² ‘Wer einen Zugang zur tragischen Dichtung der Griechen sucht, muss von den materiellen und geistigen Verhältnissen ihrer Umwelt ausgehen’: Kuch, H., in H. Kuch (ed.), Die griechische Tragödie in ihrer gesellschaftlichen Funktion(Berlin, 1983), p. 7. S. Said observes, ‘Scholars from various countries and different political opinions now fully agree about the political character of Greek tragedy…. But this broad consensus on the political nature of tragedy disappears as soon as one tries to [make] explicit one's definition of “politics” and “political”’ (loc. cit., n. 1, pp. 3–4).Google Scholar

³ Edd. Winkler, J. J. and Zeitlin, F. (Princeton, 1990).Google Scholar

⁴ Pelling, C. in C. Pelling (ed.), Greek Tragedy and the Historian(Oxford, 1997)Google Scholar, p. v. The other book named here is Sommerstein, A. H., Hailiwell, S., Henderson, J., AND Zimmermann, B. (edd.), Tragedy, Comedy, and the Polisxg54:1217(Bari, 1993)Google Scholar. Also important, but too recent for Pelling, is Silk, M. S.(ed.), Tragedy and the Tragic(Oxford, 1996).Google Scholar

⁵ ‘The formal and stylistic preferences of the three great tragedians display considerable divergences, which are hard to explain solely by the methods of social anthropology’, observes Bain, D., JHS 113(1993), 187.Google Scholar

⁶ Seaford, R., Reciprocity and Ritual: Homer and Tragedy in the Developing City-State(Oxford, 1994).Google Scholar

⁷ Croally, N. T., Euripidean Polemic(Cambridge, 1994).Google Scholar

⁸ G. Cerri does in fact go so far as to speak of tragedy as being ‘vero e proprio apparato ideologico di stato’, Legislazione orale e tragedia greca (Napoli, 1979), p. 269.

⁹ We do nowadays sometimes find her presented in a very negative light: see Sourvinou-Inwood, C., ‘Assumptions and the creation of meaning: reading Sophocles’ Antigone', JHS 109 (1989), 134–148CrossRefGoogle Scholar. That goes with the view that her sex (she is a ‘Bad Woman’) told so heavily with an Athenian audience as to outweigh the fact that she is right. Cf. the recent judgement on Lysistrata,that ‘more emphasis is needed on the fact [sic]that any message of the play wouldproblematic because it comes through a woman’ (A. M. Bowie in JHS113 [1993], 168). One thinks of the Sibyls, and the Delphic prophetess, and the Spartan princess Gorgo (Herodotus 5.51.2), and Cassandra in Agamemnon,and the prayer of Jocasta to her sons not to fight (Eur., Phoenissae528ff.), and the prophetic commands of Medea in the Fourth Pythian(13fF.), and even perhaps of the goddess Athena herself; and one wonders whether that fact is really so secure.

¹⁰ Republic3.394dff. argues that the citizens should each have a single and not be influenced by the sinister versatility of tragedy into adjusting to several: (397el–2), and that is why we must not permit tragedy in our ideal city. Presumably that means that tragedy is being seen as a menace to homogeneity. With some subtlety B. Heiden, in Tragedy, Comedy and the Polis(1993), p. 164, argues that Goldhill's mistake is not to see that Athenian citizens were meantto be versatile and ‘The Athenians regarded themselves as typically changing roles and thus as existing in something like a permanent state of transition’: the theatrical experience prepared them for this, and so it did not subvert but reinforce the values of the city. The point is perhaps a little too subtle for a city to have grasped and kept in mind for a century.

¹¹ J. Gould remarks, mildly, on the way that such statements by Longo are ‘framed as a simple and trenchant assertion of established fact’, see Silk, Tragedy and the Tragic,p. 219.

