reading Qss

The Trojan War Author(s): M. I. Finley, J. L. Caskey, G. S. Kirk and D. L. Page Source: The Journal of Hellenic Studies, Vol. 84 (1964), pp. 1-20 Published by: The Society for the Promotion of Hellenic Studies Stable URL: http://www.jstor.org/stable/627688 Accessed: 26-02-2018 13:57 UTC

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide

range of content in a trusted digital archive. We use information technology and tools to increase productivity and

facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at

http://about.jstor.org/terms

The Society for the Promotion of Hellenic Studies is collaborating with JSTOR to digitize, preserve and extend access to The Journal of Hellenic Studies

This content downloaded from 128.172.48.178 on Mon, 26 Feb 2018 13:57:30 UTC All use subject to http://about.jstor.org/terms

THE TROJAN WAR

MR FINLEY'S article is an elaborated version of a talk first broadcast in October 1963. It was offered for publication with the intention of stimulating discussion of a problem which has been exercising archaeologists and historians. For this reason its author does not here answer the provisional criticisms and comments offered by Professor Caskey, Mr Kirk and Professor Page which are also printed below. It is hoped that this presentation will help to define for readers the very varied issues which attend the problem of the date and nature of the Trojan War.

I. THE TROJAN WAR,1 by M. I. Finley

In concluding the first chapter of his Troy and the Trojans, Professor Blegen writes (p. 20): 'It can no longer be doubted, when one surveys the state of our knowledge today, that there really was an actual historical Trojan War in which a coalition of Achaeans, or Mycenaeans, under a king whose overlordship was recognised, fought against the people of Troy and their allies.' Whatever 'the state of our knowledge today' may be, or may be taken to mean, one must insist that there is nothing in the archaeology of Troy which gives the slightest warrant for any assertion of that kind, let alone for writing 'it can no longer be doubted'. Blegen and his colleagues may have settled, insofar as such matters can ever be determined with finality by archaeology, that Troy VIIa was destroyed by human violence. However, they have found nothing, not a scrap, which points to an Achaean coalition or to a 'king whose overlordship was recognised' or to Trojan allies; nothing which hints at who destroyed Troy.2

Mainland Greek archaeology and the Mycenaean tablets are equally devoid of any information on that central question. What is effectively new in the state of our knowledge today, as against the state of knowledge two generations ago, is the tangential, but nonethe- less important, testimony in documents from the world outside the Achaeans and Trojans, and a radically new appreciation of the nature and techniques of oral poetry. But the base of the whole structure of current belief about the Trojan War obviously remains the Iliad and Odyssey. That is a platitude, but it needs to be reasserted and underscored, and I propose to argue that we have not advanced very far in a rigorous, critical assessment of the poems as evidence for the historical narrative of the Trojan War; that all statements of the order of Professor Blegen's 'the tradition of the expedition against Troy must have a basis of historical fact' are acts of faith not binding on the historian; that there is evidence which, though far from decisive, at present weighs the balance the other way.

The first problem of analysis is an operational one. Everyone is agreed that the Iliad as we have it is full of exaggerations, distortions, pure fictions and flagrant contradictions. By what tests do we distinguish, and, in particular, do we decide that A is a fiction, B is not (though it may be distorted or exaggerated) ?

There is, of course, a first test which we all apply: we eliminate as pure fiction the scenes on Olympus, the divine interventions and all the rest of that side of the story. Yet I am

1 The main argument of this paper was first 12, 51) or of the sunken pithoi. Even if one accepts presented, naturally in very different form, on the Blegen's not wholly convincing deduction (Troy and Third Programme of the BBC on October 24, 1963 the Trojans I56) that the pithoi show 'that there was and then published in the Listener on November 7th. an emergency of some kind', they reveal nothing

2 I hope no one will remind me of the single about the source of the danger. bronze arrowhead found in Street 710 (Troy iv (1958) VOL. LXXXIV B

This content downloaded from 128.172.48.178 on Mon, 26 Feb 2018 13:57:30 UTC All use subject to http://about.jstor.org/terms

not being frivolous when I suggest that this is at best an equivocal first step, one which makes the rest of the operational analysis more difficult. The 'Homeric' picture of the gods is admittedly widely divergent from the thirteenth-century one, both in its omissions and in its innovations. Many of these divergences touch the core of religious belief and of ritual. By what reasoning do we permit oral transmission so much latitude with the supernatural side of the story while denying it equal freedom with the human side ? The answer is that we impose our own evaluation of what is and what is not credible on the ancients. We treat the human side of the tales as possible fact, the supernatural side as certain fiction. But did the bards and their audiences (and many Greeks in later times)-the men who were doing the transmitting and the manipulating-draw this distinction ? Were the scenes on Olympus less 'real', less 'factual', to them than the miracles of the Bible are to many today? The operational analysis must work with their conceptions in these matters, not with ours; that is why I suggest that the human-supernatural test is a misleading one.

How much latitude of divergence we allow is the decisive question. Everyone allows a good deal, but nearly everyone then stops short and agrees that 'the tradition of the expedition against Troy must have a basis of historical fact'. In the absence of literary or archaeological documentation, there is no immediate control over this will to believe. But there are oblique ways of getting at the possibilities, first by examining three other heroic traditions which we can check. There is a difficulty here because these others developed

mentation which may have acted as a contaminating influence. I shall return to that point briefly at the end; now I shall look at the Song of Roland, the Nibelungenlied, and the South Slav traditions about the battle of Kossovo as if the specific difference of total illiteracy did not exist.

In the year 778 Charlemagne invaded Muslim Spain. On the way home the rear of his army was ambushed and massacred at Roncevaux in the Pyrenees by the Basques, who were Christians. The incident was humiliating but without long-term significance. It is mentioned briefly in several chronicles of the age and that should have been the end of it. Instead, the incident, or rather Count Roland, one of the men who fell, burgeoned into an heroic tradition all over Europe, one which is still alive in very odd ways. It was brought to Sicily by the Normans, and even today in Sicily there are puppet shows about Roland and the other paladins of Charlemagne, and the same scenes appear on their decorated donkey-carts. Roland competes in peasant culture with the Sicilian Vespers and Garibaldi. The latter are obviously appropriate, Roland just as obviously is not-except as a champion of Christendom against the infidel, a completely unhistorical role into which he had been transformed at a date which cannot be fixed precisely. The earliest known text of the Song of Roland is a 4,00o-line poem written about I 150. By then the ambush at Roncevaux had become an heroic battle of the paladins of Charlemagne against a Saracen host of 400,000 led by twelve chieftains, some of whom had Germanic or Byzantine names. The courtly atmosphere of the poem is not that of Charlemagne but rather that of the First Crusade, whereas the political geography fits neither period but the tenth century. In sum, the poem seems to' have retained precisely three historical facts about Roncevaux and no more: that Charlemagne led an expedition into Spain, that the expedition ended in disaster, and that one of the victims was named Roland.3

In the approximately contemporary and also widely travelled Nibelungenlied, of the central characters Gunther and Atli-Etzel (Attila) are historical but had no actual relation- ship with each other. Gunther was king of the Burgundians on the Rhine from 411 to 437, when he was killed by invading Hun mercenaries in the Roman imperial service. The Hun

3 For all this, see P. LeGentil, La Chanson de Roland 'The Fall of Troy', Antiquity xxxvii (I963) 6-i i, to (Paris, 1955) chs. i-iii. The Roland tradition is also support his argument that 'Homeric Troy' is Troy VI. used, I think in the wrong way, by C. Nylander,

M. I. FINLEY 2

This content downloaded from 128.172.48.178 on Mon, 26 Feb 2018 13:57:30 UTC All use subject to http://about.jstor.org/terms

THE TROJAN WAR

kingdom, of which Attila did not become ruler until 445, was not involved. The Nibe- lungenlied turns the invasion of Burgundy by the Huns into its reverse, a complicated move initiated by the Burgundian princess Kriemhild, wife of Attila. For this there is no basis whatever; nor is there for the existence of Kriemhild, who is the one character tying the whole epic together, or of Siegfried and Brunnhilde, the key figures in the first half. On the other hand, Gunther and Attila are drawn into contact with the Ostrogoth Theoderic, disguised as Dietrich von Bern, who ruled most of the western Empire from 493 to 526, with Piligrim, bishop of Passow from 971 to 991, and with many minor figures, equally anachro- nistic or fictitious. The Nibelungenlied, in sum, retains even less recognisable or coherent history than the Song of Roland, if it can be claimed to retain any at all.4

