![[LOGO]](Logo.png){kind=logo}
European literature certainly, and perhaps European history also, begin with Homer’s long epics, the Iliad and the Odyssey. The Iliad is about an episode during a war between the “Achaeans” – men from an area of what we now know as Greece – and Troy, a city-state near the mouth of the Dardanelles in north-west Asia Minor (nowadays in Turkey, but the Trojan War dates back long before the Turks arrived from Central Asia and invaded what had been Greek lands). One of the Achaean leaders was Odysseus – the Romans called him Ulysses. After a siege lasting many years, the Achaeans finally conquered Troy; the Odyssey is about Odysseus’s journey home after the war, during which he has many fantastic adventures but finally gets back to his home island of “Ithaca”, where his wife Penelope is waiting and fending off the attentions of numerous men keen to move in on her.
There is an island called Ithaca (Ιθάκη, in modern Greek Itháki) today, one of the Ionian Islands off the west coast of mainland Greece, and the assumption has been that (if there really was a man called Odysseus) that must have been where he lived. But a fascinating new theory has been put forward which says that this is wrong. Odysseus really was a historical figure, but the place he called “Ithaca” was not modern Itháki: it was what is now a peninsula (called Palikí, Παλική) attached to the western side of the large island whose modern name is Cephallonia (Kefalonia, Κεφαλληνία). What is more, Palikí contains fragmentary remains which could plausibly be the ruins of Odysseus’s palace. We might actually be able to touch today traces of walls and pavements which the Homeric hero walked among.
This theory is intriguing and very attractive. But it suffers from one large problem which its authors have never seriously discussed, so far as I have seen. The aim of this web page is to put a case against identifying Odysseus’s Ithaca with Palikí. I am not saying that I am certain the new theory is wrong; but I should like to stimulate discussion of the problem with it, because until this is addressed properly I cannot think that it is reasonable to believe that Ithaca=Palikí.
Nobody knows whether the Trojan War described by Homer actually happened – which is why I said above that Homer is certainly the beginning of European literature but only possibly the beginning of European history. But the world in which Homer describes that war as taking place appears to be the Greek Late Bronze Age. Assigning even a rough numerical date is itself nowadays controversial. Traditionally it was held that the Trojan War would have occurred some time round 1200 BC; but there is now an argument (documented in Peter James, Centuries of Darkness, Cape, 1991), which seems rather solid and persuasive to me, according to which all traditional Ancient World dates before about 700 or 800 BC are about 250 years too high – if that is true, then the Trojan War might have been about 950 BC. Homer himself (if he was an individual man) lived later than that, perhaps in the eighth century, but the gap in time between Homer and the events he describes in detail is explained by saying that Homer’s role was essentially to edit into written form epics which had been handed down orally by bards for generations before him – as often does happen in preliterate societies.
Archaeologists have been trying to confirm the identity of Odysseus’s Ithaca with modern Itháki for two hundred years, but there have been long-recognized difficulties. As Martin Young puts it in his Travellers’ Guide to Corfu and the Other Ionian Islands (Cape, 1971), virtually all archaeology on the Ionian Islands up to at least the Second World War was aimed at
the solution of some aspect of what has come to be known as the “Ithaca Question” [, which] arises from the difficulty of reconciling Homer’s account of Ithaca’s position with the geographical situation of the island now called by that name.
Modern Itháki just does not match the descriptive information about Ithaca given by Homer. Thus, Homer says that Odysseus’s Ithaca is the furthest out to sea and westernmost of a group of islands, whereas Itháki is east of and closer to the mainland than the much larger adjacent island of Cephallonia. And there are other large contradictions.
On the other hand, if Ithaca was not Itháki, where was it?
