15. Bazin (1948).

16. Ramsey (1987: 102-3, 139-40, 236-7). Strictly speaking Cantonese has nine tones, having added one more split.

17. Hashimoto (1986) argues a little too desperately that Chinese was effectively 'Altaicised' in the north, but his evidence is confined to transitory pidginised states of the language in Beijing, and a deviant contemporary dialect in Qinghai, where speakers are probably bilingual in Tibetan.

18. Norman (1988: 20).

19. w.a.n.g (1992: 11).

20. Hall (1981: 212).

21. Coedes (1968: 37). See Chapter 5, 'Sanskrit in South-East Asia', p. 204.

22. w.a.n.g (1992: 16).

23. Grousset (1970: 66).

24. Mote (1999: 25, 980).

25. The figures for Egypt are derived from Dollinger (2002), and for China from Barraclough (1978: 80, 127). McEvedy and Jones (1978) suggest a rather lower figure for Egypt in Roman times, 5 million. They simply dismiss the estimate in Diodorus (i.31) of 7 million for Egypt in 300 BC as 'too high'. For China, they point out that the AD 2 census figures are actually for 11.8 million households. They estimate that China's population then stayed close to 50 million until the beginning of the second millennium AD, when it began to rise with the greater cultivation of rice in the Yangtze valley, reaching 115 million in 1200, but then falling back in the Mongol era and not recovering until 1500. None of the above affects the general point about the exceptionally high population density of Egypt and China in the pre-modern world.

26. Figures derived from Russell (1958).

27. Pritchard (1969: 415).

28. Arnett (1982: 45-7).

29. Sallier 2,9,1 = Anastasi Papyri 7,4,6, quoted in Erman (1894: 328).

30. Anastasi Papyri 5, 10, 8ff., quoted in Erman (1894: 328).

31. Ramsey (1987: 121-3). See Chapter 5, p. 209.

32. Norman (1988: 257-63).

33. Wilkinson (2000: 735).

34. The Economist, 9 March 1996, p. 4, cited in Graddol (1997: 37).

35. Karlgren (1954). Its principles are set out succinctly in Norman (1988: 34-42).

36. Pritchard (1969: 440).

37. Wilkinson (2000: 723).

38. Translated by Mote (1999: 156), from Lin Tianwei (1977): Bei Song jiruo de sanzhong xin fenxi. Song shi yanjiu ji 9, 147-98.

39. Gao (1991: 145).

40. Ramsey (1987: 224).

5 Charming Like a Creeper: The Cultured Career of Sanskrit

1. Rig Veda, vii. 103.

2. ibid., x.34.

3. Mahabhaya, i.1.

4. Ojha, Bharatiya Pracina Lipi Mala, 14, no. 6, attributed to Canakya-niti.

5. Caesar, De Bello Gallico, vi. 14.

6. Martin Prechtel, personal communication.

7. Plato, Phaedrus 275A.

8. Mahabharata, quoted by Kesavan (1992:3).

9. Brough (1968:31).

10. Deshpande (1993: 24), quoting Mahabhaya, i, p. 2.

11. Patanjali, Mahabhaya on Panini, vi.3.109, trans. Deshpande (1993: 62).

12. Manu, ii. 18-22.

13. Deshpande (1993: 86).

14. ibid.: 16; Rajasekhara, Kavyamimamsa, iv.

15. Strabo, xv.1.21.

16. ibid., xv. 1.64.

17. Milindapanha, i.9.

18. Fo-Kwo-Ki, x.x.xvi (in Beal 1884: lxxi).

19. ibid., xl (in Beal 1884: part 1, p. lxxix).

20. ibid., xl (in Beal 1884: part 1, p. lx.x.xiii).

