Writing the biography of a living person, there is the occasional but disconcerting sense of re-living your own experience of the twentieth century through another's eyes. One advantage, of course, is that it is often possible to explore your differing perspectives through conversation or interview. At the age of ninety, Edwin Morgan still feels disappointment at the reception of The New Divan (Carcanet, 1977). Perhaps anything would have been a disappointment after the remarkable success of From Glasgow to Saturn (Carcanet, 1973), the collection by which he is probably still best known beyond Scotland. He is willing to admit that the title was wrong (his editor Michael Schmidt used to call it Nude Ivan), and that he himself had mainly argued for the organisation of the book as it stands, without realising the difficulties that the opening long poem, also called 'The New Divan', would pose.

He had perhaps been misled by the ready acceptance of many of its 100 stanzas in a variety of poetry journals - but those editors were only seeing interesting fragments of a forthcoming work by a recognised writer, and so were happy to have them. Between 1974 and 1976, he had published most of the poem in sections of between three and ten stanzas in Ambit, Aquarius, The New Review, The Honest Ulsterman, Lines Review, Poetry Book Society Supplement, Meridian, Broadsheet, Poetry Review, Akros, Oasis, Sphinx and Iron. But these sections did not follow the order of composition, which is to say the order in which we have the poem in The New Divan, and they provided a sample of its rhetoric without any sense of its ultimate form. The stanzas were about sonnet length, sometimes a couple of lines more or less than that, but packed with a dazzling Middle Eastern imagery of desert and city.

Morgan had been dwelling upon his wartime experience since the beginning of the 1970s, when he was fifty years of age. During military service in the Royal Army Medical Corps from 1940 to 1945, stationed in Egypt, the Lebanon and Palestine, he had been too involved in new life experience within ancient landscapes to write about this effectively. He was then only in his early twenties, developing a new understanding of life rather than writing about it. But in middle age, memories and flashbacks began to disturb him. As he describes the 1970s in his poem 'Epilogue', which is a survey of the decades of his life, the writing of his long poem of that war becomes the major artistic concern of those years:

At fifty I began to have bad dreams
of Palestine, and saw bad things to come,
began to write my long unwritten war.
I was a hundred-handed Sinbad then,
rolled and unrolled carpets of blood and love,
raised tents of pain, made the dust into men
and laid the dust with men. I supervised
a thesis on Doughty, that great Englishman
who brought all Arabia backin his hard pen. ('Epilogue')

Looking back in 2009, Edwin Morgan claimed that the 1970s were a 'blank' in his mind, compared with the marvellous 1960s. This was the infamous decade 'when the lights went out' in Britain, the years of the miners' strike in the winter of 1972, of the three-day working week in factories following the disruption of fuel supplies during the Yom Kippur War in the Middle East, of a remorselessly rising death toll in Northern Ireland, of tumult in the economy and strikes in essential services that left refuse uncollected and the dead unburied in the 'Winter of Discontent' of 1978-9. So there was indeed much to blank out in the social and political world of the time. Like the rest of the adult population, it was a matter for Morgan of just trying to get by - struggling to get his examination papers marked in the intervals between power cuts, or discovering the problems of typing by candlelight (an excess of shadows).

But there was more than just coping with inconvenience. This is perhaps signalled in 'Epilogue' in the odd way he transposes into the 1970s 'a thesis on Doughty' that was actually begun in the mid-1950s and presented in the early 1960s. Mohamed Kaddal was an Egyptian student who was forced to return in haste to be with his family during the Suez Crisis of 1956. All of the Egyptian students at Glasgow University were leaving, and one had lost his entire family in an air raid on Cairo. Morgan gave Kaddal his own copies of books by Doughty to take with him, in the hope that he might complete his thesis in Egypt. He finally returned to Glasgow in April 1960 and, helped by frequent meetings with his tutor, completed the thesis in less than two years, submitting it in February 1962.

The original title of the research was C.M. Doughty: Search for the Form of an Epic, and it is this focus on epic form, as well as the Arabic background of Doughty's best known work, Travels in Arabia Deserta (1888), that links this research supervision with Edwin Morgan's own later attempt in the 1970s to find a suitable quasi-epic way of writing about his own war experience. This would be his 'hundred-handed' long poem, 'The New Divan'. He had been unable to use this experience artistically at the time, and now its memories were returning to haunt his imagination.

