The treasures of the Huntington Library are internationally known to scholars and bibliophiles, but to a working-class kid from Southwest L. A., they seemed more mysterious abstractions than objects of intrinsic interest. A nerdy autodidact, I was precociously well-informed on painting, but my literary taste ran to Edgar Rice Burroughs, Ray Bradbury, and H. P. Lovecraft. I shuffled dutifully from display case to display case in the manner of a twelve-year-old boy–squinting, gawking, and gaping. I was indifferent to the Ellesmere Chaucer. The Gutenberg Bible impressed me only as a vague rarity. Shelley's notebooks and the manuscript of Walden got hardly a glance. But one exhibit held me spellbound. In a glass case in the main hall lay a fair copy of Edgar Allan Poe's "Annabel Lee" written in the author's small, meticulous hand.

I knew the poem well. Indeed, I knew most of it by heart. "Annabel Lee" was a favorite of my mother, who often recited it with obvious feeling. One purpose of poetry is to give us words to articulate our joys and sorrows without revealing them. Whenever my mother spoke Poe's words, I also heard between his lines the unspoken losses that had scarred her life.

For the moon never beams without bringing me dreams
Of the beautiful Annabel Lee;
And the stars never rise but I see the bright eyes
Of the beautiful Annabel Lee;
And so, all the night-tide, I lie down by the side
Of my darling, my darling, my life and my bride
In her sepulchre there by the sea–
In her tomb by the side of the sea.

But did I really know the poem? Staring at the manuscript, I found myself seeing it from a different perspective. I was overwhelmed by its sheer beauty–the elegant script in brown sepia with every letter perfectly formed, the way Poe had skillfully glued two pieces of paper together to make a single tall sheet so that the poem could be presented on one page. I knew also that Poe had intended for me to be overwhelmed. His wife dead, his career in tatters, sunk in debt, drinking heavily in the last year of his life, Poe had obviously labored to make this keepsake–probably given to some casual acquaintance–as beautiful an object as possible. For the first time in my life I had a sense of the complex, even contradictory human reality out of which art grows. I also saw how delicately the text of a poem balanced between two lives–the author and the reader. "Who touches this," Walt Whitman wrote of his own work, "touches a man." Looking at the manuscript through the glass, I wished I might someday make something so beautiful.

Since that afternoon I have been interested in literary manuscripts, especially autograph versions of poems. This interest has never been a consuming passion, and I have resisted the temptation to collect. (Any manuscript I could afford, I probably wouldn't want.) But in the intervening years I have rarely passed up the opportunity to attend a library exhibition, and in working on major critical projects, I examined whatever archival sources I could find, even if they did not bear directly on my topic. More recently I have spent considerable time studying many of the extraordinary manuscripts in the Berg Collection of the New York Public Library. In the process of this desultory research, I have come to wonder about the odd world of literary archives.

A great deal has been written about literary manuscripts, but nearly all of it has been bibliographic, textual, or commercial. Bibliographers assiduously catalogue and classify archival material. There are directories that list every known manuscript of Wordsworth, Coleridge, Shelley, and Yeats. Textual scholars customarily cross-reference manuscript variants in compiling definitive editions. Bookdealers and auctioneers publish lavish catalogues describing and often illustrating the literary manuscripts they hope to sell. The body of descriptive material about manuscripts is immense, but one finds virtually no critical or theoretical writing on the subject. There is little informed analysis or speculation, that is, on the problematic place of manuscripts in the broader literary culture; nor is much serious consideration given to their special status versus printed texts. If scholars have customarily discussed manuscripts in utilitarian terms, critics and theorists have ignored them altogether as a subject of independent inquiry.

In the absence of serious critical or theoretical interest, the terms for discussing manuscripts have mostly been set by the marketplace–the world of dealers, auction houses, and collectors. Although the private sector recognized the importance of manuscripts long before scholars and librarians made any systematic effort to retrieve and conserve them, the marketplace has done little to explore the real value of its material. It would not be altogether unfair to suggest that, like Oscar Wilde's cynic, the marketplace "knows the price of everything and the value of nothing." Of course, one might make the same complaint about the New York Stock Exchange. What the marketplace does is price things. Questions of value belong elsewhere.

The lack of speculative thinking and critical analysis, however, has not slowed the bustling world of literary manuscripts. Few areas of literary life are as full of purposeful and engaged activity. Thousands of librarians, curators, dealers, and scholars–as well as numerous private collectors–are currently building huge archives of primary materials. No other culture in history has even approached the level of activity now routine in America. Our institutions spend vast sums to acquire and house manuscript material. Large crowds view public exhibitions while scholars fill reading rooms to study it in detail. Bibliographic journals, exhibition catalogues, and auction catalogues pour from the presses. Newspapers review major shows and cover significant auctions. Without fully articulating why, the culture has agreed that, even in an era of shrinking resources, collecting and preserving literary manuscripts is a priority. Two hundred years ago no one would have bothered to save the contents of most contemporary archives. Why have literary manuscripts become so widely valued? How did the change come about? And what do current assumptions about manuscripts reveal about modern literary culture?