21. Coedes (1968: 81-2).

22. Si-Yu-Ki, ii.9 (in Beal 1884: part 1, pp. 77-8).

23. Gidwani(1994).

24. Rig Veda, ii.20.7.

25. Chatterji (1966: 78).

26. Si-Yu-Ki, x.9-11 (in Beal 1884: part 2, pp. 204-8).

27. Pancatantra, v. 31.

28. Keith Taylor, in Tarling (1999: 195).

29. Kamara, Podouke and Sopatma, 'lying in a row', are quoted in the first century AD Periplous of the Erythraean Sea (ch. 60). Of these, the first two are presumably on the delta of the Kaveri river and at Pud.u.c.h.erry (better known as Pondicherry).

30. Yule and Burnell (1986: 456): 'It is a saying in Goozerat,-"Who goes to Java Never returns. If by chance he return, Then for two generations to live upon, Money enough he brings back Rss Msls, ii.82 (1878 edn: 418).

31. Majumdar (1975:21).

32. Coedes (1968: 26-7, 36, 275).

33. ibid.: 37, 276.

34. Majumdar (1975: 13).

35. ibid.: 19-20.

36. ibid.: 48.

37. Mahabharata, Aranyakaparva, 173; Majumdar (1975: 25-7).

38. Coedes (1968: 369).

39. Fo-Kwo-Ki, xl (in Beal 1884: part 1, p. lx.x.xi).

40. Coedes (1968: 17); Bechert and Gombrich (1984: 147).

41. Ramsay (1987: 121-4).

42. For the details of the Tibetan script and its origin, I have been dependent on Beyer (1992:40-50).

43. There is some evidence that Tibetans could write earlier than this. There are extant contemporary annals of the period 650 to 747, and for the year 655 we find: The King stayed at Mer-khe, and the prime minister Ston-tsan wrote the text of his commands to Ngor-ti.

In fact the introduction of the script is traditionally (i.e. in a history from the fourteenth century) credited to a Tibetan scholar and government minister, Thon-mi Anui-bu, said to have been sent on a mission to India in the mid-seventh century. But Thon-mi may have been an invented figure, since he is omitted from genuinely ancient records of Tibet found in central Asia, while the earliest grammatical works on Tibetan are also attributed to him.

44. Beyer (1992: 36-7).

45. As conjectured in van Leur (1955: 113) and discussed in Hall (1981: 231-3).

46. Basham (1967:491).

47. Rangarajan (1992: 18-21).

48. Si-Yu-Ki, ix (in Beal 1884: part 2, pp. 171-2).

6 Three Thousand Years of Solipsism: The Adventures of Greek

1. Old Oligarch, Athenian Const.i.tution, ii.8: kai hoi men h.e.l.lenes idiai mallon kai phbar;nei kai diaitei kai skhmati khrbar;ntai, Athenaoi de kekramenei ex hapanton tbar;n h.e.l.lebar;non kai barbaron.

2. Herodotus, viii.144 (quoted in the epigraph to this chapter).

3. ibid., iv. 183.4. They lived along the Red Sea coast, according to Strabo, xvii.1.2.

4. Aeschylus, Agamemnon, 1050-1.

5. Thucydides, ii.35-46.

6. tebar;n polin pabar;san tbar;s h.e.l.lados paideusin enai: Thucydides, ii.41.

7. Menander, fragment 72, ed. k.o.c.k.

8. Herac.l.i.tus, fragment 119.

9. Aristophanes, Knights, 1169.

10. Hesiod, Catalogues of Women (Loeb edn, fr. 4).

11. Thucydides, iii.38.4.

12. Buck (1955: 10-14).

13. Strabo, vi.1.2.

14. Segs 30.1664 and 20.326 (Greek-Aramaic Buddhist text), Schlumberger et al. (1958). See Chapter 5, 'The character of Sanskrit', p. 187, and Chapter 3, 'Aramaic-the desert song: Interlingua of western Asia', p. 84.

15. Salomon (1998: 265-7). Heliodoros comes out as Heliodora-, but Antialkidas as Amtalikita.- Very much in the Asoka tradition, it contains gratuitous urgings to Buddhist virtue. See Chapter 5, 'Outsiders' views', p. 192.