His mother had died suddenly in 1970, and we might now think that her death possibly intensified his memories and dreams of wartime Palestine, since that had been his first real separation from her. The 'hundred-handed Sinbad' of the 'Epilogue' seems to revert to a childhood love of fantastic tales, but the hundred stanzas of 'The New Divan' would also unroll 'carpets of love and blood' and raise 'tents of pain'. With such serious adult aims, the poet was nonplussed that what seemed to him a major work, exploring time, history, homosexual and heterosexual relationships and the Eastern arts of storytelling and verse, did not achieve immediate recognition.

Nor has it still. I believe that this has something to do with its form. Morgan's emphasis in subsequent interviews on its 'Arabic' method of interweaving images, echoes and quasi-autonomous delights may well have distracted attention from whatever structure this long poem actually does possess. He argued that it possesses 'a kind of structure' but that it also uses

a kind of randomness in the sense that one is not following a story that really goes forward step by step. Characters appear and reappear. You're not certain whether the characters are autobiographical or not... And, if you like, the idea of non-structure is almost a structural idea in itself - in a sense that a great deal of poetry of the Middle East (which I got to learn something about when I was in the Middle East) deliberately is anti-structural... In Arabic or Persian poetry they're rather fond of the idea that a 'divan' as they call it, a collection of poems, is something that you enter; you move around; you can cast your eye here and there, you look, you pick, you perhaps retrace your steps.
(Nothing not Giving Messages)

Although he admits that this procedure is not something a critic could easily analyse in structural terms, nevertheless there would be 'something that in a mysterious, subterranean sense would be a structure, an emotional structure, a structure perhaps relating to the life of the person who had written it'.

I don't find this argument wholly convincing, but can admit to some bias here, since the topic of my own doctoral research with Edwin Morgan in the 1970s was on form and structure in the modernist long poem. I was focusing on the poetry of Basil Bunting, whose long 'Sonatas' were underpinned by a quasi-epic structure (I argued), although this tended to be hidden by the musical analogy he used. My supervisor was too polite to tell me his own view of Bunting at the start - as I now learn from a letter of 20 November 1971 to his Carcanet editor:

Bunting is also not one of my absolute favourites, though I can see he has a gluey music all his own - but I like a verse that moves. It's not enough to have images, fine though many of his are. On the other hand, his quirky crabbitness is also a 'northernness' and in some moods this attracts just because at least it is not smooth... From my window I am watching the snow fluttering past the orange streetlamps like - like - come on Bunting - like
           doom-grated
           orange-peel
           puffed through
           fine fans.

Bunting was a conscientious objector in the First World War but an enthusiastic fighter in the Second, rising to the rank of Squadron Leader in the Intelligence services, based in Persia, Italy and North Africa. He knew the Persian language well, and his long war poem 'The Spoils' (1951) deals specifically with Islamic culture and the North African campaign in its separate sections. This may have spurred Morgan on to emulation or rivalry with his own long poem of North African experience in 'The New Divan'.

I like to think he had altered his view of Bunting, to some extent at least, by the end of our continuing conversation on literary archetypes and plot structures. It was a protracted discussion at a distance, since I was teaching full-time in Dumfriesshire and helping to raise a young family. So I would send him parts of the thesis to read, and they would be discussed during holiday visits to Glasgow. These tutorial visits, I now know, coincided with how he wrote 'The New Divan': in bursts during the Christmas, Easter and Summer vacations between December 1973 and July 1975. On one occasion he showed me the section he had just completed, not asking for any comment but just because that was what he was working on at the time.

He thought the completed sequence was significant enough to use it also as the title of his third collection, The New Divan. As editor, Michael Schmidt needed some convincing about that. He said it reminded him of Canova's painting of a mistress of Napoleon III in the Galleria Borghese in Rome. But Morgan had had enough confidence in the work to begin publishing sections of it in various magazines even when it was still far from completion, and it seemed to have been welcomed by editors. In their correspondence of the time, it emerges that Michael Schmidt also liked the Middle East very much, enjoying Egyptian and Arabic culture. Both poets were pro-Arab, the younger poet-editor having once been confronted at close range by an armed Israeli soldier. On 14 April 1972 Morgan commented: 'How strange that we should both be feluccaphils. Or Middle Easters. Of all these countries Lebanon is my favourite.'

What then of the structure of what Edwin Morgan clearly saw as a major work? It had been bubbling up in dreams and nightmares, as well as in correspondence with other ex-servicemen, for several years (for example, with Alex Frizzell, his publisher at Castlelaw Press who published his Glasgow Sonnets in 1972). There was an artistic and personal need to make a full statement about that wartime experience, and to replace his surrogate 1950s translation of the Beowulf tale of warrior bands against northern monsters with a more autobiographical work. But how were the varied memories and experiences of this Mediterranean story to be organised?