The South Slav heroic tradition about Kossovo, a really decisive battle, is in some respects more securely anchored in history than either the French or the German, though it is more difficult to control since it remains scattered in collections of shorter poems, never (except artificially) brought together in one long composition. The Ottoman invasion and the shattering defeat of the Serbs under Prince Lazar at Kossovo in 1389 remain fixed in the tradition. But then the variations and inventions begin, of which only two need be mentioned: the conversion of Lazar's son-in-law and chief support, Vuk Brankovic, into a traitor (perhaps under the influence of the Song of Roland, which Serbs in later times would have learned in Ragusa), and the heroisation of Marko Kraljevic ('the uncrowned king of heroic poetry'), a curious figure, unimportant in real life, who certainly did not fight on the Serbian side at Kossovo and who seems to have accepted Turkish suzerainty quite cheerfully both before and after.5

We must therefore reckon with three possibilities of fundamental distortion (apart from pure invention): (i) that a great heroic tradition may be built round an event which itself was of minor significance; (2) that the tradition may be picked up by regions and people to whom it was originally, as a matter of historic fact, utterly alien and unrelated; (3) that the tradition may in time distort (not just exaggerate) even the original kernel so that it is neither recognisable nor discoverable from internal evidence alone. I am suggesting not that all three always happen, but that they may, and sometimes do, occur. Working backward, how can we tell? The 'facts' in the Song of Roland, the South Slav songs and the Nibelungenlied all look alike. They bear no stigmata which distinguish the wholly fictitious from the partly fictitious. Suppose all documentation about the period of Charlemagne were lost: How should we then be able to determine which bits of the Song of Roland were historical, which not? How could we know whether the battle of Roncevaux was or was not fought against the Muslims ? Indeed, how could we know that there had been a battle there at all? It is only from external evidence that we know how to answer the last two questions. Schematically stated, the Song of Roland has the right battle but the wrong enemy, the Nibelungenlied has the right enemy but the wrong battle (and a wildly wrong battle-site), whereas the Serbian tradition has them both right.

Archaeology has settled the question, Was there a Trojan War? It has failed to suggest who attacked and destroyed Troy. Which model shall we then follow among the heroic traditions? I submit that the possibility cannot be ruled out that the Iliad (and the whole

4 It does not matter for my purposes which school weitgehende Umdeutungen, ohne Gewaltsamkeit of Nibelungenlied scholarship one prefers; see either und inneren Krampf durchzufuhren and blieben A. Heusler, Nibelungensage und Nibelungenlied (5 ed., damit unbefriedigend.' Darmstadt, 1955), or F. Panzer, Das Nibelungenlied 5 See M. Braun, Das serbokroatische Heldenlied (Stuttgart, I955), esp. chs. vii-viii. It is not without (G6ttingen, I96I) ioo-2; D. Suboti6, Yugoslav malice aforethought that I quote the latter's final Popular Ballads (Cambridge, I932) ch. ii. Subotic sentence (p. 285) dismissing all efforts to find writes (p. 87): 'It remains a mystery why the Yugo- historical roots for Siegfried and his family: 'Die slav heroic poetry should have made him [Marko] Verselbigungen von Personen und Vorgangen des out to be the greatest national hero, while converting Epos mit geschichtlichen waren doch nirgends ohne Vuk Brankovic into a traitor.'

3

This content downloaded from 128.172.48.178 on Mon, 26 Feb 2018 13:57:30 UTC All use subject to http://about.jstor.org/terms

Greek tradition) is wrong on this question; that the possibility must be seriously considered that the better analogy is with the Roncevaux tradition rather than with the Kossovo tradition. The fact that the Greeks themselves accepted the historicity of the tradition has no probative value. It is impermissible to defend the tradition on the ground that men like Thucydides 'may well have based their beliefs on a greater body of surviving oral and written evidence than that which has come down to us' (my italics).6 They had no written evidence whatsoever, and the validity (not the quantity) of the oral 'evidence' is precisely the point at issue. All Europe once accepted the Roland tradition as history, too. For centuries there was neither wish nor motive to challenge or check the tradition. 'Historical consciousness', in Jacoby's words, 'is not older than historical literature.'7 By then it was too late. All that Thucydides could do was sit down and think hard about the tradition. We can do better, thanks to archaeology and thanks to the written evidence from Egypt, North Syria and the Hittite archives.

The next line of investigation is to look at the documents. On the Hittites I need hardly do more than summarise certain of the conclusions in Page's History and the Homeric Iliad: (i) the Achaeans are mentioned in some twenty texts ranging from the late fourteenth to the end of the thirteenth century B.C.; (2) the Achchijawa with whom the Hittite rulers were concerned was not across the Aegean on the mainland but near at hand, an inde- pendent island or coastal state, most likely based on Rhodes and possessing some territory of its own in Asia Minor; (3) Troy is absent from the extensive Hittite archives save for one possible reference; and (4) a fortiori the archives provide no direct information on the relations, if any, between Troy and Achchijawa. One text from the final half-century of the Hittite Empire reports the rise of a kingdom of Assuwa in western Asia Minor which led a serious, but unsuccessful, coalition war against the Hittites. The southernmost member of the coalition was Lycia, the northernmost may have been Troy. A second text mentions both Assuwa and Achchijawa, but it is too fragmentary to be intelligible. Professor Page has devoted the third chapter of his History and the Homeric Iliad to a most ingenious and intricate reconstruction of this complex situation, the 'background of the Trojan War', from which there emerge two vital suggestions: that Assuwa and Achchijawa were eventually brought into direct conflict with each other, and that Troy VIIa fell in this context.

From this or any other reconstruction of the few relevant Hittite texts the only conclusion one could possibly draw is that the Trojan War was an exclusively Asiatic affair. It is the Iliad which causes trouble. Page writes about his reconstruction (p. i i i) that between the Hittite annals and the Iliad 'there are large and obvious differences. The Iliad's league of natives is led by Troy, not Assuwa; and the Achaeans who attack them are not the Achaeans familiar to the Hittites .. but an expeditionary force from the mainland.' The question is thus squarely put. Professor Page says 'these are differences, not disagreements'. I prefer the alternative view, that, to return to my analogy, they are fundamental disagreements exactly like those between history and tradition over who fought at Roncevaux. Page concedes that this is a serious possibility, which he then rejects because of the Catalogues in the second book of the Iliad. Before turning to them, however, I want to consider the so-called Sea Peoples.8

The last half-century of the Hittite Empire was filled with rebellions and wars in Asia Minor, but the Empire was actually destroyed, by 1 200 or 1I90, by invaders from the north. By 1190, too, Troy had fallen; so had most of the great fortresses in Greece and important

6 Blegen et al., Troy iv 10. enter into any controversial matters except on the 7 F.Jacoby, Atthis (Oxford, 1949) 201. identification with the Achaeans of the Akiyawasa 8 See P. Mertens, 'Les Peuples de la Mer', Chr. or Akawash of the Merneptah stele, on which see

d'Eg. xxxv (1960) 65-88. In what follows I shall cite Page, op. cit., 21 n. i. neither sources nor modern literature as I do not

M. I. FINLEY 4

This content downloaded from 128.172.48.178 on Mon, 26 Feb 2018 13:57:30 UTC All use subject to http://about.jstor.org/terms

THE TROJAN WAR

local states in northern Syria like Ugarit and Alalakh; there was turbulence in the west, in Italy, Sicily and Libya; there were repercussions as far east as Babylonia and Assyria. It would be going too far, on present evidence, to link all this widely scattered activity into a single unified operation, but there is a case for thinking that a significant, perhaps the main, generating impulse was a massive penetration over a longish period by migrating invaders from the north. A number are named in Egyptian texts. Identification of the various peoples remains highly controversial, but in any event it would be wrong to believe either that the 'Sea Peoples' were a coherent, firm coalition moving in a single sweep, or that the Egyptian lists are either complete or wholly accurate. 'Northerners coming from all lands', says the Merneptah stele. The evidence suggests to me the analogy with the Germanic migrations into the Roman Empire: broken in rhythm, confused in the interrelationships among the migrants, confused even in the motives. Like the Germans, too, these northerners, when they had finished, had considerably altered both the ethnic composition and the political situation in a large area from western Asia to the central Mediterranean.

Given this context, the most economical hypothesis is that Troy VIIa was destroyed by, or in association with, the marauding northern invasions.9 This is no more speculative an hypothesis, after all, than Professor Page's, for no Hittite text says that Achchijawa and Assuwa came to blows and no Hittite text says that Troy fell at the hands of the Achaeans or anyone else. The documentation about the northern invaders-it is really necessary to stop employing the misleading term 'Sea Peoples'-is still very thin, but as it grows, the scale and range of their destructive activity grow apace. Even the Hittites, it begins to appear, may have been affected, though perhaps only indirectly, decades before their Empire was actually smashed.10 This could scarcely have been guessed from the Hittite archives, and it is therefore no objection to my hypothesis to note the lack of textual evidence regarding Troy, for which there is no documentation of any kind.

