arabic) Bozarslan, M. Emin, 3rd printing (Istanbul: Hasat Yayinlan, 1990), 34Google Scholar.

¹⁷ Gökalp reaches this judgment on the testimony of Şerefname; Gökalp, , Kürt Aşiretleri, 95Google Scholar.

¹⁸ Ibid., 24–25.

¹⁹ Seliç, H., “Zaza Gerçeği,” Tercüman, 26, 27, 28, 29, and 30 06 1993Google Scholar.

²⁰ MacKenzie, David N., “The Role of the Kurdish Language in Ethnicity,” in Ethnic Groups in the Republic of Turkey, 541Google Scholar; Minorsky, Kürtler. This is also the view of Izady, The Kurds, 167, 169, 173–74Google Scholar. Kreyenbroek argues that North Kirmanci (also called Bâhdinâni), spoken mainly in Turkey and South Kirmanci, and Sorâni, spoken in Northern Iraq, are also two different languages, differing in grammar, vocabulary, and pronunciation: Kreyenbroek, Philip G., “On the Kurdish Language,” in The Kurds: A ContemporaryOverview, 71–72Google Scholar. Hassanpour, however, leans toward the view that Zaza is a Kurdish language; See Hassanpour, Amir, Nationalism and Language in Kurdistan 1918–1985 (San Francisco: Mellen Research University Press, 1992)Google Scholar, passim; see also his review of Nader Entessar's book Kurdish Ethnonationalism, idem, “Kurdish Studies: Orientalist, Positivist and Critical Approaches,” Middle East Journal 47, 1 (Winter 1993): 120Google Scholar.

²¹ Andrews, , “Catalogue,” 122Google Scholar; Bruinessen, Martin M. van, “The Ethnic Identity of the Kurds,” in Ethnic Groups in the Republic of Turkey, 613Google Scholar; Minorsky, , Kürtler, 76–80Google Scholar; Kreyenbroek, , “On the Kurdish Language,” 70Google Scholar; Izady, The Kurds, passim.

²² Bruinessen, Van, “Ethnic Identity,” 619Google Scholar.

²³ Ibid., 618, Gökalp, , Kürt Aşiretleri, 32, 127–28Google Scholar.

²⁴ Van Bruinessen casts doubt on the census figures for two reasons. He argues that Kurds were either not counted at all, or some of the Kurds were excluded from enumeration as Kurds because of a narrow definition of “Kurd.” He is possibly right in the first case, but is wrong in the reason he gives for underenumeration. Census officials are not sent out from the province centers, as he claims, but from the district centers, from which the villages are more easily accessible. See Bruinessen, van, Şeyh, Ağa ve Devlet, 33–34 and n. 4 in Section I, 389Google Scholar.

²⁵ According to estimates by Shorter and Macura, the gross fertility rate—or births per thousand population—was 56 in 1945, 59.8 in 1955, and 56.2 in 1965 in the east, where the Kurds mainly liveGoogle Scholar; it was 45.2, 46.7, and 42.3 for Turkey as a whole in the respective years. See Shorter, Frederic C. and Macura, Miroslav, Türkiye'de Nüfus Artişt (1935–1975): Doğurganlik ve Ölümlülük Eğilimleri (first published in English in 1982) (Ankara: Yurt Yayinlan, 1983), 57, 90Google Scholar.

²⁶ McCarthy, Justin, Muslims and Minorities: The Population of the Ottoman Anatolia and the End of the Empire (New York and London: New York University Press, 1983), 150 (n. 6), 160Google Scholar.

²⁷ These are Mardin and Kahramanmaraş. McCarthy is of the opinion that in the 1927 censusGoogle Scholar, Kurdish population was undercounted in Hakkari, Elaziğ, and Mardin provinces; Ibid., 107.

²⁸ The progress of the times and increase in welfare is reflected in the decrease of the infant-mortality rate from 245 in 1951, to 189 in 1961, to 156 in 1967Google Scholar; see Shorter, and Macura, , Türkiye'de Nüfus Artişt, 72Google Scholar.