Although he disliked analysing his own poetry, feeling like many poets an almost superstitious respect for its mystery, he knew how to make a long poem in the structural sense, out of sound and thematic patterns. He had edited the Albatross Book of Longer Poems (Collins, 1963), and had teased out in the Introduction to that volume some of the issues involved: 'One might say that its average length would be between 100 and 1000 lines, and that it tended to have a narrative structure; that it came somewhere between the compressed intensity of the lyric and the ambitious grandeur of the epic' (p. 9). It was often seen as a self-challenge to the poet, 'a test of Invention' (as Keats had stated when writing Endymion), but also, to quote Edgar Allan Poe in his 1948 lecture on 'The Poetic Principle', as 'a flat contradiction in terms', where length and plot demands overwhelm the true lyric intensity and unity. Poe thus considered that 'at least one-half of the Paradise Lost is prose'.

Morgan returned to these issues in 'Long poems - but how long?', his 1995 W.D. Thomas Memorial Lecture in the University of Wales. Surveying contemporary approaches to the long poem, he cites Lyn Hejinian's Oxota (1991), with its 270 chapters each of which is a fourteen-line poem like an unrhymed sonnet, as an example of the contemporary option of the long poem as a sequence. He also mentions Hughes's Crow, Berryman's Dream Songs and Lowell's Notebooks as similar approaches to the challenge. He refers to Wolfgang Iser's discussion of the need for 'indeterminacies or blanks', where the absence of connections is crucial for the constitution of an imaginary object in the art of reading. All of this reflects upon his own approach in 'The New Divan', although articulated some twenty years after the writing of his own long poem.

He had begun, as he told Michael Schmidt, with twenty poems. These were written over seven days of the midwinter season from 28 December 1973 to 3 February 1974, as noted in his archive in Glasgow University Library's Special Collections Department. It seems at least feasible that he would use this twenty-stanza pattern as a building block.

But before even looking at his original typescripts where the dates of composition are given, I re-read the poem itself, in the Carcanet edition of 1977 which Edwin Morgan gave me in July of that year, and tried to spot the structural elements he had used. Overall, of course, he was using an 'Arabian Nights' technique in keeping with the Middle Eastern setting. There, the marvellous tales of fantastic journeys and the unrolling of many a magic coincidence are set within the life-crisis of the teller, the clever Scheherazade. In order to save herself from execution by a king who hates all humankind, every evening she tells him a story and leaves it incomplete, promising to finish it on the following night. Hence, The Thousand and One Nights or The Arabian Nights Entertainment, the alternative titles. The king, anxious to hear the ending, postpones her execution from day to day, until finally he abandons his cruelty. I am not sure when Morgan first encountered The Arabian Nights Entertainment, but I do know that he was reflecting on traditional tales in 1974, because he gave my children his annotated review copy of Iona and Peter Opie's The Classic Fairy Tales (1974), after he had completed 'The dark side of Fairyland' for the Times Literary Supplement (6 December 1974). There are several comments in this book regarding the Middle Eastern tales.

Looking at 'The New Divan' stanzas not as a series of separate stories but as a series of strong sense impressions and images of contrasting cultures in conflict (flashbacks, as it were, from several different narratives from different eras) I read it again. The marvellous journey is, of course, the standard 'epic' device, whether in Dante, Basil Bunting or Ezra Pound. The story of that journey is the poem, but with the storyteller's voice and stance also brought to the fore. In 'The New Divan', the journey moves towards a combination of visionary clarity and autobiographical testament in its final stanzas, giving a sense of uplift that abides beyond the pain of wartime loss that must be endured.

But how was this positive emotion to be arrived at? Stanzas 1-20 set the scene, or rather the scenes, of the tale. It opens with the Persian poet-figure Hafiz as the 'old nightingale' who has, in the historical past, achieved his great poetry within the 'wretchedness' of life. The modern poet now must enter history like a masked dancer, or take on multiple identities like an actor, keeping 'no form but water' to deal with complex reality. Eastern princes, a dancing bear, rough angels in their unmade beds, slow caravans and white sails on the turquoise Gulf of Arabia are images that provide movement and a sense of unknown destinations. There is a ship, a voyage: 'We pound and shudder through the dark. / A drunk at my cabin door thinks it's his. / Everything's confused ... But there is 'No going back, no one to go back with, / back to.' Then 'Halfway down the road', as Dante was at the start of his epic (although unlike Dante in the company he keeps), the speaker meets a visionary creature, possibly 'just a shade', a ragged wraith under a date-palm. He cackles that 'We're / on the same road'. The scene is set, as in a play. But this is a poetic theatre with multiple stages - possibly more like an editor's suite in a television studio where different stories are playing simultaneously on many screens, and we are flicked from one to the other.
 
