1. The figures break down as follows. In the Loeb, king occurs 26 times; queen 9; sovereign, prince, royal 4 each; palace 2; crown, throne, empire 1 each: in the Penguin, king 29 times; queen 3; prince and palace 2 each; empire, crown, majesty, and royal 1 each: in the Budé, roi 15 times; trône 7; prince 6; regner and palais 5 each; souvérain 3; couronne 1.

2. ‘Anax’ and its cognates occur 27 times in O. T., fourteen of them in reference to Oedipus. But it is also used for addressing Creon, Teiresias, and the members of the chorus. It is not therefore a royal word.

Translators who gratuitously royalize the language may suppose that they are simply updating court etiquette, in the spirit of Dacier's curious remark in 1693 in his translation of O.T that it was ‘la manière ancienne de parler aux Rois měme, en les nommant par leur nomme’.

3. Aristophanes, Thes. 295–311Google Scholar; Ath. Pol. 16.10 (on which see Rhodes, P.J.The Athenian Boule (Oxford, 1972), p. 37Google Scholar, and Ostwald, M., ‘Athenian Legislation against Tyranny’ TAPA 86 (1955), 113–14Google Scholar); Thucydides 1.18.1 (also 6.89.4 where Alcibiades takes it for granted that Spartan opinion is adverse to tyrants); Thucydides 1.122.3.

4. In P. V., lyrannos and words formed from it are reserved for Zeus, the unjust oppressor; in the Oresteia they describe the gods of the lower world and Agamemnon's enemies, Priam, Aigisthos, and Clytemestra. At Ch. 479 Agamemnon himself is explicitly said not to have died as a tyrant, the implication being that Orestes and Electra are entitled to their birthrights and are not legally disqualified as a tyrant's children would be. See my remarks in Greek Tragedy and its Legacy: essays presented to D. J. Conacher (Calgary, 1986), pp. 14–15Google Scholar.

5. Ant. 1056; O.T. 380; fr.88.3 R.: Ant. 506; O.T. 408; O.C.851; fr.873.1 R.: O.C. 419: O.T. 873.

6. E.g. Med. 348, Hipp. 1013, Suppl. 429.

7. O.T. 925. In the superficially similar El. 661 it is Orestes' aged servant asking for the house of the usurper Aigisthos, and a morally neutral tone cannot be assumed.

8. O.T. 514, 1097; 1043; 939.

9. On this see Knox, Bernard, Oedipus at Thebes (Yale, 1957), pp. 67–77Google Scholar.

10. Gorgias B11, B11a (DK⁶ ii, 289.8; 297.23, 299.18).

11. Kritias B25.6 (DK⁶ ii, 386.26).

12. E.g. Ajax 1350; Ant. 506, 1056; O. T. 380, 408, 873. Add too the occasions where a tyrant or his dependants are said to look the part – e.g. El. 664, Ant. 1164, Tr. 316 – for this means they look as if they had power, just as we say of people that they look as if they had money.

13. O. T. 128, 514, 535, 799, 925, 1043, 1097. On Electra 661, see note 7.

14. Denniston, J. D., Greek Particles (Oxford, 1954)Google Scholar, does not mention the τε of this passage, but it fits comfortably into the quasi-appositional category I (1) iii (e) page 502, exemplified i.a. by Aesch. Ag. 10, and Eur. I.A. 1454.

15. O.T. 14, 64ff. (Oedipus' caringness); 93 (his open government); 384 (his election to supreme office).

16. Plato Rep. 573d–576c shows particularly clearly what, in popular Greek imagination, tyrants did.

17. ‘A peccadillo?’ wonders Brink on the occasion of a similar discrepancy between what Jebb puts in his commentary and in his translation – and answers himself ‘No, a cast of mind which shows itself again and again.’ Brink, C. O., English Classical Scholarship (Cambridge and New York, 1985), p. 223Google Scholar.

Creon here talks of ‘kratos’ without ‘arche’. It is equally possible, though Jebb does not point it out, to envisage ‘arche’ without ‘kratos’. When Ajax complains that the Atreidai have rejected his κράτν (Ajax 446), he does not mean his power of command but his rights or position as a colleague. Again, when the Creon of the Antigone says (in line 485) that his κράτν) will be prostrate if Antigone gets away unpunished he does not mean that he will lose his office or his power of command in Thebes, but that his personal prestige will be undermined.

18. Denniston, J. D., The Greek Particles (Oxford, 1954), pp. 478–9Google Scholar.

19. Before anybody protests that a μᾶλλον ῆ of positive preference should not be immediately followed by one of double rejection let him consider Ajax 1310–12, 1313–15, where exactly the same thing happens.