²⁹ Keleş, Ruşen, Yerinden Yönetim ve Siyaset (Istanbul: Cem Yayinevi, 1992), 268–74Google Scholar; see also Yayin, Teoman, Merkezi Hükümet ile Mahalli İdareler Arasindaki Mali İlişkiler (Ankara: DİE, 1971)Google Scholar.

³⁰ A good example of overstatement of population by 100 percent is provided by the case of Kinkkale in the 1985 censusGoogle Scholar; see Keskin, Y., “Kirikkale İI Olmaya Hazir,” Dünya, 9 06 1989, 12Google Scholar.

³¹ The number of provinces was 67 from the 1955 to the 1985 censusesGoogle Scholar. By the 1990 census, however, there were 73 provinces.

³² The gross reproduction rate is the average number of female children a woman would have if she survived to the end of her childbearing years, and if, throughout, she were subject to a given set of age-specific fertility rates and a given sex ratio at birth.

³³ Mean age of fertility is the average age at which a mortality-free cohort of women bear their children according to a set of age-specific fertility rates.

³⁴ The total fertility rate is the average number of children that would be born per woman if all women lived to the end of their childbearing years and bore children according to a given set of age-specific fertility rates. This is also referred to as total fertility.

³⁵ The model life table is an expression of typical mortality experience derived from a group of observed life tables. A life table lists the number of survivors at different ages—up to the highest age attained—in a hypothetical cohort subject from birth to a particular set of age-specific mortality rates. It is commonly accompanied by other features of the cohort's experience, such as expectation of life at each age and the probability of surviving from each age x to age x + n.

³⁶ Shorter, and Macura, , Türkiye'de Nüfus Artişi, 89–90Google Scholar. For the definition of the various demographic terms used here, see Cox, Peter, Demography, 5th ed. (Cambridge: Cambridge University Press, 1976CrossRefGoogle Scholar).

³⁷ Her figures are corrected total fertility rates. The correction factor is P3/F3. For an explanation of Pi and Fi, and for 1985 total fertility rates, see Cerit, Sevil, Türkiye'de Nüfus, Doǧurganlik, Ölümlülük (Ankara: Hacettepe Üniversitesi, 1989), 45, 87–88Google Scholar.

³⁸ Teşkilati, Devlet Planlama, Sosyal Yapi-l: Türkiye'de Sosyal Yapisindaki Gelişmeler ve Uluslararasi Karşilaştirmalar, DPT no. 2054-SPB:396 (Ankara: Devlet Planlama Teşkilati, 1986), 165–66Google Scholar. For 1990, both the unadjusted and adjusted total fertility rates are given; the adjusted rates are used in this study.

³⁹ Bakanhǧi, Saǧhk ve Sosyal Yardrm, Okulu, Hifzisihha, Türkiye Nüfus Araştirmasindan Elde Edilen Hayati İstatistikler 1966–67,(Ankara: Hacettepe Basimevi, 1970), 39, 46Google Scholar.

⁴⁰ Hacettepe Üniversitesi, Enstitüsü, Nüfus Etüdleri, Türkiye'de Nüfus Yapist ve Nüfus Sorunlan: 1973 Araştirmasi (Ankara: Hacettepe Üniversitesi Yayinlari, 1978Google Scholar); Hacettepe University, Institute of Population Studies, Turkish Fertility Survey, Vol. 1 (Ankara, circa 1982Google Scholar); idem, 1983 Turkish Population and Health Survey (Ankara: 1987Google Scholar); idem, 1988 Turkish Population and Health Survey (Ankara, 1989Google Scholar); Enstitüsü, Devlet İstatistik, Türkiye Nvfus Araştirmasi, 1974–75, yayin no. 841 (Ankara, 1978Google Scholar); idem, Turkish Demographic Survey 1989, yayin no. 1483 (Ankara, 1991Google Scholar).