Nevertheless the Dante comparison is not irrelevant. The poem does end with a 'paradise in prospect', although it may be a mirage. And the violence of war has its infernal aspects, just as the pain of lost love creates a period of purgatory that must be endured. What is lacking is the certainty of Dante's categories. Here the images flicker and rewind and intersect. Love has both divine and demonic tendencies in Morgan's modern vision of things.

Section Two's stanzas 21-42 were composed between 7 and 11 April 1974, when ten poems were written, and then between 28 May and 16 June 1974, fairly intensively in bursts of three or six poems. That looks as if Easter and then examination marking have interrupted the April flow. The theme of this second section is Exploring the Arabian world - striking more deeply into its weathers, markets, boats, young Arab gangs ('Real nomad cats / with nose-rings, knives, horsewhips ...'). There are cockerels, calabashes, parakeets and a hunting horn, all brought on board by 'the master of ungathered things'. There are see-through demons and ruins under snow, and a desert excavation of a king's body: 'His whole retinue had been burnt like coal / to warm his afterlife'. Ancient civilisations and living crowds jostle in the same space, and within this human chaos the key question is posed: 'Where can love, in this world, ever lodge?' In an image that recalls Bunting's poem 'The Spoils', Faruz, a female Arabian singer, laments on the radio 'the final / flake and loosened quiver, winding down, of love'. This is where the journey has taken us so far.

Although he had completed stanzas 40-42 in that preceding period, my reading of the third section placed its starting point at 40. A frightened gerbil being stroked by a child is so anxious that 'it trusts its cage more than love, that like wax / can melt to fearful shapes and suffocate'. This third movement of the poem, 40-59, deals in more detail with Dimensions of love, both heterosexual and homosexual. Earlier images of voyage and archaeological excavation recur, floating by like the clouds 'boiling in swarms'. This is also a 'speeded-up universe', a time-travelling test of endurance where we enter gardens, rooms, and lives. Looters in the market squares and desert bandits co-exist with the cosmic perspective of mysterious calm observers of human activity: 'This world, they say, has only action.'

Set against that world of impulse and desire, the fourth section, stanzas 60-80, explores more deeply The force of commitment, in its political, religious and also artistic forms. The storyteller's role acts as a sort of hinge, in fact, in stanzas 67 and 68, where stories within stories mark a place where the poet had to begin again after a month's gap caused by the new university term. It is almost as if he is reminding himself of what the poem is about, and what his own artistic role must be within it. Stanzas 62-66 were written between 20 August and 18 September. Then he manages another three stanzas between 30 September and 1 October. After a month's gap for the start of term, he completes nine stanzas between the end of October and mid-November.

This section opens with 'an activist grown old', walking in his garden: 'the admired, / limping, white-haired child of the revolution, loved / by the soft generations smells blood, not roses, still.' Cheap religious statues for Christian pilgrims are mocked ('nothing looking / blander ever shot from the mould') and 'demons fill the cracks with smoke'. There is often a bitter cast to the results of political commitment in wartime violence: 'By searchlights men were made the maggots of / a turned furrow, each waving with his / scattered O of panic.' Somewhere there may be a recording angel, but he is only writing down a 'vast perhaps'. The final nine stanzas, completed after the month's break, are much more physical, with a sexual atmosphere and sharply observed domestic detail that perhaps sets up the poem for the final loss soon to come: 'That day was an onion of skins, / it came apart in temporary / hours, distinct, sweet, pungent and good.'

The last section of the poem, stanzas 81-100, was written mainly in two halves: 83-90 between 24 March and 10 April, and 91-100 between 25 June and 20 July 1975. Again this is writing in the breaks from teaching. It offers A final vision: 'Paradise in prospect / after sand after sand after sand.' It may be a mirage, but the pilgrimage goes on. Sages smoking their bubble-pipes in an anteroom of heaven, recording angels, and a flying carpet snapping its fringes are all intersected with demonic images of death and torture ('in the form of dragons, worse. Winged leather, metal beat about us'). Love, unrequited but faithful, is the end of the journey, although it is 'Not in King's Regulations, to be in love'. Cosgrove is named, recreated and recalled forever: 'yesterday, tomorrow / he slumbers in a word.'

The physical fears and excitements of war are suddenly brought into full light. The poem places the individual transient life of the young soldier-poet, with its intense friendships and its painful encounters with the dead and injured on stretcher duty, together with his later dream-visions and survivor nightmares, confidently but quietly alongside the great Persian poet's work. The final image is of a shred of sailcloth, washed up after a shipwreck, that can still be used to bind a collection of stanzas, a 'divan' in the Persian style.