Neither hypothesis requires, or indeed allows for, a mainland Achaean coalition whereas both provide a proper historical context and a motivation for the siege and destruction of Troy, which the Greek tradition does not."l This question of motive is customarily, though uneasily, pushed aside. Presumably no one any longer accepts the rape of Helen as a sufficient cause of the Trojan War. But what are the possible alternatives which would explain a large-scale attack from the mainland? Troy was a powerful fortress. No ordinary booty raid, like those described by Nestor in II. xi 670-84 or by Odysseus in Od. ix 39-42 and xiv 229-85, would have had a hope of being effective, and

9 Neither this suggestion nor some of the argu- ments which follow are new; see, e.g., briefly A. Heubeck in Gnomon xxxiii (196 I) I 5; Nylander, op. cit. (who unnecessarily complicates matters by an unimpressive dating argument, among other things); C. G. Starr, The Origins of Greek Civilisation (London, I962) 66 n. 3.

10 See H. Otten, 'Neue Quellen zum Ausklang des Hethitischen Reiches', MIitt. d. Deutschen Or.-Gesellsch. zu Berlin xciv (1963) 1-23; cf. briefly C. F. A. Schaeffer, Ugaritica iv (1962) 39-41.

11 The need to fit the Trojan War and the events of mainland Greece into a general eastern Medi- terranean context is persistently overlooked. Thus, Mrs Vermeule, in her review in Gnomon xxxv (1963) 495-9, rightly criticizes the new CAH chapter (ii ch. 36), 'The End of Mycenaean Civilization and the Dark Age', by V. R. d'A. Desborough and N. G. L. Hammond, because neither author 'really grapples with the question of their relation to con-

temporary destructions in the east'. Her own article, 'The Fall of the Mycenaean Empire', Archaeology xiii (1960) 66-75, makes a serious attempt to do so and comes to very different conclusions from mine on the central question, largely, I believe, because she does not abandon the Greek tradition, even to such pseudo-problems as trying to reconcile the archaeology with the tradition of 'the mutual exhaustion of the Trojan War'. On this general question see Starr, op. cit., 66-8. Desborough's The Last Mycenaeans and their Successors (Oxford, I964) appeared just as this article was going to press. It seems to contribute nothing new to this particular discussion. His conclusion that the Trojan War took place between I250 and I230 is based not on any archaeological evidence for these two decades but on the argument that, if the tradition is to be preserved, no other dates are compatible with the archaeology (pp. 220 f., 249).

5

This content downloaded from 128.172.48.178 on Mon, 26 Feb 2018 13:57:30 UTC All use subject to http://about.jstor.org/terms

no booty raid on the necessary scale can be exemplified, so far as I know, nor, in this instance, given any plausibility (as I shall argue later). As for a commercial war, I frankly refuse to take the idea seriously until one of its proponents offers a reasonable explanation why mainland Achaeans should have organised and mobilised themselves on a great scale in order to destroy a centre to which they had long been sending a continuous supply of pottery and from which, we are told, they received horses, which they needed, wool, which in fact they may not have needed, and perhaps gold.

On any hypothesis, I am confident that the explanation of the Trojan War must be either political (in the sense in which war was incessantly being waged in Asia Minor and the Near East for political reasons) or 'accidental' (that is to say, external) incursions into the area from outside for reasons which must be sought (and cannot be found at present) to the north. Neither kind of explanation excludes the possibility of an Achaean share in the operation (as distinct from an Achaean initiative or monopoly). Page has conjectured a political struggle between Achchijawa and Assuwa. I prefer the hypothesis that Achaeans joined a marauding force of northerners, just as they had been part of the mercenary force engaged by the Libyans when they attacked Egypt in the reign of Merneptah (1220 B.C. or thereabouts). We do not know who those Achaeans were or where they came from, and I have no suggestion to make about the Achaeans who, on my speculation, shared in the destruction of Troy. They could have come from Asia, from the Aegean or from the Greek mainland. The essential point is that the half-century or more of migration, invasion and marauding was one of general disruption, precisely like the age of the Germanic migrations, during which allegiances and alliances were shifting and blurred. It would be an obvious guess that, when their own society was under such severe pressure, bands of Achaeans took to buccaneering and mercenary service, sometimes as allies of the invaders. The Merneptah stele makes it unnecessary to guess, as does, in a different way, the career of Attarssijas, so dramatically described in Page's third chapter. (If new texts should confirm Otten's recent suggestion of a direct link in Alasija (Cyprus) between the 'Sea Peoples' and the marauding of Attarssijas,'2 I should feel myself on very firm ground indeed.) The invaders themselves, again like the later Germans, ultimately sought to settle, but on the way they looted and burned, detoured, played the mercenary, as circum- stances directed. For them to smash Troy, with or without Achaean supporters, is an altogether different, and far more intelligible, manoeuvre than the Homeric tale.

The archaeology is not inconsistent with a smash-and-grab raid, though an unusually devastating one. Life was then resumed in Troy: the citadel was reoccupied, 'new houses were superposed over the ruins of their predecessors', 'the fortification wall evidently still continued to stand, or was repaired'.'3 It is only in Troy VIIb 2 (which Professor Blegen puts 50 or 60 years after the destruction of VIIa) that we find novel architectural features and the Knobbed Ware which points unmistakably across the Hellespont. Does that mean that a foreign population did not enter Troy until then? The question is at present unanswerable because we do not yet properly understand the significance of Myc. IIIC pottery, which far exceeds IIIB in number in Troy VIIb i. In considering the same problem for the Greek mainland, where the uniform IIIB style of the great centres was replaced by locally varied IIIC after their destruction, and only later by proto-Geometric, Desborough concludes: 'It might be argued that some one of the variations of the new pottery style should belong to newcomers, and this is not impossible, though it is not provable, as in each district L.H. IIIC pottery seems to be clearly linked at the outset with the preceding style.'14 In sum, the pottery finds in Troy are in the present state of our knowledge compatible with any explanation of the Trojan War. If the Philistines could 12 Op. cit., 2I. 14 Op. cit., 5-6. 13 Blegen, Troy and the Trojans I65-6 and ch. viii

generally.

M. I. FINLEY 6

This content downloaded from 128.172.48.178 on Mon, 26 Feb 2018 13:57:30 UTC All use subject to http://about.jstor.org/terms

THE TROJAN WAR

sit down at once to make Myc. IIIC pots so could new occupants of Troy, especially if potters survived the attack and went on working, as they obviously did in many other places. Alternatively, if the attackers moved on after wreaking all the damage they could, then the continuity is no problem at all, nor is the probability of a further incursion, this time for permanent settlement, 50 or 60 years later.

There is one archaeological argument, however, which, in my view, is more compatible with my hypothesis than with any other, and that bears on the date of the destruction of Troy VIIa. Blegen and others place it near the middle of the thirteenth century (Blegen himself tending to take it further back all the time, even to 1270), arguing from Furumark's chronology of the pottery and from a hypothetical tempo in the development and change of the relevant styles, IIIB and IIIC. Most archaeologists, I believe, now tend to reject both the reasoning and the date. Mrs Vermeule, for example, has put the matter squarely: 'It must be emphasised that the general character of the pottery in all these destruction levels is similar, whether at Troy and Ugarit or at Mycenae and Pylos. There is real difficulty in making any distinctions of date among them.'l5 Imported IIIB pottery was still current in Ugarit and Alalakh when they were destroyed about 190 by the northern invaders.16 And so it was in Troy, too, to be followed there by the immediate emergence of IIIC, and that argues for a date nearer I 190 than I250 for the fall of Troy VIIa.

It is obviously very convenient for my argument to get the destruction of Troy down in date into the heart of the invasion period. But it is not altogether essential, at least not for my rejection of a mainland coalition. Suppose the dates are moved back, provided they are all moved together as they must be. The argument from motive would still stand. On any dating, is it reasonable to imagine that, just when the Achaean states of the mainland were faced with grave difficulties and even total destruction at home, they would take it into their heads to join forces in a wild and risky venture overseas, committing their man- power to go after booty, captive women or whatever ? It is surely more reasonable to think that when their own world was threatened, bands of Achaeans left to join the marauders in the search for booty or new 'homes or just escape and hope (provided one feels the necessity of getting some mainland Achaeans into the story at all, which I do not much care about one way or the other).17

And now, finally, the Catalogues. As part of the Iliad they are a mess on any interpreta- tion. Again I need not go into details-about the central role of the Boeotians, the irreconcilable conflict between Catalogue and narrative over the kingdoms of Agamemnon, Achilles and Odysseus, the numerous disagreements in other matters-since they are all laid out in the fourth chapter of Page's History and the Homeric Iliad. I agree fully that the Catalogues and the narrative in the Iliad as we have it developed separately in the oral tradition and were eventually joined mechanically at a time when they had acquired their many contradictory and irreconcilable elements. There is only one question to be con- sidered: Does the Achaean Catalogue, for all its distortions and fictions, retain a large, hard core of Mycenaean reality which compels us to believe in the existence of a mainland coalition against Troy?18 Page and others answer in the affirmative, essentially on the single argument that a substantial number of the place-names fit known Achaean sites and that a small but still substantial number were gone in post-Mycenaean times and were 15 Op. cit., 68. phase (by or about 1200). If that distinction is 16 W. C. Hayes et al., 'Chronology', rev. CAH i untenable, as other experts on the pottery say, then

ch. 6 (I962) 67-8 (Rowton), 75-6 (Stubbings); cf. the whole structure of his chronological argument Desborough, op. cit., I2. falls. 17 Blegen is acutely aware of this difficulty and he 18 I do not propose to waste time on other, fanciful,

tries (Troy and the Trojans I63-4) to get round it by possibilities, such as the existence and preservation of dating the destruction of Troy VIIa midway in 'the a written Order of Battle; see Page's review of ceramic phase IIIB' (about I260), the destruction of Jachmann, Der homerische Schiffskatalog und die Ilias, the mainland centres 'toward, or at the end of', the in CR, n.s. x (I960) I05-8.