¹² Ethical, I think, rather than political. The question of the status and function of the chorus is interestingly discussed by J. Gould and S. Goldhill in Silk, Tragedy and the Tragic,pp. 217ff., 244ff. Gould points out that choruses, often exotic in personnel and always excluded from authority and action, could not be felt as standing for the authority of the democratic polis; some choruses, composed of foreign slave women, must have been ‘perceived by the citizen body as doubly, or even triply, marginal’ (p. 220). Goldhill replies that choruses nevertheless do speak with ‘the full weight of a collective authority’. Most tragedy is not set in Athens or about Athenians: ‘allresponse to tragedy involves projection, sympathy, idealization—a negotiation of “the other” to find meaning for the self (p. 253).Google Scholar

¹³ Sometimes, for instance, they exhibit such unsocial emotions as panic in the face of the enemy Septem,(parodos), or indifference to the murder of their king by a foreign woman (Medea),or joy at the mutilation of their king by his own mother (Bacchae),or callous indifference to the suffering of the hero (Philoctetes,843–64), or a short-sighted and hysterical rejection of the hero that has to be corrected by a worthier representative of the community (O.C.,parodos).

¹⁴ For criticism of Winkler's suggestion, see Vidal-Naquet, P. in PCPS 32 (1986), 137f.Google Scholar

¹⁵ Kuch, H. advances the view, not easy to accept, that the dithyramb, a backward-looking form, was favoured by ’die konservativen Aristokraten‘, the democrats favouring the progressive form, tragedy (p. 28). Only by some such assumption can political significance be got into the matter of the dithyramb. But B. Zimmermann, Dithyrambos: Geschichte einer Gattung Hypomnemata98 (Göttingen, 1992), 36, has an opposite theory, more in line with the views discussed in this paper: he speaks of the form as possessing a ’demokratische, egalitare Charakter‘. See also Zimmermann in Tragedy, Comedy, and the Polis,p. 43: the dithyramb competition came in with the ten Clisthenic tribes, to strengthen the sense of solidarity, ’das Zusammengehdrigkeitsgefuhr, in the new, democratic, units.Google Scholar

¹⁶ Frogs1021ff. shows that it could be said in a comedy that some tragedies had the effect of encouraging warlike spirit. The Septemcertainly has a martial atmosphere; but the reflective spectator will hardly have interpreted the climactic duel of the doomed brothers as an encouragement to make war.

¹⁷ The chorus of Choephoroeare mourning women, whose physical exertions can hardly have been great. It is true that the binding song of the Erinyes at Eumenides307ff. must have been quite lively; note especially 372–6; and perhaps the satyrs in Proteuswere active.

¹⁸ Aristophanes, Acharnians71–2. On the question of exemption from military service for choreutae, see Csapo, E. and Slater, W J., The Context of Ancient Drama(Michigan, 1995), p. 352CrossRefGoogle Scholar.

¹⁹ The centrality of war to the classical polis (except Sparta) can be exaggerated: see Humphreys, S. C. in JHS91(1971), 191–193; B. Heiden in Trag.Com.Pol,165, n. 40.CrossRefGoogle Scholar

²⁰ For what it is worth, I agree with the authors of the most recent discussion, Csapo and Slater, The Context of Ancient Drama,p. 286: ‘In our opinion the testimony of ancient authors [mostly Plato, I observe] shows clearly that women (and boys) were present in the audience.’

²¹ Croally, Ewipidean Polemic,p. 50.Google Scholar

²² See Griffin, J., Homer on Life and Death(Oxford, 1980), pp. 103–143Google Scholar; Macleod, C. W., Commentary on Iliad 24(Cambridge, 1982), pp. 1–16.Google Scholar

²³ Not only war in Homer brings tears as well as glory.

²⁴ This argument of Goldhill is criticized by R. Friedrich, and defended by Goldhill, in Tragedy and the Tragic,pp. 263ff

²⁵ The view in the DDR was that ‘tragedy included democratic elements’, e.g. H. Kuch in Die gr. Tragödie(above, n. 1), pp. 27ff: specifically, ‘Mit den immanenten Elementen demokratischer Natur wie Diskussion, Verantwortung, Entscheidung, vielleicht schon sozialer Phantasie, entsprach das Dialektisch-Dialogische der Tragödie dem Trend zur Demokratie’ (p. 28).