20. ‘Act’ or ‘behave like a king’ or ‘wield a king's power’ is the sense given by Jebb (1883), Young (1906), Polites (1936), Bignone (1937), Hunt (1937), Roussel (1940), Grene (1942), Errandonea (1959), Kino (1962), Willige (1966), Wilson (1967), Gould (1970); ‘live like a king’ is the sense given by Sheppard (1920), Mazon (1965), Berg and Clay (1978). Fitts and Fitzgerald (1951), ‘a king's rights’ as opposed to ‘a king's power’, and Banks (1956) ‘royal freedom’ as opposed to ‘a royal title’ are perhaps to be classed in this group. I have found only two translations in the last seventy years that break rank. Yeats (1928), thinking perhaps that it adds nothing to what Creon has just said, leaves the line out altogether! Paduana (1982) writes ‘Io no preferisco essere il padrone al comportarmi da padrone …’. thus becoming the only translator to avoid bringing in royalty. But he has no systematic conscience on the matter: elsewhere he often calls Oedipus ‘re’ and Jocasta ‘regina’.

21. Victorius, Petrus, Variae Lectiones (Florence, 1572)Google Scholar. The essay on our passage is the fourth in Book 6.

22. Seneca Oed. 687 ‘Solutus onere regio regni bonis fruor’ is cited by Erfurdt on our line, but Seneca's words do not seem to translate it. The Greek phrase that comes closest to ‘regni bonis’ is τὰσύν κέρδει καλά in O.T. 595.

As for Eustathius, after quoting the line he adds ‘just as being a king would be distinct from acting as one (βασιλκὰ πράττειν)’. This could conceivably have suggested the idea of ‘quae faciunt reges’ to Vettori even though it does not itself imply anything about Sophocles having used the word tyrant to mean king.

23. So Gabia (1543) and Winsheim (1546). Winsheim's extremely pedestrian translation continued to be reprinted until well into the eighteenth century. The way that ‘imperata facere’ for τραννα δρᾶν was explained to pupils is perhaps shown by an ink-written note in a copy of the 1665 Cambridge edition in the Bodleian, ‘to do your acts of asking’.

24. George Rataller, in a much more spirited translation than that of Winsheim published at Antwerp in 1570, gave the line the same meaning but phrased it in a more general way. Creon would prefer not to rule (‘imperare’) but to be a subject (‘imperio obsequi’). Retailer commented in the margin that the speech gives a fine comparison between the exercise of power (‘regnum’) on the one hand and a life of private independence (‘privata vita cum dignitate’) on the other, τραννα δρᾶν was rendered ‘ubidire’ by Giustiniano in an Italian version in 1610.

After that Vettori began to win out, e.g. Grotius in 1626 (‘regni vox vana quam res regia’), Dacier (1693) ‘ce n'est pas le nom de Roy que j'ambitionne, mais d'en faire les actions’, and thereafter in the same vein Theobald (1715), Adams (1729), Franklin (1759).

But the pre-Vettori interpretation was kept alive by the frequent reprinting of Winsheim, and there was even a new version by T.Johnson (London and Eton, 1746), which kept away from royal words and contrasted ‘tyrannus esse’ with ‘tyranni iussa exsequi’. This was denounced by Brunck as absurd.

25. The story is told by Polybius (10.40.5) and by Livy (27.10.5), but neither in fact writes anything quite equivalent. In Polybius Scipio does not want to be thought ‘basileus’ but ‘basilikos’ in Livy he wants to be credited with a ‘regalis animus’. There is nothing about kingly deeds, ‘quae faciunt reges’.

26. Aristotle, Poetics 6 (1450b 5)Google Scholar.

27. EN. 9.3.4 and 10.3.12 = 1165 b 26, 1174 a 2.

28. Pol. 5.7 = 1327 b 24.

29. deAn. 2.3 = 414 b 18, 415 a 8; G.A. 2.6 = 744 a 30.

30. Po. 6 1449 b 38; N.E. 6.3.5 = 1139 b 5; M.A. 6 = 700 b 17.

31. Po. 6 = 1450 b 5–7.

32. This is how it is taken by, among others, Bywater and D. W. Lucas.

33. At Rh. 3.12 = 1414 a 11 rhetoric is said to be least in place where the audience is fewest.

34. This is the interpretation proposed by Gerald Else. See the note in his translation of the Poetics (Ann Arbor, 1967)Google Scholar.

35. ‘Komizomai’ is the standard term for recovering what belongs to you by right, and is frequently used in the orators for the recovery of dowries. The person named as recovering is normally the woman's male next-of-kin into whose guardianship she returns when her marriage is dissolved by death or divorce, but can be the woman herself (e.g. Dem.40.7.1).

36. For κλύειν's slight nuance of ‘to hear is to obey’ see Ajax 1072, 1352; Ant. 666; Ph. 925.