⁴¹ The figure is from an empirical study by Yücetürk. See Yücetürk, Alper, “Sex Ratio and Monthly Distribution of Births,” in Turkish Demography: Proceedings of a Conference, ed. Shorter, Frederic C. and Güvenç, Bozkurt (Ankara: Hacettepe University, 1969), 167–73Google Scholar. In the most recent demographic survey, the ratio is slightly higher; Enstitüsü, Devlet İstatistikTurkish Demographic Survey 1989, 43 (n. 2)Google Scholar.

⁴² In the model life tables, gross and net reproduction rates are coupled with the MAFSs of 27, 29, 31, and 33Google Scholar. In employing the tables to estimate growth rates, the MAFS nearest to the provincial MAFS was used.

⁴³ Shorter, Frederic C., The Decline of Infant and Child Mortality: Estimates from the Turkish Census, MEAWARDS Regional Papers, no. 34 (Cairo: The Population Council, West Asia and North Africa, 1989,) 19–20Google Scholar; UNICEF, Analysis of Children and Mothers: Situation in Turkey, draft 2 (Ankara: UNICEF, 28 02 1990) 2–12, 2–13Google Scholar; Government of Turkey-UNICEF Programme of Cooperation, The Situation Analysis of Mothers and Children in Turkey (Ankara: UNICEF Turkey Office, 04 1991), appendix Table 6–1, 130Google Scholar.

⁴⁴ A study on Turkish mortality patterns has found that it is best represented by the East Model Life Table among the Coale-Demeny Life Tables; see Shorter, and Macura, , Türkiye'de Nüfus Artişt, 70–71, 92–93Google Scholar; Coale, Ansley J. and Demeny, P., Regional Model Life Tables and Stable Populations (Princeton, N.J.: Princeton University Press, 1966Google Scholar). There is, of course, no perfect fit between the two, and at least around 1960Google Scholar, the early child mortality in Turkey was considerably higher than that shown in the East Model Life Table; see Demeny, P. and Shorter, Frederic C., Estimating Turkish Mortality, Fertility and Age Structure: Application of Some New Techniques (Istanbul: Istanbul University Press, 1968Google Scholar). This discrepancy decreased as infant mortality fell in Turkey in the 1970s and 1980sGoogle Scholar. Hence, the growth rate and population estimates for the 1970s and 1980s, using the East Model Life Tables, should give progressively more accurate figures.

⁴⁵ The GRR in the Turkish Demographic Survey of 1989 for Region 1, which includes İstanbul, İzmir, and Bursa, is below the rate which would make for zero population growth under the mortality conditions in the mid-1980s; see Cerit, , Türkiye–de Nüfus, 55Google Scholar.

⁴⁶ Enstitüsü, Devlet İstatistik, Genel Nüfus sayimi, 12.10.1980. Daimi İkametgaha Göre İç Göçler (Ankara, 1985Google Scholar); idem, Genel Nüfus Sayimi, 20.10.1985. Daimi İkametgaha Göre İç Göçler (Ankara, 1989Google Scholar). Immigration and emigration data by province for 1970–85 are from Teşkilati, Devlet Planlama, Sosyal Yapi-l, 82–84Google Scholar; 1985–90 migration data were obtained from the State Institute of Statistics' computer files.

⁴⁷ In the private file of Ayşe Gedik of Middle East Technical University, Ankara. She obtained them from the SIS in the 1970s and kindly made them available to me.

⁴⁸ Gedik, Ayşe, “Türkiye'deki İç Göçler, 1965–85; Bazi Varsayimlann Sorgulanmasi,” in Kent, Planlama, Politika, Sanat: Tank Okyay Amsina Yazilar, ed. Tekeli, İlhan (Ankara: METU, 1994), 220–23Google Scholar; Tümertekin, Erol, Türkiye'de İç Göçler (İstanbul: İstanbul Üniversitesi Yayini, 1968), 25–47, 60–74Google Scholar.

⁴⁹ In addition to legal emigrants, there was a substantial stream of asylum seekers in European countries in the 1980s. According to one estimate, the number of asylum seekers was 57,913 in 1980Google Scholar; 20,020 in 1989; and 22,082 in 1986, 1987, and 1988; see Bakanhǧi, Çalişma ve Sosyal Güvenlik, Müdürlüǧü, Yurtdişi İşçi Hizmetleri Genel, 1990 Yih Raporu (Ankara, 1991), 8–9Google Scholar; Simon, G., “Migration in Southern Europe: An Overview,” in The Future of Migration, OECD (Paris: OECD, 1987), 269Google Scholar.