In a large and detailed book, Odysseus Unbound (Cambridge University Press, 2005), an enthusiastic amateur, Robert Bittlestone, with the help of the classicist James Diggle and the geologist John Underhill, argues that what is now the Palikí peninsula could once have been separated from the rest of Cephallonia by a sea channel; and that, if that was so, in other respects Palikí would fit Homer’s description of Ithaca in such minute detail that the identification is virtually certain. The Ionian Islands are an area of intense seismic activity, where the African plate is being overridden by the Eurasian plate – only fifty years ago there was an earthquake which destroyed almost all buildings on Cephallonia and produced large landslips. The reason why Palikí is not an island today, according to Bittlestone, is that three thousand years of earthquakes and landslips have blocked the channel that used to separate it from the rest of Cephallonia.
One obvious objection – if Odysseus’s Ithaca was Palikí, why do present-day Greeks use the name for a different place? – is probably not as serious as it sounds. Down the centuries the Ionian Islands have been ruled by speakers of many different languages (for much of the nineteenth century they were part of the British Empire), and have borne many alternative names both in the rulers’ languages and in Greek. No references to “Ithaca” occur anywhere in literature which has survived from the classical (i.e. post-Homeric) period. Furthermore, Itháki has not been inhabited continuously; five hundred years ago, for instance, it is known that no-one lived there. So it is not out of the question that a name could have got swapped from one place to another, particularly when it was so culturally important for Greeks to be able to point to an island called Ithaca but (if Bittlestone is right) the original Ithaca had disappeared as an island.
The problem I see is quite different. How realistic is it that there could ever have been the kind of channel which Bittlestone postulates as having separated Palikí from the rest of Cephallonia three thousand years ago? It would have been about 4½ miles long, yet only a hundred yards or two wide (working from the indications of scale shown on the illustrations in Odysseus Unbound, the channel depicted seems closer to 100 than 200 yards wide for most of its length, and nowhere wider than about 350 yards – see also Figure 14 in James Underhill’s article “Quest for Ithaca” in Geoscientist, September 2006). If we encountered a phenomenon like that in the modern world, we would assume that it was an artificial canal. Could it really arise naturally?
It is crucial for Bittlestone that Palikí should have been a true island, surrounded by water on all sides. One part of the argument in Odysseus Unbound involves a claim that the Greek word for “island”, νῆσος, could be used in an extended sense to mean “peninsula”, as we call a peninsula in London the “Isle of Dogs”. But Bittlestone is not suggesting that Odysseus’s Ithaca was that sort of “island”. Much of his exercise in matching up the geography of Palikí with Homer’s story depends on postulating that one could get through the hypothetical channel in a boat, and that ferries were used to transport men and beasts between Palikí and eastern Cephallonia. Currently, the theory of Odysseus Unbound is being actively pursued by sinking boreholes at the presumed position of the channel, in order to discover whether the surface of the bedrock is lower than sea-level, as the theory requires.
John Underhill, one of the three collaborators on Odysseus Unbound, is a geology professor at Edinburgh University. He puts considerable effort into arguing that, if the postulated channel did once exist, then it might plausibly have been filled in through repeated seismic episodes. This is no small claim in itself, because the area of the channel is not flat today: to get from one end to the other one has to climb about 600 feet above sea level. Nevertheless, to a layman like me it seems relatively easy to grant Underhill this part of the argument. We know that earthquakes can sometimes lead to dramatic reshaping of landforms. Immediately to the east of the hypothesized channel, Mt Imerovigli rises over 3000 feet, so there may have been no shortage of land to slip down and bury a previous channel.
Remarkably, though, the many pages which Underhill devotes to arguing that if the channel existed, then geology could have converted it into the present-day terrain, are not balanced by so much as a sentence to make it seem plausible that such a channel could indeed have existed in the first place. What could have created it? The only wording I can find in the whole book which comes near to addressing that issue is on p. 545, where Underhill suggests that
the most likely reason for its original formation was that it was created by subaerial erosion during the last glaciation (c. 30–20,000 years ago) when sea levels are thought to have been more than 100 m[etres] lower than at the present day
– after which rising sea-level over the next 10,000 years drowned it and created the continuous sea link between the two sides of the isthmus.
