In my book The Echoes of Love, the hero, Paolo, is a something of a storyteller – though he is not a writer, he certainly has the mind of one. Thus he takes an interest in the local folklore of the places to which he travels. So when he tells Venetia of one of his favourite Italian islands, Sardinia, he can’t resist saying:

‘In between myth and history, there’s the theory that Sardinia could be the lost Atlantis.’

Venetia’s reaction is to laugh. For Paolo is teasing her. Surely? Atlantis is a fictional island, after all…

Actually, Paolo is quite serious!

It was Plato who first mentioned Atlantis, in his 360 BC dialogues Timaeus and Critias. It was, he said, located beyond the Pillars of Hercules, 9,000 years before his time, but was destroyed by a natural disaster, possibly a tsunami.

The majority of scholars believe that the city of which he spoke was never meant to be construed as a real place – he invented Atlantis as the antagonist that provoked, and was subsequently defeated, by Ancient Athens. It was part of an allegory, a story meant to impart wisdom.  But the notion of Atlantis inspired many writers who followed, and a small but persistent school of thinkers believe Plato was in fact writing of an actual place. Sardinia?

Italian journalist Sergio Frau argued so. The basis of his theory is as follows:

Plato said Atlantis was beyond the ‘Pillars of Hercules’. The ancient writer Erathosthenes wrote that the Pillars of Hercules were at the Straits of Gibraltar. That would mean the city was in the Atlantic Ocean – clearly impossible; hence the ‘city under the ocean’ legend.

But what if Erathosthenes got it wrong? Frau believes that the Pillars of Hercules were actually on Sicily. As one newspaper reported:

Frau had his brainwave after seeing a print of two maps of the Mediterranean as it was in the Bronze Age. One showed Tunisia and Sicily almost touching; the other, of the Straits of Gibraltar, was remarkably similar . Frau thinks Erathosthenes moved the pillars because in the 120 years between Plato’s era and his, the Greek world changed dramatically, and the strait between Sicily and Africa was no longer at the outer reaches of the Empire. Furthermore, geological shifts and rising sea levels widened the distance between Tunisia and Sicily, contributing to Erathosthenes’ mistake and reinforcing it over time.

If the Pillars of Hercules really were in Sicily, Sardinia would be the obvious location for Atlantis.

Interestingly, moving the location of the Pillars of Hercules means that other classical writings make more sense as well. And the existence of the Nuragic civilisation on Sardinia during the Bronze Age may add credence to the theory: they lived on Sardinia 900 years before Plato’s time, and were wiped out by a natural disaster, possibly a tsunami. Was 9,000 a simple error?

Frau’s theory has caused heated controversy among historians – some 250 of them came together to denounce his claim. Yet his book sold 30,000 copies in Italy, and was the catalyst for a UNESCO symposium on the issue.

Certainly, I think the theory is worth serious consideration. Just as we don’t assume that all Bible stories are fictional allegories with no basis in true events, why dismiss Atlantis as fiction, not fact?

I wonder, though, how we would feel were Atlantis to be wholeheartedly declared the true Atlantis. Sardinia is a wonderfully romantic island, of course – such a beautiful place. As Paolo says in The Echoes of Love, ‘It’s an earthly paradise, with a wealth of secluded places and open spaces to take your breath away, really.’ But it is unequivocally real – tangible. It’s no ruined city on the seabed, nestled in the blue and entwined with seaweed, kingdom of the mermen and mermaids.

Perhaps Sardinia is Atlantis. But if it is, it is Plato’s Atlantis – not our twenty-first century Atlantis, born of centuries of imagination and dreaming. But then that Atlantis isn’t lost at all: it exists in myriad books and poems and paintings and films. We lost Plato’s city, but we built our own in its stead.