⁵⁰ Bulutay gives the complete series on emigration of Turkish workers, emigration of population, stock of Turkish workers, and stock of population for the period 1961–1990Google Scholar. His series is used in this study. See Bulutay, Tuncer, External Migration in Turkey, draft (Ankara, 15 05 1992Google Scholar). For outflows of workers to Arab countries, use is also made of İş ve İşçi Kurumu, Bulma, 1991 Statistical Yearbook (Ankara: İİBK, 1992Google Scholar).

⁵¹ The annual data on Turkish births and deaths in the Western European and Arab countries by country are available for 1982–1990 in SOPEMİ reportsGoogle Scholar; see OECD, SOPEMİ (Systeme d'Observation Permanante des Migrations) (Paris: OECD, 1978–1991)Google Scholar; also, Müdürlüǧü, Yurtdişi İşçi Sorunlan Genel, Yurtdişi İşçi Sorunlan '81, (Ankara: Başbakanlik Basimevi), 18Google Scholar.

⁵² The following contain data for various years on the proportion of Turkish workers who were married, and on the ratio of married workers who had their spouses with them: Abadan-Unat, N., Bati Almanya'daki Türk İşçileri ve Sorunlan (Ankara: Devlet Planlama Teşkilati, 1964), 64Google Scholar; Teşkilati, Devlet Planlama, Yurt Dişindan Dönen İşçilerin Sosyo-Ekonomik Eǧilimleri Üzerinde Bir Çalişma (Ankara: Devlet Planlama Teşkilati, 1974Google Scholar), Tables 4 and 5; Kudat, Ayşe, Stability and Change in the Turkish Family at Home and Abroad: Comparative Perspectives (Berlin: International Institute for Comparative Studies, 1975), 32–33Google Scholar; İş ve İşçi Kurumu, Bulma, '72 Örnekleme Araştirmasi: Almanya'da Çalişan Yabanci İşçilerin İstihdam. Aile ve İkamet Sorunlari, yayin no. 117 (Ankara: İş ve İşçi Bulma Kurumu, 1974), 12–13Google Scholar; Gökdere, Ahmet, Yabanci Ülkelere İşgücü Akimi ve Türk Ekonomisi Üzerine Etkileri (Ankara: Türkiye İş Bankasi Kültür Yayinlari, 1978), 54Google Scholar; Heijecke, H., “The Labour Market Position of Migrants in Selected European Receiving Countries,” in The Future of Migration, 183Google Scholar. Most of these contain information on the situation in Germany only; the conditions in Germany were assumed to hold for all European countries.

⁵³ The data on the origins of workers in Germany by home region for 1961–73 are from ÖzkanAyşe Kudat, Yilmaz Ayşe Kudat, Yilmaz, and Öncü, Yaşar, “Yurtdişina İşçi Göçünün Yöresel Boyutlari,” The 2nd Turkish Demography Conference, Tables 12 and 13, Map 1Google Scholar; Gökdere, , Yabanci Ülkelere İşgücü Akimi, 278Google Scholar. The regional data were disaggregated to provinces on the assumption that the number of workers and their dependents emigrating from a province in a particular region was proportioned to the ratio of that province's population to the region's population. For the period 1974–1977Google Scholar, for which no data by origin exist, it was assumed that the provincial pattern of emigration of workers was the same as that of 1978–1980Google Scholar. The totals were distributed to provinces accordingly.

⁵⁴ The data on the origins of workers by province for 1978–1990 are from İş ve İşçiGoogle ScholarKurumu, Bulma, İstatistik Yilliǧi, annual issues for 1978–1990Google Scholar.