When I first read this I misunderstood it as suggesting that the hypothetical valley was scoured out by an Ice Age glacier, but that is not what Underhill means. A paper “Ithaca theory gains support” by Ted Nield in the February 2007 issue of Geoscientist explains that ice never reached as far south as Greece; “the last glaciation” is relevant only because, by locking up so much of the world’s water in glaciers further north, the Ice Age caused sea levels everywhere to be low. Nield quotes Underhill as saying
Marine waters reached the upper part of the gulf as recently as 5000–6000 years ago, after the enhanced upstream erosion that took place due to the lowering of local base levels.
In other words, the hypothetical valley that was destined to be drowned when the sea-level rose was created, as valleys often are, by rainwater running off hills and forming rivers. The caption to the schematic illustration in Nield’s paper refers to “fluvial incision” in what is now the isthmus linking Palikí to the rest of Cephallonia.
But water flows downhill, not horizontally for miles. If runoff from Mt Imerovogli to the east and the lesser hills to the west had begun to erode lower ground between them, one would surely expect the eventual outcome of that process to be a valley higher in the area immediately between the two ranges and lower further north and further south, as the streams descended towards the bottoms of the bays which now adjoin the two ends of the hypothesized channel. In that case, though, either the rise in sea-level would not be enough to form a continuous sea connexion from bay to bay, or, if it was, then the ends of the channel where they meet the bays would be far deeper (and, since the valley sides can hardly have been perfectly vertical, far broader) than Bittlestone suggests. Bittlestone and his collaborators do not believe that their channel was deep anywhere. For instance, Figure A2.23 in Underhill’s appendix to the book, showing a hypothetical cross-sectional profile of the strata at the southern end of the channel, show the bedrock sloping down to just below current sea-level (under hundreds of feet of hypothetical earthquake-generated debris).
It is clear from the shape of the postulated valley that it would have had to be produced by water rather than ice flow. Valleys shaped by glacial action are U-shaped, with broad floors, unlike the V-shaped valleys created by water flow. The Figure just quoted, and other material in Odysseus Unbound, make it explicit that the valley containing the authors’ hypothetical channel was not just V-shaped but extremely steep-sided. Traversing it by boat would have been a dramatic experience, perhaps a bit like passing through the Corinth Canal.
I would sum up my puzzle by asking where, in the world as we know it today, is there a sea channel (not created artificially by Man) which comes even close to being fifty or sixty times as long as its typical breadth?
The Dardanelles or the Bosphorus do not approach that proportion, but in any case I understand that they are thought to have been created by river-like water flow, when the Mediterranean broke through and over a long period filled a basin which became the Black Sea. Nothing like that could have happened at Cephallonia; the hypothesized channel linked two arms of the same sea, so while the sea-level may have fluctuated over the geological aeons it will always have been the same at both ends of the channel. And although the diagrams of the course of the postulated channel are schematic, and do not pretend to be precise with respect to each little kink and bend, the general property of extreme narrowness relative to length seems to be required by the geological discussion which projects the buried valley sides from the stratal formations observable at the surface (referring again to Figure A2.23).
If, as I believe, there is nothing at all like Bittlestone’s channel anywhere today, then how realistic can it be to postulate such a thing in order to interpret a piece of literature from the dawn of civilization, with all the uncertainties and obscurities which it unquestionably contains?
Being no expert myself, I am very open to being shown by those who are that my scepticism is ill-founded. But the objection I have raised seems so obvious that I would have expected it to be addressed, if there is a good answer to it. In a large-format book of almost 600 pages it is never touched on. Nor is it mentioned in the book reviews I have seen so far. Peter Green of the University of Texas seems to speak for many reviewers in a long and thoughtful article, which suggests at several points that Bittlestone risks letting his enthusiasm run away with him, but nevertheless concludes that the Ithaca=Palikí idea “is almost certainly correct”. He does not allude even in passing to the strangeness of the hypothetical channel.
Until someone takes this point seriously and explains why it is less of a problem than it appears, I must draw the opposite conclusion. Bittlestone has constructed a lovely theory, and reading the details of how he and his collaborators have followed the idea up is enormous fun. I wish it could be true. But, alas, it is almost certainly wrong.