²⁶ ‘What seems certain is that it was in the sixth century that the festival [the City Dionysia] became important, probably through the policy of Pisistratus’: Pickard-Cambridge, A. W., The Dramatic Festivals of Athens,2nd edn, rev. J. Gould and D. M. Lewis (Oxford, 1988), p. 58Google Scholar. The date of the introduction of competition in tragedies at the Dionysia is unfortunately uncertain (Ibid., p. 103). Thespis is said to have won a prize for tragedy in 535/2 B.C.; Gould and Lewis list it under ‘City Dionysia’ (p. 124). M. L.West points out that these assertions about the very early years of tragedy are not firmly reliable: CQ39 (1989), 251–4. But there seems to be no evidence that tragedies were first introduced by the democracy. An interesting speculative case is made for that, and for a post-tyranny date for the City Dionysia, by Connor, W. R., ‘City Dionysia and Athenian democracy’, C&M 70(1989), 7–32.Google Scholar

²⁷ So Csapo and Slater, The Context of Ancient Drama,pp. 103–4: “The purpose of the new festivals was to foster and display the power of the unified state, centered politically upon the city and ultimately upon the tyrant himself, and to promote a common cultural identity and a system of values consistent with the new political reality.‘ Nothing here about democracy; but otherwise the story is the same. Evidently the Dionysia, and tragedy, could serve democratic and tyrannical regimes and ideologies equally well.

²⁸ E.g. Ferguson, W. S, Hellenistic Athens(London, 1911), pp. 290fGoogle Scholar; Schneider, C., Kullurgeschichte des Hellenismus(München, 1971), vol. II, pp. 188,245. New tragedies continued to be performed, and to be an important part of the festival. ’The competitions went on, and proclamations of honour were made and crowns bestowed at the Dionysia down to the first century B.C.’ (Pickard-Cambridge, The Dramatic Festivals of Athens,p. 82).Google Scholar

²⁹ It is of course not even the case that all tragedies were put on at the Great City Dionysia. Some were staged at the Lenaia, at which the allies were not present to be impressed; though we are not well informed about that, and it seems always to have been the minority option (Pickard-Cambridge, The Dramatic Festivals of Athens,p. 41). There were also many performances at Rural Dionysia: they are known in thirteen demes. See conveniently Csapo andSlater, The Context of Ancient Drama,p. 107. Fans of the theatre could go round from one performance to another: Plato, Republic475d.

³⁰ Goldhill, Cf. S., ‘Battle narrative and politics in Aeschylus’Persae', JHS 108 (1988), 188Google Scholar–93: p. 192, the play ‘may not demonstrate the ironic questioning of an Euripides, but it is not hard to see it investigating attitudes within the polisto the recent victory’. Such attitudes, I think, would be religious, rather than political. For Goldhill, the fact that Greeks are not named, while Persians are, is because of the ‘subsumption of the individual into the collectivity of the polis’(Ibid.). That, too, seems to me to have an important religious element: it is the gods who have done this, not individual Hellenes. We recall the words of Themistocles in Herodotus, 8.109.3: Nowadays the polis and its collectivity gets in everywhere; but Herodotus' Themistocles meant not one polis, with its collectivity contrasted with its individual aristocrats, but the whole of Hellas, as mortal men contrasted with the divine.

³¹ Cf. the memorable rebuke given to Achilles by the wise Nestor, Iliad11.762–4:Cf. also 16.29ff., 18.97fF. I agree with Hainsworth, J. B., The Idea of Epic(California, 1991), p. 8, that the Iliadis ‘about heroism, specifically about heroic honour, and its effect and price’.Google Scholar

³² This was of course the view of Homer taken in antiquity by the poet Horace in Epistles1.2: the Iliadis a better teacher of ethics, and specifically of the control of the passions, than theprofessional philosophers. Quidquid delirant reges, plectuntur Achivi.Cf. the comments of Richardson, Nicholas, Commentary on the Iliad,vol. VI (Cambridge, 1993), pp. 14–19.Google Scholar

³³ Greek Tragedy and the Historian,p. 234. We may add a reference to the statement of Isocrates, that by parading them the city ‘revealed to the rest of Greece the number of the orphans and the disasters caused by this policy of imperialism’, , Isocr. 8 (On the Peace),82.

³⁴ What we do find, of course, is the allegation in comedy that in the good old days of Aeschylus the citizens learnt from tragedy to shut up and not question orders; only with Euripides did horrid questioning come in. So Ar. Frogs1013ff., esp. 1071–3: [Aeschylus] ‘Now, thanks to you, Euripides, the rowers on the war-ships answer back to their officers; when I was alive, all they knew was to call for their grub and shout “Yo heave ho!”’ Also Clouds998f: in the days of the men of Marathon, boys never answered their fathers back. All this questioning was by no means felt to be clearly democratic and laudable. But see Euben, J. P. in Euben, J. P. (ed.), Greek Tragedy and Political Theory(Berkeley, 1986), p. 24: tragedy helped to ‘define, sustain and questiondemocratic culture’ (my italics).Google Scholar

³⁵ Loraux, Cf. N., L'invention d'athenes(Paris, 1981). A different genre, admittedly, but a form intended for essentially the same audience.Google Scholar

³⁶ Except in the Platonic Menexenus,which is not a real example of the genre.