"All literary manuscripts," Philip Larkin once observed, "have two kinds of value: what might be called the magical value and the meaningful value." No one today doubts the scholarly or "meaningful" importance of literary manuscripts. Auction houses and dealers sell them at ever-higher prices. Libraries announce the acquisition of a famous writer's papers with the same blare of publicity that accompanies a medical breakthrough. The public rationale for collecting manuscripts is that they provide information not usually found elsewhere. This scholarly argument is cogent if–as we shall see later–also incomplete. Manuscripts can solve factual problems like dating a poem or establishing an accurate text. It is not uncommon for a scholar studying the manuscripts of a literary work to discover some obscure line in the standard printed text contains a mistranscription. Herman Melville's mysterious image in White Jacket of the "soiled fish of the sea"–much praised by F. O. Matthiesen–became, upon examination of Melville's original, the more literal "coiled fish of the sea." A famous critical puzzle turned out to be merely a misprint.

Manuscripts also illuminate the broader meanings of a literary work. Seeing what a poet cut out often helps clarify what was left in. Drafts can elucidate an author's intention in a particular piece. Diaries, letters, notebooks, and other documents add to the knowledge of a writer's life and milieu. Ezra Pound's compressed, cranky, and cryptic letters from Rapallo, Italy prepare the reader for the allusive and increasingly idiosyncratic Cantos he wrote concurrently. The "meaningful value" of manuscripts often transcends their purely verbal contents. Sometimes even the physical materials used suggest certain insights about an author's life. Emily Dickinson's carefully composed and fastidiously recopied packets of poems, usually handbound in thread, not only reveal the poet's class position in the New England gentry, which could afford the materials and leisure to prepare such fascicles; the packets also show the essentially private medium that constituted Dickinson's "publication." Not even her family knew the extent of her literary output. After Emily died in 1886, her sister Lavinia was astonished to see how many poems had been stashed away in the drawers of the bedroom bureau.

Best of all, manuscripts often contain new work. Few authors publish everything they write. The Nachlass (to use the German term for an author's literary remains) almost inevitably reveals unknown material. The final version of John Keats's sonnet, "Bright Star, Would I were Stedfast as Thou Art," was found in the author's hand, inscribed in his copy of Shakespeare's poems. In extreme cases like Dickinson or Gerard Manley Hopkins, virtually all of the work emerges posthumously from unpublished papers. Only seven of Dickinson's 1775 known poems appeared during her lifetime. Moreover, in the first editions of her work, her poems were rewritten by her editors into more conventional styles. Later editors returned to the manuscripts to restore her authentic versions. Hopkins published virtually nothing after joining the Society of Jesus–only three comic triolets and three Latin versions of English poems. His mature work did not appear until 1918, thirty years after his death. His literary executor, Poet Laureate Robert Bridges, carefully put together the first collection from four manuscripts he had assembled–like a classicist preparing an edition of Horace by collating and comparing the surviving manuscripts.

The manuscripts of a poem can be divided into three general categories–the working drafts, the final manuscript, and fair copies. Each type of manuscript affords certain insights into the author and the work. The working drafts (or worksheets) of a poem reveal the author's creative process. If all the worksheets survive, they track the poem's development from the author's initial impulse to the text's final form. Many authors, however, discard their drafts. Among the seven thousand items contained in the Huntington Library's Wallace Stevens archive, for example, one will find no worksheets. Working drafts also demonstrate the often significant differences in process between poets. Lord Byron seems to have written with remarkable swiftness and self-possession. His poems usually achieve some recognizable version of their ultimate shape in the first draft. The folio-size worksheets of Don Juan in the Berg Collection show whole passages being composed and decisively polished to their final form on the same sheet of paper. Elizabeth Bishop, however, could fret for years over a single poem. The surviving drafts of "The Moose" show that Bishop worked on the 168-line poem for more than twenty-five years before publishing it in The New Yorker in 1972. (She then revised it again before putting it in a book.)