— After placing the above material on the Web, at Robert Bittlestone’s suggestion I posted a message briefly outlining my objection and referring to this web page on the Odysseus Unbound forum, in order to invite discussion by the community of people interested in the theory. This drew an immediate response from someone identified as “Demodocos, Site Admin” (so apparently in a position to write with authority), who said:
This is a very interesting geological issue and one which was by no means clearly understood by the project team when the book was published in October 2005.
Demodocos points out that “industry-scale resources” are now being used to look below the present-day surface of the site where the hypothesized channel ran, “and this has broadened the team’s understanding of this issue … further details of this research will be announced in the September 2008 issue of Geoscientist magazine. John Underhill will also be speaking on Oct 2 2008 at the Geological Society [London]”.
To my mind this response makes the theory seem even less plausible than before. It comes to much the same as saying “The Paliki=Ithaca case which we made at great length in 2005, which has attracted worldwide attention and convinced numerous reviewers for respectable newspapers and journals, contained a gaping hole and should not have convinced anyone; but we have extra evidence now, and if you can hang on till September perhaps you will find that the argument works after all.”
Perhaps we shall. But if the team convinced themselves sufficiently to publish their 2005 book on the basis of an argument which, to a friendly but objective outsider, seems so obviously flawed, then I wonder what the chances are that evidence they bring forward three years later will truly change the picture (though I look forward to reading the September article in due course). [Added in Oct 2008: However, now that I have read Underhill’s paper “Testing classical enigmas” in the September 2008 Geoscientist, it turns out to be an interim report which cannot really be said to take the argument further.]
Demodocos’s phrase “industry-scale resources” suggests that, because the theory is now being investigated via expensive engineering techniques, this ought to make us more persuaded of its truth. But of course the relative cost of obtaining particular evidence has no bearing in itself on the weight that should be assigned to it – that depends entirely on the logical relationship between the findings and the proposition to be proved, and since we are told nothing yet about the results of the project sponsored by the Dutch geotechnical firm FUGRO, we have no reason yet to find the Ithaca=Palikí theory more convincing than before.
The real difference introduced by moving to resource-intensive industrial research methods is that it creates strong motives for not allowing the theory to die, whatever the true balance of evidence may be. While a handful of knowledgeable people explore an attractive idea by reading the texts and walking the terrain with cameras, favourable and adverse evidence can both be given their true force; if in the end it seems the theory is mistaken, no-one has lost anything beyond a pleasant dream. After real money has been invested, people will be needing to see real returns on the investment. Industrial-scale research leads to deployment of industrial techniques of marketing and public relations to maintain the value of the research product. (Already the way that we are made to wait for months until all is revealed about the FUGRO findings smacks of the arts of showmanship rather than the sober methods of scholarship.)
Summing up: as I write in June 2008, the Ithaca=Palikí idea is supported mainly by an attractively-produced and widely publicized book whose argument contains a glaring hole that, we are now told, its authors were not aware of; and further research findings are said to be on the way, but we are given no hint about whether they add much (or anything) to the case. What is more, the truth of the theory would require us to believe that Odysseus’s Ithaca was the site of a natural phenomenon without parallel in the present-day world; if it did exist today, it would presumably be as famous as, say, Niagara Falls.
The reasonable conclusion must be that it was a good try but sorry, Ithaca ≠ Palikí!
— So much for the geological side of the “Odysseus Unbound” theory. But, in case any reader feels unconvinced by my argument about the implausibility of the channel that is said to have made Palikí an island, let me add that one can take issue with Robert Bittlestone’s case even on textual grounds, without going into geological questions. Bittlestone quotes passages from the Odyssey which fit his theory, but there are other passages which do not. In Book IV, for instance, when Telemachus is visiting Menelaus in Sparta, he declines the gifts of horses and a chariot which Menelaus offers by explaining that he could not use them in Ithaca:
In Ithaca we have no broad riding-grounds, no meadow land at all: of these our islands which rise rock-like from the sea, not one is fit for mounted work, or grass-rich: least of all my Ithaca. (T.E. Lawrence trans.)