7

This content downloaded from 128.172.48.178 on Mon, 26 Feb 2018 13:57:30 UTC All use subject to http://about.jstor.org/terms

unknown to Greeks of the historical period. I accept both statements though I believe they are exaggerated. But I draw a very different inference. The fact that Greeks from the eighth or even the ninth century on had lost all trace and memory of Dorion or Aepy or twenty more such places has no relevance to what may have been remembered two or three generations or even two centuries after the destruction of Troy and of the lost places.19 'Destruction' is a dangerous word. Few places were ever so destroyed that no life continued or returned there, and anyway people lived on with memories, even if they moved elsewhere.

It was in the post-destruction, post-Mycenaean generations that the traditions about the heroic age and the Trojan War took shape.20 That seems to be characteristic of 'heroic ages' nearly everywhere21: they are a looking back after a break-down, and the past itself moves along with the generations of the present. Witness the Boeotians in the Achaean Catalogue. If in the early Dark Age the idea of a mainland coalition were invented, that is, if the main attackers in the Trojan War were shifted from Asia to Europe, that could have been done only by Achaeans (perhaps I should say 'ex-Achaeans') who looked to the mainland, and primarily to the Peloponnese, as their original homes (wherever they were now living). It would then have been simple enough, and indeed inevitable, for the specific place-names to be selected, in the first instance, from the place-names of Mycenaean civilisation.

We have a choice of explanations, neither of which is easy for us to visualise opera- tionally in our kind of world. One is that a very long muster-roll was passed on orally, generation after generation, either unattached to poems about the war itself or attached to versions very different from the one which finally survived, gradually distorted and in particular acquiring a wholly false Boeotian colouring. The other is that the very idea of a coalition and the appropriate catalogues were both built up without historical foundation during the generations after the Trojan War. Neither explanation gets round the grotesqueness of the final interpolation, and the choice between them is subjective and not a very happy one. My own choice is determined, negatively, by the failure of the 'Mycenaean geography' argument to carry conviction; positively, by everything else I have said thus far.

It is actually possible to narrow the field for subjective decision a bit further. If one agrees with Professor Page that it is 'certain that the Catalogue was originally composed in Boeotia' and if one accepts the tradition, repeated by Thucydides (i I2), that the Boeotians migrated from Thessaly sixty years after the Trojan War, then it must follow that 'a considerable period of time may have elapsed between the Trojan War and the making of the [Achaean] Catalogue' (p. 152). The interval, Page continues, was never- theless 'too brief to allow us to regard as fictitious the expedition with which the Catalogue is connected'. The point of disagreement is therefore the length of time required for it to become permissible to believe in total invention. (The Achaean Catalogue, we recall, is the only ground for deciding that there are 'differences', not 'disagreements', between the Iliad and the Hittite annals.) Obviously we cannot pinpoint the interval between the War and the Catalogue at exactly sixty years; it might have been a hundred. But even sixty is, in my view, long enough.22

I do not underestimate the strain it puts on the imagination to suggest that 'unofficial' Achaean participation in a marauding operation was twisted and magnified into our heroic Trojan War. But I do not believe the strain is any greater than that imposed by the

19 This point has been made in a review by A. 22 It is enough to cite the classic article of R. H. Parry and A. Samuel in C lvi (I96o) 85. Lowie, 'Oral Tradition and History', reprinted in

20 The basic discussion is now G. S. Kirk, The his Selected Papers in Anthropology (Berkeley, 1960) Songs of Homer (Cambridge, 962) chs. vi-vii. 202-10.

21 See Sir Maurice Bowra, The Meaning of a Heroic Age (Earl Grey Memorial Lecture, Newcastle, 1957).

8 M. I. FINLEY

This content downloaded from 128.172.48.178 on Mon, 26 Feb 2018 13:57:30 UTC All use subject to http://about.jstor.org/terms

THE TROJAN WAR

transformation of Christian Basques into Saracens and of a Hun invasion of Burgundy into a Burgundian invasion of the Hun kingdom. Yet we know those things happened; we can also suggest, after the fact, the psychology which underlay the transformations, and it would not be difficult to spin out an explanation of the rise of the Achaean epic. At this stage, however, when the suggestion is still only hypothetical, that would be a pointless gesture.

It can, and no doubt will, be argued that all comparisons with Roland and the other heroic traditions are false because they have all been contaminated by chronicles and other written documents (which is certainly true), whereas the Greek tradition was purely oral and a proper professional tradition of creative oral poetry is conservative and therefore tends to be more accurate and 'historical'. (A similar objection can also be raised against the undeniable evidence of the worthlessness of non-poetic traditions in illiterate societies in the Americas and Africa: there it is not writing which contaminates but the lack of a poetic tradition.) I do not know how one meets such an argument for the obvious reason that it is impossible to study a strictly oral poetic tradition over a long enough period of time. Control can come only from written documents, and the very existence of the latter automatically removes a culture from consideration.

In the end, the one hope for progress from hypothesis to verification (of any of the various explanations of the Trojan War) is that new Hittite or North Syrian texts may yet produce direct evidence. Until then, I believe the narrative we have of the Trojan War had best be removed in toto from the realm of history and returned to the realm of myth and poetry. The Song of Roland tells us much about feudalism in the eleventh century, nothing of any value about Charlemagne's court and the battle of Roncevaux. The Iliad and Odyssey, likewise, tell us much about the society in the centuries after the fall of Troy and scattered bits about the society earlier (and also later, in the time of the monumental composers),23 but nothing of any value about the war itself in the narrative sense, its causes, conduct, or even the peoples who took part in it.

No one in his right mind would go to the Song of Roland to study the battle of Roncevaux or to the Nibelungenlied to learn about fifth-century Burgundians and Huns. I do not see that the situation is any different with respect to the battle at Troy. True, we have nowhere else to to turn at present, but that is a pity, not an argument.

II. ARCHAEOLOGY AND THE TROJAN WAR, by J. L. Caskey

Material evidence from the site of Troy has indeed not proven that the place was captured by Mycenaean Greeks. Proofs, of the kind that Mr Fin ley demands and that we all should like to have, rarely come to light in any archaeological excavation. When they do appear they must be in the form of written documents, and in a language that can be read. At Troy these have not been found. If the argument is to be based on that fact, let us clear the ground altogether by asserting another negative: the physical remains of Troy VIIa do not prove beyond question that the place was captured at all. An accidental fire, in unlucky circumstances, on a day when a strong wind was blowing, might account for the general destruction that is known to have occurred. Furthermore, if this citadel was not sacked-and indeed if it was not sacked by Greeks under Agamemnon-we are left without a compelling reason even to go on calling it Troy.

All this is familiar to students of the problem and scarcely calls for a detailed restatement. The archaeological evidence, like the literary and historical, is incomplete and inconclusive.

23 I do not propose to re-enter that controversy or to repeat the reasons for my view that one may here (for the latest critical survey of the literature, legitimately reject the narrative as fiction but not the see P. Vidal-Naquet in Annales xviii (I963) 703-I9), social and cultural institutions.

9

This content downloaded from 128.172.48.178 on Mon, 26 Feb 2018 13:57:30 UTC All use subject to http://about.jstor.org/terms

One may take refuge in a declaration of doubt, but not many will choose to do so; Homer will not let us rest. And if an experienced archaeologist, who is not ignorant of Greek literature or of historical method, presents a short account of Troy for the benefit of non-specialists, surely he may be permitted to include a concise statement of his considered opinion upon this subject without adding a cloud of provisos and reservations that would defeat his audience.

The facts observed in the course of excavations at Troy remain unchanged. The Sixth Settlement ended with an earthquake of great violence. In Troy VIIla the people used the same fortifications, only slightly modified, and almost the same kinds of pottery and implements, but built their houses in a different way from those of Troy VI, different enough to make us certain that their manner of life had changed also. Settlement VIIa perished utterly in a fire. Of the succeeding settlement, VIIb I, only a few remnants were discovered and tested in the campaigns of 1932-38. The walls of its houses and the objects of daily use resembled those of VIIa, and there is no evidence that the inhabitants were of another stock. To what extent the fortifications were again restored is not certain, nor can we say positively how VIIb I came to its end. Troy VIIb 2 appears not to have been fortified. Again there are signs of continuity, but innovations appear also. The socles of the house walls often are built with stones set upright, a new structural feature, and about one-half of the pottery is of a distinctly different, handmade, foreign type.