³⁷ Croally, Euripidean Polemic,pp. 44f. Cf., for example, Kuch (n. 2 above), p. 38: ‘Die attische Tragodie wurde so zu einem Organ fur das Selbstverstandnis der Polisburger’, etc.

³⁸ Plato, Apology38a5.

³⁹ Pelling, however (Greek Tragedy and the Historian,pp. 224ff.), accepts this view of ‘ideology as question’. He goes on, ‘This is a very self-critical and self-analysing people. We have already noted Athens’ pride in herself as a home of discourse‘ (p. 229). I think an uncharacteristic looseness of thought mars this passage of Pelling’s distinguished essay.

⁴⁰ Sommerstein, A. in Greek Tragedy and the Historian,pp. 65ff. The theorikonseems to have been a fourth-century innovation, and the charge in the second half of the fifth century to have been two obols: not inconsiderable! For the evidence, see Pickard-Cambridge, The Dramatic Festivals of Athens,pp. 265–268.Google Scholar

⁴¹ English translation: Polity Press, 1993.

⁴² P. Easterling mentions ‘such things as tension between loyalty to the city and loyalty to the family, to mention one of the many conflicting claims that we talk of nowadays as constituting the subject-matter of tragic discourse’ (Trag. Com. Polis,561). Such tensions, of course, are not specifically democratic; they would be no less likely to come up for members of powerful and prominent families under oligarchy or tyranny.

⁴³ The argument of A. T. von S. Bradshaw, that Sophocles‘ Ajax fails in aidos,and that that failure is relevant to the relations of Athens with her allies (in D. C. Pozzi and J. M. Wickersham [edd.], Myth and the Polis[Cornell, 1991]) is well criticized by C. Sourvinou-Inwood, CR46 (1996), 82.Google Scholar

⁴⁴ Or that, for instance, of the Athenian general Eurymedon at Corcyra, Thuc. 3.81.4: ’In the seven days that Eurymedon was there with his sixty ships, the Corcyreans butchered their enemies‘.

⁴⁵ It is suggested by Osborne that the competition between tragedies was itself important: ’Just as the city's promotion of other forms of competition implies a confidence that the political effects of promoting particular individuals can be controlled, so the city's promotion of dramatic competition implies a confidence that the political effects of promoting issues can be controlled‘ (Tragedy, Comedy, and the Polis,pp. 34f.). The point is an interesting one for Athenian social history, but it does not greatly illuminate the plays themselves.

⁴⁶ Theseus stays within his own country and repels invaders in Oedipus at Colonus;he arrives after the killing of the usurper in Heraclesand offers the shattered hero shelter in Attica. Aegeus, in Medea,similarly has n o trace of violence or aggression; he simply offers the heroine sanctuary (an offer of course, which she will abuse; but that is n o t his fault). Nowhere d o these edifying Athenian rulers show a trace of imperialism.

⁴⁷ 489, 491. This prattle about tyranny, says he, is absurd. Kingship versusdemocracy: Aeschylus, Persians241 ff., Eur. Suppl.399ff. with C. Collard's Commentary (Groningen, 1975). Collard notes, following J. de Romilly, that in the closely similar discussion in Herodotus 3.80–3, all three alternatives are mooted and discussed: democracy, monarchy, and aristocracy. Tragedy prefers to omit one, and the one more relevant to contemporary Athenian politics. Plato, in Republic8, takes as basic all three forms: monarchy, aristocracy, and democracy.

⁴⁸ See n. 7 above.