The final manuscript is in one sense merely the author's last draft of a poem or collection of poems. Although it represents the end of the private creative process, it also marks the beginning of the work's public life. The final manuscript leaves the writer's desk and enters the public realm of editors, publishers, lawyers, and censors. For earlier writers this final manuscript was written by hand. For modern authors it has usually been a typescript. (Even in the age of personal computers, most poets continue to send along a "hard copy" in addition to a digital disk.) The final manuscript is uniquely valuable because it is the version of the text the author submits for publication. In artistic terms, it stands as the work achieved; in scholarly terms, it often best reveals the authorial intention. Literary works are often cut, revised, repunctuated, and even censored before publication for reasons that have nothing to do with the author's wishes. The punctuation of The Poetry of Robert Frost, the standard complete edition of Frost's poems, for instance, differs–according to Donald Hall–in over 1100 instances from the texts the poet himself put into print. By recasting Frost's poems into conventional punctuation, the book changes the rhythm and sometimes even the meaning of lines. If the best text of a poem is the one that represents the author's probable intentions, then the final manuscript provides a starting point for establishing that intention.

The comparison of the completed manuscript with the printed book can provide real revelations. E. E. Cummings's papers, for instance, show how the poet's language was sometimes censored for the publisher's concerns about obscenity. "No words changed," he wrote Boni and Liveright, about the manuscript of his prose narrative, The Enormous Room (1922). He then specificed objectionable words should be deleted rather than replaced by euphemisms: "In cutting use dash ––– e.g. '––– it,' said etc. (not 'chuck it,' said etc.)" His publisher nonetheless substantially revised his book without his approval–dropping character sketches, changing punctuation, translating French phrases into English, and removing supposed indecencies. The author's true version did not appear until 1978. The manuscript of Robinson Jeffers's 1948 collection, The Double Axe and Other Poems, now at the Humanities Research Center in Austin, reveals how Random House asked the author to delete ten poems and change several others for political reasons. Publisher Bennett Cerf and editor Saxe Commins were offended by the poet's criticism of Franklin Roosevelt. Although Jeffers reluctantly agreed to their changes, Random House insultingly placed an unprecedented publisher's disclaimer at the beginning of the volume announcing their disagreement with the author's political opinions. While boasting of Random House's commitment to "the writer's freedom," the note makes no mention of the omitted poems and many emendations, which remained unknown until a scholar examined the author's manuscript decades later. Jeffers's original versions were not published until 1976–fourteen years after the poet's death.

Authors, of course, also make changes after their work has been accepted for publication. "When a man knows he is to be hanged in a fortnight," Samuel Johnson observed, "it concentrates his mind wonderfully." The prospect of imminent publication clarifies an author's intentions. "Final" manuscripts frequently reveal last-minute changes by the author. The Berg's typescript of Jeffers's Solstice (1935) shows small revisions in the poet's hand–changes made for artistic rather than political reasons. Robert Lowell's typescript for Imitations (1961) in the Berg contains massive reworking for many of the author's free translations. The proofing stage, which is the traditional point where a text moves from the author's handwriting or typing to mechanical reproduction (although the advent of word-processing and electronic publishing have radically changed that process), also invites authorial changes. Seeing a text in the cold light of print changes an author's perspective. James Joyce repeatedly rewrote Ulysses in proof. Lowell's galleys of Imitations reveals his obsessive urge to revise. He once admitted that he could not see one of his own poems without feeling the need to change it.

Sometimes authors have the satisfaction of restoring their original versions. In the Berg's manuscript of Vladimir Nabokov's Poems and Problems (1970), the author's pleasure is tangible in the unhesitating revision of the final line of "On Translating Eugene Onegin." Striking out the bland ending foisted on him by the then prudish New Yorker , Nabokov reinstates his forceful original, a merciless but magisterial self-assessment of how his own English version of Pushkin, which occupied him for fourteen years, stands up to the Russian original: "Dove-droppings on your monument."

A fair copy is an autograph version of a poem written out or typed by the author after its completion or publication. (A "fair" copy contains no revisions or disfiguring corrections in contrast to a "foul" copy, which shows changes and emendations.) A fair copy, therefore, is often as much a keepsake as a literary artifact. Often intended as a gift or presentation piece, a fair copy usually displays the author's finest hand. Such manuscripts were immensely popular in the last century. Famous poets–and many poets truly ranked as celebrities then–were barraged with requests for autograph copies of their anthology pieces. The correspondence of 19th century American poets reveal many such requests and replies. The Berg Collection, for instance, has a charming note from Henry Wadsworth Longfellow to Walt Whitman that accompanied some fair copies the Good Gray Poet had requested for two friends. The Collection also possesses no less than five fair copies of Oliver Wendell Holmes's patriotic mega-hit, "Old Ironsides."

To the confusion of archivists and scholars, worksheets, final manuscripts, and fair copies often end up in different locations. If worksheets are not discarded, they usually remain in the author's possession and are eventually placed with his or her papers. Working drafts, however, may also be sent to literary friends for comment and thereby end up in another writer's possession. Completed manuscripts almost inevitably remain in the files of editors or publishers. Fair copies are scattered among their recipients. It is a rare circumstance when an author's entire body of drafts and manuscripts–not to mention letters, journals, and diaries–end up in one institution. Elizabeth Bishop's worksheets reside mainly at Vassar. Many of the final typescripts of her poems and collection remained in the files of The New Yorker and Farrar Strauss and Giroux, both of whose archives are now in the New York Public Library. Assorted other manuscripts are at Harvard, Washington University (St. Louis), and the University of Delaware as well as in private hands.