Probably none of the Ionian Islands were good places for horses, but it seems odd to say that Palikí was least so of all (περὶ πασέων); Bittlestone has been at pains to point out that Palikí is relatively low-lying (because, in another passage, Homer describes Ithaca as low-lying). One might think that Telemachus’s words to Menelaus would fit Itháki better. Certainly the writer Lawrence Durrell took them to do so:
Ithaca [i.e. modern Itháki], which reverberates with the Homeric legend, is a delightfully bare and bony little place, with knobbly hills, covered in holm oak, which come smoothly down into the sea … Nothing could convince you more that this was the island of Odysseus than recalling it while actually on the spot: “It is a steep little island impracticable for horses …” (The Greek Islands, Faber and Faber, 1978)
Durrell is a novelist and does not pretend to be making a scholarly contribution to the “Ithaca question”, but as a common-sense observation by someone intimately familiar with the Ionian Islands his remark is worth respect. (Durrell was well aware that not everyone accepted the identification of Homer’s Ithaca with modern Itháki.)
Similarly, Homer repeatedly describes Ithaca as κραναός or τρηχύς, words which Liddell and Scott’s standard Greek dictionary glosses as “hard, rugged, rocky” and as “rough, rugged, rocky” respectively. A low-lying island may be stony (the southern end of the Palikí peninsula is notably stony today), but it is surprising to find Palikí described as distinctively rugged compared to much steeper islands such as Itháki. Book XVI of the Odyssey describes Penelope’s suitors as
The pick of all the island gentry from Dulichium and Same and leafy Zacynthus (not to mention rugged Ithaca [κραναὴν Ἰθάκην] itself)
Would that list not be easier to understand if Ithaca were taken to be modern Itháki?
A passage which Bittlestone does discuss (on his pp. 89–90) is the one in Book XV where the goddess Athene is advising Telemachus to avoid capture by the suitors on his return to Ithaca by getting his crew to land him at the “first cape” or “first beach” of the island (πρώτην ἀκτὴν Ἰθάκης), before they sail on to the town. According to Bittlestone, this meant that Telemachus was to be dropped off at a bay in the north-west of Palikí, after which the crew were to continue round the northern end and down through the hypothesized channel. But when Athene gives Telemachus this advice, they are in Sparta, far to the south-east of the Ionian Islands, and his ship will approach Cephallonia and Palikí from the south-east. From a common-sense point of view, the “first cape/beach” of Palikí would be somewhere at its southern end. Bittlestone’s Figure 10.6 instead represents Telemachus’s ship as following Athene’s advice by sailing past that end of Palikí and almost all the way up its long western side before reaching the bay where he thinks Telemachus landed. Something is surely wrong there, if Homer knew the geography of Ithaca as intimately as Bittlestone supposes.
But that is the point. All these inconsistencies, including the fundamental one about whether Ithaca was to the west or to the east of its island neighbours, are only troublesome on the assumption that Homer was familiar with the Ionian Islands. Bittlestone argues that Homer was actually an Ithacan himself, writing primarily for his fellow-islanders, who would have been delighted to hear allusions to the individual paths and smallholdings of their home. That is quite contrary to the traditional view, though, according to which Homer came from the eastern edge of the Greek world – either from the island of Chios, or from Smyrna (now Turkish Izmir). This tradition could of course be wrong; but experts tell us that various features of Homer’s language suggest an eastern background, and would not fit a birthplace in the Ionian Islands. (I believe that one example would be the word τρηχύς for “rugged”, quoted above; the vowel of the root would be alpha rather than eta, if Homer were a westerner.)
If Homer did come from Chios or Smyrna, then for him the Ionian Islands were about as remote as anywhere in the Greek world. If he wanted to take Odysseus’s homecoming as his theme, he was forced to place it there, because the body of legends he was drawing on presumably said so. (Legends about the Trojan War, sections of which Homer used as the basis for the Iliad and Odyssey, were clearly well known in outline independently of Homer; for instance, although Homer alludes briefly to the Wooden Horse episode, for the details one has to go to other writers.) But if Homer was not personally familiar with the Ionian Islands then he would have had to describe the geography as best he could on the basis of hearsay, and it would have been natural enough for him to make mistakes. Perhaps he just forgot which was further west of Ithaca and Cephallonia. He could not check it in an atlas, as a modern writer would.