⁴⁹ It is a question whether there actually was cult of Oedipus at Colonus, outside the pages, or the influence, of Sophocles. There is a dearth of evidence; there is another reported cult in Athens, below the Areopagus; and the insistence in the play that nobody except the king and his successors in each generation should know the place of the grave is hard to reconcile with any public cult. F. Pfister, in his classic work Der Reliquienkult im Altertum(Berlin, 1909), pp. 1 lOff., is confident that Sophocles cannot have invented the Colonus cult. Kearns, E., The Heroes of Attica (BICSSupplement 57, London, 1989), p. 50, is more sceptical but concludes her thorough discussion with the judgement: ’I am inclined to feel that there is a little too much evidence connecting Oedipus with Kolonos to be explained simply by the influence of Sophocles‘ play' (p. 209).Google Scholar

⁵⁰ Tragedy and the Tragic,p. 175.

⁵¹ That cult of Philoctetes is mentioned in his play is suggested by Harrison, S. J., JHS 109 (1989), 173–175. His argument is that w. 1418–22, spoken from the machina by Heracles, which promise Philoctetes a glorious existence like his own after all his sufferings, ’subtly suggest‘ posthumous cult. If so, this cult must be one to which Appian once alludes, ’on an island near Lemnos‘; this island must be Chryse, a few miles from Lemnos; Lemnos itself had been acquired for Athens by Miltiades and was an Athenian possession. Thus ’there is at least some possibility that this cult of Philoctetes on Chryse existed in the fifth century B.C. and was known to Sophocles and his audience’ (p. 175). The argument is subtle, but several steps in it are speculative, and the supposed allusion is fleeting and elusive; to make it a central plank in the interpretation of the play must be very bold. And even if it is accepted, the awkward fact remains that Sophocles has placed such allusion as there is to cult entirely on the recipient. There is no mention of the worshippers or of ‘collectivity’.CrossRefGoogle Scholar

⁵² No cult, for instance, at the end of Sophocles‘ Oedipus Tyrannus, Electro,or Antigone(an eminently impossible heroine); nor of Alcestis, Andromache, Hecuba, Troades.Other plays end with predictions that are not of cult: the vote of Athena (Eur., Electra1267), or the origin of the place name Oresteion (Orestes1647), or the future glories of the descendants of Ion (Ion1571ft), or the future sufferings of Hecuba and Agamemnon (Hecuba1259ff.). Prophecy of something, rather than establishment of cult, seems to be the point. And even when a ritual is announced, it may give no key to the action of the play: the Corinthian ritual in memory of Medea's children (Medea1381) does very little to help us understand Jason and Medea (both, of course, still alive).

⁵³ It is an obvious difficulty that the cult made possible at the end of the Bacchae,and partly lost in the lacuna, is not of the hero Pentheus but of the god Dionysus. Seaford tries to meet it by suggesting that Pentheus, too, is to receive cult (Reciprocity and Ritual,pp. 295, n. 68, and 312, n. 124): ‘Pentheus is explicitly connected with cult only in Pausanias 2.2.7’, but ‘Pentheus may in fact have received cult along with Dionysus’. At Paus. 2.2.5–6 what we read is that outside Corinth a shrine contains two xoanaof Dionysus, one named Lysios, the other Baccheios, made of the wood of the tree which Pentheus climbed to spy on the Maenads. The Delphic oracle told the Corinthians to find that tree and to honour it like the god: that is why the statues were made of it. There is no suggestion of honour being paid to Pentheus.

⁵⁴ Cambridge Ancient History, 2nd edn (Cambridge, 1992), Vol. V, pp. 75.Google Scholar

⁵⁵ Tragedy and the Polis,p. 34. Csapo and Slater assert (p. 105), that in refusing a poet a chorus ‘factors other than quality control clearly [sic]came into play’; but they produce no evidence for this but Cratinus, PCGf. 17, which seems to me incapable of proving it.

⁵⁶ Poetics1449b27,1453b12.

⁵⁷ On this tradition, not recorded in evidence as early as we should ideally like, see e.g. Merkelbach, R., Untersuchungen zur OdysseeZetemata 2 (München, 1969), pp. 239–262Google Scholar; West, S. R. in the Oxford Commentary on the Odyssey,vol. I (Oxford, 1988), pp. 36–40.Google Scholar

⁵⁸ On this point I have a lot of sympathy with the challenging book of Malcolm Heath, The Poetics of Greek Tragedy(Duckworth, 1987); but in my view he goes too far in the opposite direction, allowing too little space for anything other than emotion and pleasure in his analysis of tragedy. The plays do contain a serious vision of the world and of the relation of men and gods; that cannot simply be disposed of by declaring that the taste for such things is a mere modern fad (p. 78). But he is right to say that tragedy, like other forms, may be ‘political’ in the sense of glorifying the polis(p. 64), not necessarily in any other, and he speaks well of tragedy as providing ‘intense but ordered emotion, controlled not by intellectual interests, but by the coherence of the whole simply as an emotional experience, by the aesthetic satisfaction which the audience receives through its experience of the emotions as an ordered sequence’ (p. 80).

