Imagine a noble building in ruins, ruins so ruinous that its original purpose is unclear. And what style is it in? Even the floor plan is half erased by time. In the late 19th century restorers get busy on the house. They tudorise it, classicise it, gothicise it and finally, in a wild spirit of appropriation, they modernise and postmodernise it in various styles; Gropius, Wright, van der Rohe and finally Hadid all have their way with it, none of them sure of the original idiom, each following a different theory.

Such a ruin is Gilgamesh, the oldest surviving long poem so far discovered. It has been dubbed an epic. Is it? By analogy with later poems in the epic tradition – Homer’s, for example – it has been declared the product of an oral tradition. Is it? It remains a poem comprised of clay fragments, short and long, and though the desert delivers up occasional additional text, we are a long way from a whole poem.

We are a long way from a whole story, too, though there are strong narrative lines that can be followed. Gilgamesh is a great king of Uruk. He is also a philanderer who claims first-night rights over his subjects’ brides. The women enjoin a goddess to help tame him and she creates Enkidu, a creature entirely of nature, who romps with the animals, saves them from hunters and grazes on nature’s goodness. Enkidu is seduced by Shamhat, a temple harlot or hierodule, and this alienates him from his wild friends, who are repelled by the reek of the human. He goes to Uruk and confronts Gilgamesh, and they have an almighty fight, which neither wins. They become close, even intimate friends and go off adventuring, slaying Humbaba, the guardian of a great forest, and bringing building materials back to the city, alienating the gods at the same time. Thereafter, the poem becomes obsessed with mortality. Enkidu dies; Gilgamesh laments and then goes in quest of eternal life. He visits the immortal survivors of the great flood, which prefigures the one in the Bible. At last, reconciled to his fate, he returns to Uruk, a good deal sadder and a little wiser.

The fragments are in different languages, written down over more than a millennium, from around 2100 BC onwards, scattered under a territory that runs from Turkey to Iran. A lot of sand. Most of the languages in which the poem was recorded were lost and have been laboriously reconstructed, starting in Victorian times, by some of the most brilliant philologists the world has known, working from the tablets found in 1853 in the ruins of the seventh-century BC library of Assurbanipal, the last king of the neo-Assyrian empire, at Nineveh (near present-day Mosul). It has mostly been a European project, with major work done by German, French, British and latterly American and Middle Eastern scholars. And though we know a lot, volumes and volumes of scholarship, we are still a long way from understanding the poem fully.

Yet reading through English versions (whole versions and lyric riffs) that have appeared over the last century, it is startling to see what very different verse forms, dictions and themes are teased out of what is there, as though there were no text, only pretext, a pretext that allows for almost anything so long as the narrative line, such as it is, is maintained. The same, I am told, is true of versions in other languages.

But even the narrative line of the poem is subject to debate. The first and most popular English-language translation, into a prose rendered in an idiom reminiscent of the King James Bible and Book of Common Prayer by Nancy Sandars for Penguin Classics in 1960, plays false with sequence in the interests of making the poem ‘relatable’, graspable. Sandars did away with the verbatim repetitions that characterise the text. In her version, Gilgamesh is assimilated into the sweep of the familiar; it has more or less good manners and merits a seat at the high table with later classics.

As a nonspecialist writing Gilgamesh: The Life of a Poem, I was startled at how many roads led out of ancient Uruk, where the story begins and ends, where we are asked to wonder at the stability and grandeur of Gilgamesh’s city and to walk around the well-appointed walls he has built to defend and dignify it. The poem leads us, but no person takes us by the hand. It is not like Virgil or Homer, whose poems start with invocations to a muse (and the implication of a narrator). This poem voices itself, one of its wonderfully strange properties.