It is possible, therefore, to discover a formal structure present within this long sequence that helps explain why it is shaped as it is. But this formal presence does not of itself guarantee an enthusiastic readership (should poems need a critical gloss?). And The New Divan certainly sold less well than his previous collection. Sales were pretty good, his editor said, and compared with many books of poems they were excellent. But From Glasgow to Saturn remains the book by which Edwin Morgan is best known in England, it seems.

Michael Schmidt was probably correct in his initial reaction to the title. Even although three meanings of the word 'divan' are glossed opposite the book's Contents list (viz. 1. a state council, 2. a couch or bed, 3. a collection of poems e.g. the Divan of Hafiz), this attempt at definition gets lost among the Acknowledgements and publication details. In any case, it seems by its positioning to refer to The New Divan as a whole and not to the title poem it is meant to contextualise. Most readers picking up the volume in a bookshop might be expecting a domestic interior and puzzle over the 100 stanzas that open the collection. They have rhetoric, colour, concern, drive - but seem hectic and confusing at first sight. Human emotion and biographical detail do not really surface clearly until the closing section of the poem, which in any case merges with the preceding 80 or so stanzas. But once the title had been accepted, it made sense to place this magnum opus first in the collection. Late in life Morgan came to think the poem should have been published by itself as a separate volume.

What emerged was a complex whole that lacked adequate signposting for readers. There are many good things in the collection as a whole. I particularly like the sound poem 'Shaker Shaken', 'Resurrections' and the affecting 'Unfinished Poems' sequence, written in memory of Veronica Forrest-Thomson, who had died in tragic circumstances in 1975. Edwin Morgan had known and corresponded with her since the late 1960s. But there was a warning signal too, perhaps, in the inclusion of so many series of poems: 'Three Trees', 'Ten Theatre Poems', 'Five Poems on Film Directors'. There is a law of diminishing returns that can operate in poetry as much as in business, no matter the technical skill. We might surmise that readers of 'The New Divan' may experience something like this on a first reading of its 100-piece set of glittering dream-mosaic pieces. 'Give us a break!' they might say on this hectic journey through landscapes and cultures.

Or learn from Bunting, I might say. Setting the long poem in more clearly defined sections (and I've suggested one arrangement that, at the very least, fits its compositional history and Morgan's fondness for sequences) might allow readers to pause on that journey, consult the map, look both backward and forward, and take time to think what it all could mean. That's probably a Northern way of looking at things, but it might have helped. Presumably Bunting did not lead his fuel convoy of trucks all that way from Persia to Tripoli with only the stars above for direction.

There is, of course, a case to be made for Edwin Morgan's claim for 'a mysterious, subterranean ... structure, an emotional structure, a structure perhaps relating to the life of the person who had written it' in his long poem. His wartime encounters with ancient cultures, temples and monuments that he had previously only read about, and recorded avidly in scrapbooks that he kept from the early 1930s onwards, must have given him an almost archaeological sense of living as a soldier in simultaneous layers of time. This also linked with his reading of the work of J.W. Dunne, much discussed in the late 1930s and early 1940s, notably An Experiment with Time (1928, with subsequent revised editions). Basically, Dunne argued that the future as well as the past does actually exist, and that the dreaming mind of a sleeper is freed in both directions, so that precognition is possible. This idea perhaps explains the sages sitting 'in full divan / in the anteroom of heaven' in stanza 81, and the observant presences of humankind's wartime activities elsewhere.

At the end of his University of Wales lecture, Morgan returns to Arabic, Persian and Turkish poetry, which partially provided the models for his own poem, although he does not explicitly mention it. He describes how the structure of their long poems may come not from the development of themes or from chronology of any kind, but only from metre and rhyme. He refers to FitzGerald's Rubaiyat of Omar Khayyam, with stanzas that 'in the original would be to us in the West bewilderingly unconnected but to those in the Middle East delightfully unconnected and in fact mysteriously and subtly connected!' - a return to the idea of the subterranean structure of his own long poem.

He contrasts the Western use of walking or sailing as a central unifying image for the long poem, with the Eastern comparison of stringing a set of beads with a variety of fine stones and pearls. 'The New Divan' ambitiously attempts both devices. Sea gales blow up at the start of the poem, and blow themselves out by its end. But the poet pictures himself in the final line using needle and thread 'to bind the leaves of my divan'. Perhaps a smaller hand-stitched volume might have helped to convey the cross-cultural form he had aimed for.