⁵⁵ Gökdere, develops a series on returnee workers for 1962–1973Google Scholar, and Bulutay, for 1961–1990Google Scholar. See, Gökdere, , Yabanci Ülkelere İşgücü Akimi, 102, Table 11.38Google Scholar; Bulutay, , External Migration in Turkey, 115Google Scholar, table 5.5 and 119, table 5.9. However, for the years 1961 to 1973, Bulutay, whose series is used in the study, does not take into consideration the growth of the Turkish population due to births. To rectify this omission, the formula given in Appendix B of this article was used to estimate the returnee population during 1965–1973Google Scholar.

⁵⁶ One scholar observed that the fertility rates of Turkish emigrants were 2.6 to 3.2 times higher than that of the Germans during 1975–1981Google Scholar; see Maillat, D., “Long Term Aspects of International Migration Flows,” in The Experience of European Receiving Countries, OECD, 47Google Scholar.

⁵⁷ This is not entirely true because of the existence of illegal workers, students, and asylum seekers. It is highly probable that the provincial origins of asylum seekers were different from those of workers. Especially in the 1980s, according to press reports, besides real political refugees, people from the east and southeast of the country seeking work were posing as political refugees to get in. This may well have introduced biases in our calculation of the provincial origins of outflows from Turkey.

⁵⁸ The workers in Arab countries are mostly employed by Turkish contractors. When the work is completed, the workers have to return because of the visa conditions. Personal communication with the contractors indicates that a contract takes about three years to complete.

⁵⁹ The method followed here may underestimate the number of Kurds in these large provinces, because the census ratio of the Kurds to total population in any one of these selected provinces is lower than the estimated one. However, the underestimate arising from this is probably so slight as to be negligible, so no second round was attempted.

⁶⁰ The infant-mortality rate was 280 in eastern Anatolia in 1945, whereas it was 190 in İstanbul and İzmir. By 1965, it had fallen to 108 in İstanbul and İzmir. But it was still 187 in the east; Shorter, and Macura, , Türkiye'de Nü/us Artişt, 77, 101Google Scholar.

⁶¹ The number of public-housing units built for the refugees from Bulgaria in various provinces, as of August 1994, was: Adana, 50; Ankara, 1,544; Balikesir, 500; Bursa, 7,218; Edirne, 220; Erzurum, 40; Eskişehir, 1,200; İstanbul, 5,050; İzmir, 2,040; Kocaeli, 1,000; Kütahya, 44; Manisa, 240; Sakarya, 1,000; Tekirdaǧ, 1,410. The data are from the state ministry responsible for Bulgarian refugee affairs.

⁶² Burkay, , Kürtler ve Kürdistan, 24Google Scholar. In his interview with Imset, Öcalan claimed, “Half of Istanbul is now Kurdish”; see Imset, , PKK, 374Google Scholar. The perception that half of the Kurds now live in the west, despite lacking an objective basis, is very prevalent among politicians and in the press.

⁶³ The number of Kurds of Turkish citizenship living abroad has not been analyzed here. A recent article by Robins puts their number in Western Europe at 500,000. See Robins, Philip, “The Overlord State: Turkish Policy and The Kurdish Issue,” International Affairs 69, 4 (1993): 662–63CrossRefGoogle Scholar. However, an analysis of the origin of workers by region—and of the number of asylum seekers, especially in the 1980s—indicates that Robins's number is too high. The majority of workers in Europe were from the western regions where the Kurds did not live in substantial numbers. On the regional origin of workers, see the references in n. 53 of this article.

⁶⁴ If the Kurdish population increases at the 1965–90 rate, at 3.24 percent per annum, it will reach 9.74 million by the year 2000 and will constitute 13.82 percent of the total population, assuming that the non- Kurdish component of the population increases at the 1965–90 average rate of 2.17 percent per annum.

⁶⁵ In 1960, 11.45 percent of the population lived in a province other than the one in which he/she was born. This had risen to 12.80 percent in 1965, and to 20.95 percent by 1985. The percentages are calculated from the data in the 1960, 1965 and 1985 issues of Enstitüsü, Devlet İstatistik, Nüfusun Sosyal ve Ekonomik Nitelikleri, 71, 32, and 59Google Scholar, respectively by year.