⁵⁹ ‘Tragedy could claim to be the true inheritor of epic, and to have discernible links, too, with the choral traditions of the wider Greek world, at the same time as being a truly Athenian invention’ (Easterling, Greek Tragedy and the Historian,p. 25).

⁶⁰ See J. Herington's important book, Poetry into Drama(California, 1985).

⁶¹ See Griffin, J., ‘The Epic Cycle and the uniqueness of Homer’, JHS 97 (1977), 39–53.CrossRefGoogle Scholar

⁶² See Romilly, J. de, L'evolution dupathetique d'eschyle a Euripide(Paris, 1961).Google Scholar

⁶³ See the classic article by Gould, John, ‘Hiketeia’, JHS 93 (1973), 74–103.CrossRefGoogle Scholar

⁶⁴ CAH,2nd edn, vol. VI: The Fourth Century(Cambridge, 1994). That well-mannered volume says only (p. 35): ‘Critias had enough armed support at his disposal to convince the Council and force Theramenes’ execution.‘ We observe how neatly the colour, the particularity, the sense of period and place, have all been drained away; we might be reading about events of any place, at any time. That is not the way to get an insight into the history of the period, as it presented itself to the minds of those who made and endured it. The flight of Themistocles, similarly, is described, in Vol. V: The Fifth Century(Cambridge, 1992), p. 65, without reference to his enforced supplication of the queen of the Molossians (simply, ‘Admetus of the Molossi refused either to surrender him or to allow him to stay’). Those details, occurring in a first-rate source, and very fevealing for the mentaliteof the period, are presumably left to the students of ancient religionand literature. Clio, the austere Muse of serious modern historiography, draws her skirts aside; she prefers, she says, a diet of very dry bread.

⁶⁵ Locus classicus:Pausanias 1.15. It was the local heroes Phylacus and Autonous who put the Persian raiders to flight at Delphi: Herodotus 8.38–9. ‘It is not we who have done this but the gods and the heroes’, said the Themistocles of Herodotus after the Battle of Salamis (8.109.3): the heroes whom the Greeks had invoked before the battle (Hdt. 7.189; Simonides fragment 3 West).

⁶⁶ The Plataeans make all the ‘tragic’ appeals: to the rights of suppliants, to the graves of ancestors, and to the altars of gods (3.58–9). The Melian Dialogue, of course, is conducted in very different terms.

⁶⁷ Thucydides 3.52–69; 5.84–116.

⁶⁸ Some have seen a connection between these events and Euripides' Suppliant Women;for references and discussion, see C. Collard's edition of the play (Groningen, 1975), Vol. I, pp. 8ff.

⁶⁹ Lysias 2.7–10; Isocrates, Panegyricus54–8; id., Panathenaicus168–74; Demosthenes 60.8. See N. Loraux, The Invention of Athens(Eng. trans., Harvard, 1986),Google Scholar

⁷⁰ Cf. J.P. Vernant, ‘Le tyran boiteux: d'oedipe à Periandre’, in J.P. Vernant and P. Vidal-Naquet, Mythe et tragedie deux(Paris, 1986), pp. 45–78; K. Reinhardt, ‘Herodots Persergeschichten’, Vermachtnis der Antike(Gottingen, 1960), 133–74 Herodot,ed. W Marg (W.d.F.26, 1965), pp. 320–69Google Scholar; E. Wolff, ‘Das Weib des Masistes’, Ibid., pp. 668–78 Hermes92 (1964), 51–8.

⁷¹ It will be recalled that Thucydides makes Pericles speak of ‘the many rest periods from exertion’ offered to her people by the city, with its ‘competitions and sacrifices through the year’ (2.38.1); doubtless he included the tragic performances. Cf. Plato, Laws,653cd.

⁷² The ideas in this paper began to be worked out for the Stubbs Lecture, given at University College, University of Toronto, at Easter 1996. I am grateful for the invitation, and for helpful discussion of the paper.