Throughout these four periods Mycenaean wares were imported and imitated at Troy, and it is this pottery which gives the best evidence of chronology that we possess: the 'best', and requiring the closest scrutiny, but still far from perfect. Again let the facts be stated plainly, and the limitations be quite clearly understood, especially by those who are not used to handling pottery. Whole or nearly whole Mycenaean vases are rare in the strati- graphically certified layers at Troy, as at most other sites on the borders of the Mycenaean world. Sherds, which may be displaced and are therefore much less reliable for accurate determination of chronology, represent many more pots, hundreds altogether, but even so they are scarce enough. And exact dating of Mycenaean shapes and styles has not been established. Mr Finley may properly refer to this uncertainty, but having done so he may not imply that his arguments for a lower dating have been substantiated thereby.

The limitations were well known to Professor Blegen and his colleagues as they examined the pottery from Troy for the nth time about ten years ago (some twenty years after the Cincinnati excavations began). Arne Furumark's book was at hand, a monumental compilation of the whole body of Mycenaean material as known from publications before the second World War, and it was carefully consulted. But Finley's present suggestion that Furumark's methods and conclusions were uncritically accepted is an egregious mistake. So also is the studied but undocumented remark that 'most archaeologists' now reject his reasoning and his dates. Both were remorselessly examined from the start and both remain subject to criticism and correction. The fact that they also continue to command general respect is sufficient testimony to their basic value.

What then may be said today about the archaeological evidence for Trojan chronology? The ceramic style called Mycenaean IIIB had come into existence before Troy VI was destroyed. Mycenaean IIIC is not found in Troy VIIa but appears in VIIb i and there- after in VIIb 2. In spite of the uncertainties outlined above, these observations remain valid; to question them without rehandling the pottery itself is a waste of everybody's time. Assigning absolute dates to the pots and fragments, however, is another matter. That Troy VI ended near 1300 B.C. need not be doubted, but the stylistic change from Myce- naean IIIB to IIIC has not been firmly fixed, as Finley (among others) quite properly points out. In time it will be determined, but as yet we must allow a latitude of some twenty or thirty years toward the end of the thirteenth century, with a distinct probability that there was a moderate chronological overlapping of styles in some regions.

J. L. CASKEY 10

This content downloaded from 128.172.48.178 on Mon, 26 Feb 2018 13:57:30 UTC All use subject to http://about.jstor.org/terms

THE TROJAN WAR

The stratigraphy observed at Troy, specifically in Squares Eg and J5, was insufficient to reveal whether pottery of Mycenaean IIIC styles appeared early or late in Period VIIb i. If these styles reached Troy early in that period, and therefore soon after the destruction of VIIa, I should agree that it was difficult to accept Professor Blegen's most recent con- jecture of c. 1270-60 B.C. as the date of the destruction. If VIIb i lasted for, say, two generations, and Mycenaean IIIC pottery arrived toward the close of the period, there is no difficulty. This is another of the uncertainties that cannot be resolved at present.

A number of Finley's arguments are provocative and undoubtedly will stimulate lively debate. Too often, it seems to me, they end by saying, in effect: 'You cannot prove that you are right, and therefore I declare that you are wrong.' I myself feel invited to answer in the same vein by stating my own opinion (in no wise exclusively mine) that the develop- ment of Carolingian, Burgundian and South Slavic tales, fascinating as the subject may be, has very little indeed to do with that of the Iliad and Odyssey, and that 'faith' in the value of early Greek tradition is a quite respectable possession, if by the word one means the conviction at which a scholar has arrived after long and sober reflection upon all the available evidence.

But let us return to the war at Troy. If the sack of Settlement VIIa is ever shown to have occurred after the fall of Mycenae and Pylos, or at the same time, we shall indeed have to reject most of Homeric tradition. An army other than that of Agamemnon will have been the conqueror, and for its origin we may have to look despondently among those restless unnamed peoples of the north. Finley's suggestion that some dispossessed Achaeans might have joined them, and so have brought the event into Greek legends, is ingenious and would become plausible if the lower chronology prevailed.

On the other hand one can easily picture a different set of circumstances, equally practical and unromantic, which accords with the earlier date of the Trojan War and Achaean leadership, without denying a role to the shadowy tribes whom Egyptians called Peoples of the Sea.

Troy occupied an exceedingly important place, strategically, on the Hellespont. In the fourteenth century it may have been a strong bulwark against foreign intrusion into the Aegean, and as such, we may suppose, a friendly ally or at least a respected peer in co-existence with Mycenae. Then around I300 Troy VI was ruined by a natural catastrophe and the succeeding generation of Trojans, in VIIa, was obviously weaker. Perhaps it was thought to be less trustworthy as a guardian, and precisely at the time when there were reports of serious new dangers from the north. That, surely, was the moment for Achaeans, who were still strong, to improve their defences at home and to send an expedition to secure the Hellespont. The venture was ill-advised, as it turned out: Troy was more powerful, even now (in Period VIIa), than the Achaean general staff had expected; the siege was long and so exhausting that the Greek forces, though ultimately victorious, could not take advantage of their success and occupy the straits; their remnants straggled homeward. Thereafter, of course, Troy was indeed spent, and when the barbarians did finally arrive the inhabitants of the town that we call VIIb i were unable to offer serious resistance.

This is a fantasy, if you will-imagined, unsupported by positive proof-but it is not less valid than the theory that Mr Finley propounds. In my opinion an explanation of the events along these lines will come nearer to the truth.

II

This content downloaded from 128.172.48.178 on Mon, 26 Feb 2018 13:57:30 UTC All use subject to http://about.jstor.org/terms

12 G. S. KIRK

III. THE CHARACTER OF THE TRADITION, by G. S. Kirk

That epic traditions can distort historical events is commonly accepted, and rightly so. Mr Finley's thesis makes a welcome opportunity for asking what sort and extent of distortion are to be expected from traditions of different kinds. He cites distortions in the NJibelungen- lied, the Chanson de Roland and the Kossovo poems as support for the idea that the Homeric tradition could have developed out of an attack on Troy by northern invaders, perhaps with Achaean freelance helpers. The position shared by Finley with Heubeck, Nylander, Starr1 and others is that, if the incomplete archaeological phenomena do not at all points accord with the Homeric tradition about Troy, then it may well be the tradition that is at fault.

My main task in this discussion is to consider the argument from supposedly analogous and more recent traditions. I may as well say at once that I am not convinced that difficulties over the date of Troy VIIa and over the idea of a Panachaean expedition in the thirteenth century B.C. can best be met by the assumption of a major misunderstanding in the oral tradition. At the same time I see the dangers of drawing too precise a division between major and minor, unacceptable and acceptable, degrees of misunderstanding. Nilsson stated in I933 that the Teutonic epics 'raise many more problems than they can

solve', and unfortunately that is still the case.2 The Nibelungenlied, in particular, is an academic puzzle almost of the order of the Homeric one: how did it develop, what were its different regional components, how far is it oral? The stages by which the eventual twelfth-thirteenth century version was reached, the nafure of the distortions and additions due to successive centuries, are still largely unknown. Trying to construct Homeric probabilities on the basis of the Jfibelungenlied is like trying to solve metaphysical problems on the basis of the metaphysics. of Anaximander-it might be all right, if we knew what Anaximander's metaphysics were. For example, Finley's statement that 'there is no basis whatever' for the existence of Kriemhild or Siegfried means 'there is at present no known basis'. Certain other facts, it may be conceded, are more positively known: it is true that the chronological displacement of Theoderic is a notable example of historical distortion in a heroic tradition, and Finley rightly emphasises, too, the reversal of actual allies and enemies. Yet against these undeniable examples I must raise a point of central importance to which Finley himself has briefly alluded: that, although almost everything remains unknown about the processes by which events in central and northern Europe in the fifth and sixth centuries A.D. came down into the poetical form of our much later Nibelungenlied, it is known that written records, and the passage of written narrative material from one country to another, were elements in the situation.3 Curiously enough the art of writing can promote not accuracy but inaccuracy in the understanding of past events, at least until the copying of books, by well-organised scriptoria or later by printing, ensures the wide propagation of a stable version of what happened. Most of the early medieval chronicles were copied only in small numbers, if at all, and one-sided local versions proliferated. The position of Latin as a lingua franca in the literate world made the conflation of different ethnic accounts all the easier. When a copy of a regional chronicle happened to pass abroad it could be distorted, misunderstood, and conflated with other and incompatible material, with little if any control. Something similar happened with written copies of poems. The Hildebrandslied fragment,

in which Attila and Theoderic are already made contemporaries, dates from as early as around A.D. 800; and we know from Egginhard's Vita Caroli Magni (29) that Charlemagne

1 See Finley's n. 9 for references. 3 Cf. e.g., C. M. Bowra, Heroic Poetry (London, 2 M. P. Nilsson, Homer and Mycenae (London, I933) I952) 43-5; H. M. and N. K. Chadwick, The Growth

I85. of Literature (Cambridge, I932) i 482-4.