Scholarly translations are marvellous playgrounds because they stay close to the text and in their numerous footnotes puzzle at the cruxes and aporia and show the gaps and holes, how the poem is reconstructed out of fragments from different corners of the ancient world. Benjamin Foster, professor of Assyriology and Babylonian literature at Yale University and one of the commanding scholar-editors, declares, ‘I have no patience with clueless folk who think that they can translate the epic without going to the trouble of mastering Babylonian, though of course they are welcome to retell it.’ Gilgamesh has been subjected to dozens of whole and partial translations into English over the last century, only a very few of which qualify as translations in Foster’s terms. For the Assyriologist the peril of error is everywhere. The transitions between and within languages create such ambiguities that it is not quite clear whether, for example, a balag or balagu is a drum or a lyre. Are the monster’s teeth tusks? Is the flood ark ship-shaped or cubic or spherical? (The charismatic Assyriologist Irving Finkel, bearded like Noah, has had a lot of fun working that one out, and even had a to-scale version of the ark built and sailed.) Not every seeming error in the friable text is an error, and what appears common-sensical can be quite wrong. If there are two or more possible meanings, it may be that the ambiguities are intended and solutions have to be found that preserve them.

A translator attends to the call of language, and is drawn out by it. If a translator is ignorant of the original language, what authority is there? But for the translator freed from the responsibilities to language, the road ahead is clear. C H Sisson described translation as ‘fishing in other men’s waters’. The languageless translator simply drains the stream and harvests the fish by hand.

Nowhere are the poetic and moral vices and virtues of the age better demonstrated than in the translations of the passage in which Shamhat, the temple harlot, is positioned on the bank of a lake where the beasts come to drink. Her legs are spread wide and she awaits Enkidu, who duly comes. In the original poem the sexual encounter, while epic in proportion (six days and seven nights of lovemaking), is recounted relatively succinctly. The early translators were embarrassed and bowdlerised the text. In 1928, Shamhat ‘Shew’d him her comeliness’ and ‘ravish’d the soul of him’. By 1955, she is being told to ‘Free thy breasts,/Bare thy bosom that he may possess thy ripeness’. By 1970, in an awkward free-verse translation, Enkidu’s ‘hands go out, his fingers search:/And she takes his hands and presses close/And guides him into the seat of love’. Shamhat leads him on and in. Recent translations are more inventive, to the extent that some are worthy of the Bad Sex in Fiction Award, adding circumstantial detail, smells and textures. Some translations successfully privilege Shamhat’s perspective, or show her taking no pleasure in her vocation. My favourite version is one of the most recent, by Philip Terry, who re-creates the poem in Globish, a language with a limited vocabulary that was devised for trade. ‘Open | you leg | show WILD | MAN you | love box/Hold no | thing back | make he | breathe hard’.

Professor Andrew George of SOAS (my chosen guide in most Gilgamesh matters, whose translation is the most compelling) and other Assyriologists, in the detail of their scholarship, release the mischievous djinn of ‘translation at one remove’ from the fractured lamp that is the poem. Their footnotes fill would-be translators, and critics like myself, with an illusion (is it pure illusion?) that we can get close to the original and even, mouthing the transliteration, have a distant familiarity with the sound of its language or, rather, the sounds of its languages.

Gilgamesh reads us rather differently from the way Homer and Hesiod do. After all, a lot of people have at least a smattering of Greek. In the case of Gilgamesh, even the language – that most complex set of structures, rhythms, patterns, etymologies and other connections – loses force: none of the poet-translators (and few of those who write about it, myself included) has mastered it. The centuries of tradition and scholarship that inform our reading of other classics are missing, as is its creative presence in our literature before the 20th century. The poem comes to be read with an appropriative, anachronistic eye; it belongs peculiarly to the age that encounters it.

Our age lacks literary coherence and nowhere is this more apparent than in translation. Gilgamesh is welcomed in a bewildering number of ways. It is as though not one damaged and fragmentary poem has arrived, but dozens of poems, quite different in character and inflection. There is a reduced risk of charges of cultural appropriation when the source culture isn’t there to answer back.

A more challenging approach to the original, even at a few linguistic removes, is to follow the scholar’s route – rich, luminous scholarship is readily available – and approach it in an exploratory spirit, looking at what is there, to see what it proposes, humbly coming to terms with it rather than compelling it to come to terms with us.                                                                        r