(There is a large potential source of confusion in that the now-Turkish coast of Asia Minor was called “Ionia” when it was Greek. The dialect of that area was called Ionic, which sounds in English as though it ought to be specially related to the Greek of the Ionian Islands; and Bittlestone thinks it was. But these names only happen to look similar in English because we have just one letter O to represent two quite different Greek vowels. In Greek the names are entirely unconnected – Ionia in Asia Minor was named after a legendary man called Ion, the Ionian Islands after a legendary woman called Io; and the dialect of Greek classically spoken in the Ionian Islands was definitely not an Ionic dialect.)
It is much easier to believe that the Odyssey was composed by a poet who was hazy about the geography of Ithaca, because he had never been there, than that the Palikí peninsula was once separated from Cephallonia by a channel unlike anything in the modern world.
Furthermore, the geographical mistakes have no importance. They do not go to the core of what the poem is about. The Odyssey has appealed to readers for thousands of years because each of us has his Ithaca: it is what A.E. Housman called
… the land of lost content,
I see it shining plain,
The happy highways where I went
And cannot come again.
We love to hear about Odysseus, because he achieved the impossible and won his way back to his own land of lost content. The significance of the poem lies in its human resonances, as well as in things like Homer’s vignettes of nature – it does not lie in details of geography.
The Trojan War legends may have had a historical basis (though we do not know whether they did). Possibly there really was a man called Odysseus. But when Homer worked up legends about him into a long, detailed poem, I very much doubt that each footpath and headland mentioned in the poem was for Homer a particular, identifiable path or headland. They were generic landscape features, invented as the story called for them. To think otherwise seems as misguided as it would be to enquire about the precise latitude and longitude of Prospero’s island in The Tempest.
[Added in Oct 2008:]
On 2nd Oct 2008 John Underhill presented the Ithaca=Palikí argument in a very enjoyable illustrated lecture at the Geological Society. I put it to him that nowhere in the world is there any present-day parallel, other than man-made canals, to the sea-channel which he and his colleagues postulate as having made Palikí an island in Homeric times, and I asked what process he envisaged as having created such a landform. Underhill accepted my point that the channel lacks any non-artificial present-day analogue; but he argued that in the Homeric period the postulated channel might have been partly excavated by human action – if rivers had reduced the isthmus to a narrow neck, it would have been attractive to people at the time to cut through the neck and create a navigable passage.
One must be open to all possibilities, but is this idea seriously plausible? Relative to the sparse population and primitive technology of the time, creating such a waterway would surely have been a truly mammoth enterprise, even if the isthmus was lower and narrower than rockfalls have since made it. People might undertake such a task if the practical gains would be correspondingly great. In modern times the Suez and Panama canals massively reduced the length of many economically-significant shipping routes, and even the smaller-scale Corinth Canal shortens routes between many places to the west and many places to the east of the Greek region. On the other hand, the channel postulated between Palikí and mainland Cephallonia would be relevant only for shipping between places within or just outside the Gulf of Argostóli at the southern end of the channel, on the one hand, and places north or north-east of Cephallonia, such as the islands of Corfu or Lefkás, on the other. Any other water-borne traffic near Cephallonia would gain nothing by using this canal, if it existed. Could there ever have been enough relevant traffic to motivate the creation of such a canal? It is hard to believe that; Corfu is already near the northern edge of the ancient Greek world. In any case, if there really had been important trade routes northward or north-eastward from Cephallonia, then rather than leading to a shortcut being created into the Gulf of Argostóli, this trade pattern would have caused the main ports of Cephallonia to develop elsewhere than within that gulf.
Much as I enjoyed John Underhill’s Geological Society lecture, I am afraid that it did nothing to allay my scepticism.
last changed 8 Oct 2008