This content downloaded from 128.172.48.178 on Mon, 26 Feb 2018 13:57:30 UTC All use subject to http://about.jstor.org/terms

THE TROJAN WAR

himself commissioned collections of ancient and barbarous songs-what an opportunity for conflation, confusion, and error! Moreover the minstrels themselves often travelled farther afield (in ethnic terms, certainly) than their Greek predecessors; though Widsith need not be taken too seriously.4 These circumstances, together with the complexity of the migrations and invasions which in central Europe followed the disruption of the Roman Empire and accompanied the steady expansion of the Germanic peoples, helped to produce an extra- ordinary hotch-potch of quasi-historical traditions. The effects of this can often be judged by comparing English and Nordic versions of the same historical events.

Each region and ethnic group had its own version of important happenings and its own special heroes. Time-relations and personalities, even the whole trend of heroic enterprises, could undergo serious distortion in a mixed tradition. To the complications produced by the interplay of local and external versions must be added the effects of all these written records on the oral singer. It appears from analogous modern situations that he believes, ,often wrongly, that he cannot compete in accuracy with any written account; he tends in these circumstances to abandon the ideal of realism and historicity and feels free to embroider and conflate the stories he has inherited. In a true oral tradition, on the other hand, which suffers neither competition nor intrusion from literacy, the risk of serious conflation is initially slighter; and, in addition, the broad lines of heroic events handed down from earlier generations are controlled by a knowledgeable, conservative, and strongly localised audience.

For all these reasons it is of primary importance, in assessing the probability of serious distortion in a heroic poetical tradition, to decide whether the tradition in question is purely -oral or whether it is complicated by the effects of writing.

Everything suggests that the Homeric oral tradition was a pure one, free from intrusion by literate chronicles or poetic versions ossified in writing. Writing seems to have died out after the fall of the Achaean palaces near the end of the Bronze Age, to be revived only with the introduction of the simpler Phoenician system at a time when poetical traditions about Troy were already well established. As for the Bronze Age linear scripts, there is no evidence to suggest that they were used for the recording either of annals or of literature. The external evidence of the availability of scripts is confirmed by the internal evidence of the epic language itself, in which a highly developed formular structure also argues for a completely oral tradition.

The consequence is that the sort of thing that happened, in the course of transmission, to Attila and Theoderic (who becomes Dietrich of Bern) is unlikely to have happened, in anything like the same degree, to Agamemnon or even Achilles. Admittedly Ajax and Nestor may have suffered some chronological displacement; and the latter, like Idomeneus, may have worked his way into a central tradition from an outlying region. Yet the cultural and historical background remains broadly constant. When the conditions of a heroic age last for many generations (as we see from the Muslim tradition of Bosnia and Montenegro, fed by heroic circumstances which persisted from the sixteenth to the nineteenth century), it is easy for individual heroes to extend their range both in time and in space and for similar events and personalities to become compressed. What is less easy is for a hero to stray right outside the sort of heroic milieu to which he originally belonged. That is what seems to have happened with certain figures of the Teutonic saga; and here the impurity and geographical diffusion of the tradition are largely responsible. For the sake of completeness one other cause of displacement should be mentioned: if a hero possesses marked genre characteristics, perhaps originally derived from life, then he may become uncommonly popular with singers and audiences and intrude, like Marko Kraljevic, into .backgrounds and circumstances outside his own.

4 Cf. K. Malone, Anglistica xiii (Copenhagen, 1962) 77-9; Bowra, op. cit., 415 and 435 f.

I3

This content downloaded from 128.172.48.178 on Mon, 26 Feb 2018 13:57:30 UTC All use subject to http://about.jstor.org/terms

14 G. S. KIRK

Persons and events may suffer distortion for rather different reasons; and it is with events that Finley's argument is primarily concerned. Now some poems and some traditions are concerned with minor or highly localised events, others with major ones of widespread effect. So, for example, the Muslim tradition represented in the first volume of Parry-Lord, Serbocroatian Heroic Songs, takes as its main subject frontier skirmishes of the sixteenth century, while the Christian Serbian tradition focuses on the crucial battle at Kossovo in I389. Other things being equal, the latter type of event tends to become more securely fixed in its outlines than the former, in which the details of each small-scale encounter are liable to be conflated with those of many others. It is possible that the battle of Roncevaux, described in the Chanson de Roland, was of this minor type, though that is not undisputed. The encounters behind the Nibelungenlied also belong to this type rather than to the Kossovo category; and here, too, the tradition is built on a complex of battles (whatever their scale) rather than on a single event-another factor which tends towards conflation and distortion. The distinction between major and minor encounters, and singular and plural ones, must be firmly carried in mind when we try to assess distortions in the Homeric tradition about Troy (though it is true, as Mr Finley would not allow me to forget, that our only evidence for the magnitude of the attack on Troy is precisely the subject of dispute, namely the Homeric tradition itself). The Chanson de Roland differs from the Nibelungenlied in the degree of its identifiable

distortion and the number of different regional versions and components that may be involved; but resembles it in the profound obscurity of its development. Again, however, written annals and records, many of them in Latin and therefore able to cross regional and racial frontiers, were certainly available. They undoubtedly affected some stages of the developing tradition, and the use of written Gesta Francorum is even mentioned in the poem (3262; cf. I443, I683, 3742); so are 'charters of St Giles', who claimed to be at Roncevaux (2095-8).5 Yet surviving records mentioning the battle of 778 are few, and indecisive over scale and detail. They do not justify a categorical estimate of the degree of distortion in the twelfth-century poem, at least so far as the nature of the enemy and the importance of the engagement are concerned. With personalities (like the traitor Ganelon) distortion or chronological displacement is more certain, but also more predictable; and it is generally agreed that the cultural background is much affected by the more recent times of the first crusade. Elaborate theories have arisen about the formation of the poem, and even the most unlikely of them have had their vogue. My impression is that there is more to be gained from using Homer to elucidate Roland than vice versa. The South Slavic tradition is regionally more unified than the Teutonic or even the

Frankish, though complicated by the juxtaposition of different religious elements (Christian and Muslim) with opposing political sympathies. In most respects it is more closely comparable with the Homeric tradition than are the other two. The poetical distortion of the events of Kossovo is not very serious, as Finley agrees; and this really removes the South Slavic tradition from the case he is trying to make, or even places it at the disposal of his opponents. By the evidence of contemporary chronicles, both Christian and Turkish, the poems of the Kossovo cycle get the enemy, the result, and most of the main personalities right. There are slight differences between prose and poetical accounts of the manner in which Milos Obilic killed Sultan Murad, but that is a detail and does not involve the degree of distortion with which Finley's case is concerned. It is the kind of imprecision which is inevitable in any oral tradition, even a tightly-controlled poetical one. More serious distortion is seen in the interpolation of Marko Kraljevid as a conspicuous hero on the Serbian side, and in the casting of Vuk Brankovic in the role of traitor. In the latter case it may be conceded that personality and event are interwoven; and the explanation of the

5 Cf. Bowra, op. cit., 37I.

This content downloaded from 128.172.48.178 on Mon, 26 Feb 2018 13:57:30 UTC All use subject to http://about.jstor.org/terms

THE TROJAN WAR

Serbian defeat as due to Vuk's defection involves a substantial departure from historical reality. The historical Vuk Brankovic seems to have turned over to the Turkish side some years after the great defeat, as a result of a personal quarrel. Somehow this allowed him to fill a role that must have been greatly desiderated by singers and their audiences, that of scapegoat or acceptable cause of defeat. For defeat to be caused by inferior bravery or skill at fighting was intolerable in a genuine heroic tradition; betrayal, inferior numbers or divine displeasure are the acceptable thematic causes. As for Marko, I have already mentioned that this relatively obscure historical person became a legendary figure with genre or folk-tale characteristics, and so worked his way into songs and events with which he originally had no connexion. He apparently did not fight at Kossovo, but he lived at that period, and it is not altogether surprising to find him in the Kossovo songs-though it is surprising to find him so prominent there. It is unlikely that literate influences played much part in the transformation on this occasion; but written chronicles, the cultural domination of a literate priesthood, and even written copies of poems, are potentially important elements in the heroic tradition of Christian Serbia. That is why the Muslim songs provide on the whole a safer (though still distant) parallel to the Homeric tradition than those of the older but more impure and more sophisticated Kossovo cycle.

In short, the comparative material adduced by Finley does not seem to meet his requirements, mainly because that part of it which contains notable distortion is significantly different, in its manner of development and transmission, from the Homeric poetry. Against my distinction of pure and impure oral traditions he replies that we cannot know what happens in a pure oral tradition, simply because, being oral, it is removed from literate study. But the fact is that a genuinely illiterate tradition may have as its subject events recorded elsewhere by other, literate parties; and this allows some control. Thus Finley himself hopes for new Hittite documents to allow us to assess the accuracy of the illiterate Homeric tradition. Moreover the detailed operation of illiterate singers can be studied, in modern times at least, by literates like Parry and Lord. Some distortion of the picture there must be, but not necessarily enough to disguise the main characteristics of a pure tradition. What may be agreed, on the other hand, is that many of the arguments I have brought forward have been a priori in character. They are not to be ignored on that account alone, in a context where so much is hypothetical; and they are not entirely unsupported by factual evidence, like the spread of written chronicles and the intermixture of national traditions on the one hand, and the observed effects of printed songs and histories on Slavic singers on the other-effects which include the corruption not only of the quality but also of the narrative content of their songs.

Let us now look more directly at the Homeric tradition about Troy. The scale of the war is greatly exaggerated, and many Iron Age objects and customs have intruded them- selves in the course of transmission; that is admitted. Yet we must constantly bear in mind precisely what Finley's kind of theory requires us to accept: not merely that the Homeric tradition is distorted, but that it is so severely distorted that its whole picture of a united expedition from the Achaean mainland is false. A small minority of roaming Achaean refugees may have been involved (according to this theory), but the main enemies of the Trojans were northern invaders of whom the Homeric tradition gives not the slightest hint. Now that tradition is not just a vague legendary one; it preserves a great deal of accurate information, not only about Bronze Age social institutions but also about Bronze Age armour, buildings, and people-and some of this information is crystallised in recognisable relics of Bronze Age terminology. I am quite prepared to agree with Finley that the total picture of life given in the Iliad and Odyssey owes almost as much to the circumstances of the early Iron Age as to those of the real Achaean world of the late Bronze Age. Misunder- standings of Bronze Age methods of fighting, especially perhaps over the use of chariots, probably imply that there was no detailed poetical account available to the singers who

I5

This content downloaded from 128.172.48.178 on Mon, 26 Feb 2018 13:57:30 UTC All use subject to http://about.jstor.org/terms

16 G. S. KIRK

developed the kind of poetry that has survived in our Iliad. But can we believe that the interruption of the tradition, whether poetical or non-poetical, caused by the upheavals at the end of the Bronze Age can have been so severe as to destroy not merely the details but the very outlines and whole substance of events belonging to the last heroic period of the Achaean civilisation? The magnification of a heroic past is common enough; how common is the virtual creation of a great heroic enterprise, in this case based merely on the exploits of an assumed expatriate minority, and that at a time when genuine older and indigenous traditions (Thebes, Argonauts, Pylos) evidently still survived? I have emphasised else- where that the early Dark Age probably had more to do with the creation of the poetical tradition about Troy than is often assumed.6 Finley accepts this, and uses it as a means of arguing that the interval of a mere couple of generations at the end of the Bronze and beginning of the Iron Age was enough to allow the kind of distortion over Troy which he envisages. I believe there are grave difficulties in this assumption, and that the parallel referred to in his note 22 is as incomplete, on close inspection, as that provided by the Nibelungenlied and the rest. It seems highly unlikely that Achaean freebooters could turn in men's imagination into

a great coalition from the mainland, except under one or both of two conditions: first that there was an almost total break in culture and tradition between the end of the Trojan War and the early Iron Age, and second that the distorted version grew up overseas, out of the stream of mainland tradition, where curious unhistorical versions could proliferate unchecked. The first of these conditions certainly did not apply; so much is proved, if by nothing else, by the amount of certifiable Bronze Age information (quite apart from the Trojan War) which descended into the Iliad and Odyssey. The second looks more promising; but one instantly runs into a problem of motive. If the distorted version grew up overseas, it was probably in Asia Minor or the Levant, among the descendants of Achaean settlers there-descendants, most probably, of the hypothetical Achaean free- booters themselves. Yet what motive could they have had for elaborately making out that the attack on Troy was the work, not of their own direct ancestors, or even of those of other expatriate settlers and refugees, but of the mainlanders themselves? Did they wish to connect their ancestors with the main Achaean palaces, now long since collapsed? This looks possible, but too complicated; one would expect such people to exaggerate, rather than disguise, the expatriate achievements and new regional affiliations of their fathers and grandfathers. A different argument from motive is used by Finley against the Homeric picture of a

coalition expedition. Once again it is helpful to be made to see an old difficulty in a new light, but once again I remain unseduced by his objections. First of all, there are many other wars in history whose exact motive the modern historian cannot adequately explain. The Trojan War is not unique in this respect, and does not for this reason have to be abolished. The interaction of long-term motives and immediate casus belli adds to the complication of reconstruction. The Trojan War was probably not 'commercial' in the sense of being fought to win or defend markets or continuing sources of raw materials. Commerce may have been something like that in the settled days of the Achaean palaces, but it can have been so no longer by the middle of the thirteenth century B.C. It is not at all sure, however, that a large-scale booty raid is out of the question. Many of the buildings of Troy VIIa were not of themselves impressive, but the huge wall of Troy VI still remained to give the appearance of wealth. It may well have been supposed on the mainland that this wall was not adequately repaired, that Troy was now vulnerable to attack. If the Homeric tradition is right, then this supposition was not completely correct; but that is immaterial. Judging by its imports of Achaean pottery, Troy VI in its later

6 The Songs of Homer (Cambridge, 1962) ch. 6.

This content downloaded from 128.172.48.178 on Mon, 26 Feb 2018 13:57:30 UTC All use subject to http://about.jstor.org/terms

stages had been a useful commercial partner; Troy VIIa, by the same criterion, was ceasing to be so. But was it really worth attacking? Was it completely drained of all its old wealth, its metals and its horses, so conspicuous in the Iliad? Did the earthquake which ruined Troy VI destroy all of this-or, more to the point, was it known on the Achaean mainland to have done so? Did the earthquake empty the royal graves of precious objects of gold such as had been buried with the kings of Mycenae and Midea and in the rich and numerous tholos tombs of Messenia? It matters not whether there were in fact such graves at Troy (probably there were not); it is the remote speculations of mainland Achaeans that concern us here. To cut a long story short, the rumour may have been widespread in the thirteenth century B.C. that 'Troy is still as rich as anything'; and we cannot even be sure that such a rumour would have been totally incorrect.

Would the prospect of such booty have been enough, in the conditions of the times, to motivate a large expedition from the already threatened palaces of the mainland? In itself, perhaps, not so; there may well have been other reasons too, like the need for the Achaean nobility to restore its prestige or for a failing king to assert his authority. I do not underestimate the effort required: moving a contingent across the Aegean was no easy task, especially if it included horses. Perhaps only a small force went at first, to be gradually increased as a long siege proved necessary. That is just the sort of detail the tradition would simplify. As for the coalition aspect of the expedition, inter-palace ventures were nothing new in the Achaean world, judging from the Argonautic tradition and the south- Achaean alliance against Thebes. Even so, the organisation of a joint force remains somewhat surprising in the probable circumstances. Yet personally I should find it more surprising if the general tendencies of oral transmission in an illiterate society were to be seriously upset, than if events, decisions and political changes in a largely unknown Aegean world were to have been rather different from what we may at present expect. In short, the nature of the tradition about Troy is still a most serious impediment to the kind of reconstruction supported by Finley, unless we water down that reconstruction so much (for example, by arguing that the supposed Achaean freebooters had originally come from different mainland states, and so were a coalition in a way) that it differs from the tradition only in emphasis.

IV. HOMER AND THE TROJAN WAR, by D. L. Page

Many of us have been tempted in our time to look for a connection between the fall of the Hittite Empire, the collapse of Mycenaean Greece, the destruction of certain places in Syria and Palestine, and the assault of 'northerners' and 'sea-peoples' upon Egypt. Mr Finley has now put our hopes or fears into a quite definite form, suggesting that 'northern marauders' (represented in our records solely by the Delta-raiders) were 'a signifi- cant, perhaps the main, generating impulse' for all these (and some other) catastrophes; and he maintains in particular that they are likelier candidates than the Achaeans for the prize of Troy. If I disagree, it is mainly because I find the Finley-hypothesis incapable of verification at any significant point, whereas the Achaean-hypothesis is the apex of an assembly of observations which suffice to bear the construction put upon them; though it is salutary to be reminded with so much clarity and force that it is theoretically possible that they might bear the weight of quite different constructions if only we had more bricks to build with.

The evidence of Homer, that Greeks from the mainland sacked Troy (this I call the 'basic narrative'), cannot be proved to the exclusion of other possibilities. That is frankly admitted; and let it be admitted with equal candour, (a) that it has been confirmed by other evidence at certain material points, (b) that no rival account of the destruction of

THE TROJAN WAR I7

This content downloaded from 128.172.48.178 on Mon, 26 Feb 2018 13:57:30 UTC All use subject to http://about.jstor.org/terms

I8 DD. L. PAGE

Troy VIIa has any confirmation in any source. I briefly re-state some of the points on which the Homeric narrative has been confirmed:-

I. We have learnt in the present generation that the Homeric epic has a continuous history reaching back to the Mycenaean era. Its formula-system includes elements which must have entered it at a quite early phase of that era. A few of these elements can be identified with certainty, and it is prudent to allow that there are others which, for lack of evidence, we happen to be unable to identify. We all agree that the basic narrative has been greatly amplified with fictions; but the fact that the basic narrative may have begun very near the time of the events themselves, and must have begun within a few generations after them, is evidence (I do not say proof) that the basic narrative is trustworthy; for it is unlikely that poets at Court or in the market-place would tell of the participation of the great families on the mainland, if their audience knew (from fathers and grandfathers) that the great families had not participated. We make no claim to proof of anything here, only (as elsewhere) to evidence, unthought of a hundred years ago, in favour of acceptance of the basic narrative.

2. Homer says that Troy was sacked. The world smiled, when it did not sneer. Now Hissarlik has been fully excavated, and nobody doubts that Troy VIIa was destroyed within the Mycenaean period.

3. Homer says that the centre of Achaean power was at Mycenae, rich in gold. This too was surely a fable; few believed it. After nearly a hundred years of excavation at Mycenae and other palaces on the mainland, there is (I suppose) not a sceptic left.

4. Homer says that Troy was besieged by Achaeans. Archaeology has brought the Mycenaeans into very close contact with Troy; indeed it has shown that the Mycenaeans were active throughout the Eastern Mediterranean for a long time, settling in Colophon, Miletus, Rhodes, and Cyprus, trading all round the coast from Troy to Tell el-Amarna.

5. Now come the Hittite documents, and (if we accept the equation of Ahhijawa with Achaia) we can add many touches of colour to the picture of Achaeans and others on the west coast of Anatolia: Achaean kings and buccaneers penetrating quite a long way inland; north of them, a league of Assuwa, including (probably) Troy itself; the Hittites, hitherto keepers of the peace, withdrawing in the late thirteenth century. If Mr Finley has difficulty in thinking of a motive and a historical context for the Trojan War, conflict between commercial (and perhaps imperial) rivals within the vacuum created by the Hittite withdrawal from western Asia Minor will provide one likely enough; and the stage for it is set not by mere conjecture but by the Hittite documents.1

6. Homer says that Greek forces came from the mainland against Troy. Even for this (the main point of disagreement between us) there is a measure of confirmation in the fact that Homer's Order of Battle describes the mainland as it existed before the

collapse of the Mycenaean world. The Catalogue has been amplified, and the historical picture distorted; but the basic narrative is much likelier to be true than false, if (as we believe) it has its origin in a time when father would have known from grandfather whether it was true or not. Of course I recognise the room for disagreement

Mr Finley writes: 'The Greek tradition does I prefer the hypothesis that Achaeans joined a not . . . provide a proper historical context and a marauding force of northerners.' This is not so motivation for the . . . destruction of Troy'; bad simple a preference as it sounds: my conjecture is feeling between two powerful kingdoms is not, based on documents which show Ahhijawans and historically, an unprecedented motive for war. As Assuwans active in contiguous territories in the same for historical context, archaeology and the Hittite period, both warfaring peoples; Mr Finley's hypo- documents have filled the gap more than adequately. thesis postulates a marauding force of northerners Mr Finley writes also: 'Page has conjectured a for whom, at this time and in this area, there is no political struggle between Ahhijawa and Assuwa. evidence whatsoever.

This content downloaded from 128.172.48.178 on Mon, 26 Feb 2018 13:57:30 UTC All use subject to http://about.jstor.org/terms

THE TROJAN WAR

here, but I question the principle of Mr Finley's objections: he allows his audience to remember the names and descriptions of numerous places which had been unoccupied since the Mycenaean period, but does not allow them to remember whether their forefathers had sailed to Troy or not. In my view you might add to the story (Boeotians and the like), but there must have been a story to add to, and it must not be in contra- diction to folk-memory on a point of such magnitude.

Mr Finley says nothing about the Trojan Catalogue; it was an important step in my argument, and I still think it uncommonly solid ground.

The list of confirmations could be extended into numerous details, mainly matters of language and formula; I have done what I can elsewhere. The position is this: our Epic witness makes a number of historical assertions (or implications); we can only prove or disprove a few; on all the most important of these, the witness is reliable. We can confirm other assertions up to a point far short of proof; on almost all of these, the witness is reliable so far as we can judge. Excavation at Troy; excavation at Mycenae and elsewhere; Nilsson on mythology; Parry on oral technique; Hittite documents; archaeological researches throughout the eastern Mediterranean; Order of Battle in the Iliad; detailed study of language and formulas,-all have added their mite of proof or confirmation; none has given reason to doubt the basic narrative. And we have still said nothing about what seems to us the huge intrinsic improbability inherent in Mr Finley's opinion-that an Epic originating within a few generations (at latest) of the events should include or develop into a fiction of the alleged magnitude. In this respect the Chanson de Roland and the Kossovo-Epic are, on Mr Finley's showing, on our side, not his.

I have already admitted that the evidence is not wholly satisfactory. There is a big difference between acquitting a man of perjury and proving that he has told the truth. There are too many unknown factors. We should be on much firmer ground if we were not so ignorant of the sequences, time-intervals, and other relations (if any; we make no assumptions) between the following events: (i) the collapse of the Hittite Empire; (ii) the sack of Troy VIIa; (iii) the destruction of Pylos; (iv) the fall of Mycenae and Tiryns; (v) the building of Broneer's great wall across the Isthmus (and was this ever finished? Marinatos thinks not. If not, was that because the danger had receded, or because invasion came too soon? These are not the only possibilities). Again, who destroyed Pylos? Against whom was the great wall across the Isthmus built? Why was Gla abandoned (if it was) in the generation of its building (whenever that may have been) ? We do not know, and it would be most imprudent to postulate a single cause.2 We certainly have no reason at present to believe that any of these matters is an obstacle to belief in an overseas expedition to Troy. We do not know the chronological order of events, let alone the time-intervals. It remains perfectly possible that the overseas expedition antedated awareness of threats to the mainland by five years (or two years or ten years; a very short interval would suffice).

The Homeric account has been confirmed since i870 to an extent unimaginable before that time. It is very likely the true account; at least it is the only one which can claim the support of various and abundant evidence in both literary and archaeological records. But its claim to our confidence would be much stronger if we could relate it to the catastrophes on the mainland. This being so, is it not reasonable to allow marauding northerners to knock, however timidly, at the door of our conservatism ?

Where so little is known, little can be absolutely denied. For Troy and the Greek mainland, there is no evidence whatever of the activity of the Delta-raiders or of any

2 A hypothesis is not necessarily the more probable frontiers of the Roman Empire at sundry periods; for being 'economical'; imagine how misleading it or to Central Europe during the wars of Frederick would be to apply the principle of 'the most eco- the Great; or to the European occupation of nomical hypothesis' to the northern (or eastern) Northern America.

I9

This content downloaded from 128.172.48.178 on Mon, 26 Feb 2018 13:57:30 UTC All use subject to http://about.jstor.org/terms

20 D. L. PAGE

peoples directly or indirectly connected with them. The connection of the Delta-raiders with the Hittites is at present so remote and tenuous that the scales of judgment remain quite unaffected. Mr Finley seems to me to be substituting the wholly unverifiable for the partly confirmed. The future may prove him right; the present is not in his favour. If we are to follow him in principle, why go so far afield ? The most obvious candidates for the prize of Troy are the Hittites; I could make a good case for them. The next most obvious are the Thracians and their neighbours. Thirdly, for all I know, some proto- Dorians. There is no evidence that any of these peoples had anything to do with Troy VI or VIIla: they have therefore as much or as little right as any 'northern marauders' to gather round our gates. Our present tenants, Homer's Achaeans, may not be quite certain of the legality of their lease; but at least they have a lease, and those clauses which can be inspected have proved valid enough. The inquiry will continue; and those of us who look with disfavour on gate-crashing brethren of the Delta-raiders will do well to remember the words of a colleague of Mr Finley and myself: the historian, like the art-critic, must not forget 'the subtle difference between the ability to make sense and the possibility of under- standing'. We claim to have achieved the former; the latter is quite out of our reach.

This content downloaded from 128.172.48.178 on Mon, 26 Feb 2018 13:57:30 UTC All use subject to http://about.jstor.org/terms