Although the diaspora complicates certain types of research (and keeps biographers on the move), the much-lamented situation is not without its advantages. An author's manuscripts are available to a broader constituency when they are dispersed; they are seen–in private research or public exhibition–by a more diverse audience. Centralization is not an absolute virtue in the arts. Having every Botticelli under one roof might simplify the work of art historians, but it would impoverish the experience and education of countless gallery-goers. The Stanford Library had modest special collections during my student years, but in 1975 James Healy donated a small but superb collection of modern Irish literature, including W. B. Yeats's own copies of all Cuala Press books with his unpublished annotations on each flyleaf. The directors of established Yeats archives may complain that such an important collection did not belong in California remote from most related material, but the Healy Collection was well cared for and much consulted. Its presence provided a tremendous boon to Stanford faculty and students.

"Fair is foul, and foul is fair," chant the prophetic witches in Macbeth, and indeed literary archivists generally consider fair copies less significant than working drafts. Worksheets offer scholarly insights; fair copies often seem mere souvenirs. And yet fair copies sometimes contain variant readings of a text, which may reflect an author's second thoughts about a poem. The elegant copy of "Ulalume–A Ballad" that Poe gave to a teenage girl he met a month before his death now stands as the definitive text of the poem. Fair copies also sometimes constitute the only surviving authorial manuscripts of a work, especially for pre-Romantic literature. What would textual scholars give for a copy–fair or foul–in Shakespeare's hand of Hamlet? The Westmoreland Manuscript, for example, represents the only authorial source for most of John Donne's English poems, and this exquisitely penned fair copy provides the sole text for three of the Holy Sonnets. Even when several sources exist, a copy of a poem carefully written out in the author's hand–or, as in Donne's case, an amanuensis–is a persuasive endorsement of the correctness of a particular reading. Worksheets may reveal the creative process, but a fair copy helps establish an authoritative text of the finished poem.

But what is a finished poem? Is Paul Valèry correct in declaring that "A poem is never finished, only abandoned"? The study of literary manuscripts highlights a central issue of modern poetics–what constitutes a completed work of art? As Romantic philosophers shifted their attention from the analysis of a conventionally finished artwork to contemplation of the creative process, they initiated a shift in taste that continues today. Works that might have been considered incomplete and unsatisfying drafts a century earlier received enthusiastic approbation as fully achieved poems. In 1798 Samuel Taylor Coleridge, the central English Romantic theorist, published "Kubla Khan" with the subtitle "A Fragment." In 1820 John Keats allowed the text of his abandoned poem, "Hyperion," which ends in mid-sentence, to appear in a collection. Another of Keats's most famous "poems" is an incomplete, eight-line passage found in the margin of a manuscript. The lines may represent the first draft of a poem or a possible speech for a projected play. First printed in 1898, these lines seemed to a modern sensibility perfectly self-sufficient.

If the Romantic period prized poetic fragments for revealing the artist's inner genius unfettered by the external constraints of classical form and propriety, the era also fostered an unprecedented reverence for literary manuscripts. Earlier ages had viewed a poet's holograph in purely functional terms. Once the text had been more decorously preserved in printed form, there was no special value seen in preserving the usually "foul" autograph copy. Paradise Lost was recognized on publication in 1667 as a masterpiece, but neither Milton nor his contemporaries took care to save the manuscript. Only one section of the huge original survived the curatorial indifference of immediate posterity. (This fragment still bearing the ink smudges of the printer now rests in the climate-controlled vaults of the Morgan Library.) The Romantic Zeitgeist, however, endowed a page written in a famous poet's hand with a talismanic power. Matthew Arnold may have erred in prophesying that poetry would replace religion in guiding mankind, but his prediction suggests how–at least among literati–poetic manuscripts became Victorian society's equivalent of holy relics. Indeed, the luxurious nineteenth century leather bindings and satin-lined boxes that house many of the Berg's famous manuscripts are the bibliophile's versions of reliquaria.

This shift in sensibility could probably only have occurred in the nineteenth century at the height of print culture. This was the first age of mass literacy, inexpensive printing, and rapid communication. Successive waves of technological and commercial innovation–daily journalism, universal post, telegraph service, pulp paper, broadscale advertising, dime novels–gradually changed the frame of reference for human knowledge and communication. For the first time in history it seemed that most information was announced, disseminated, preserved, and codified on the printed page. Authors occupied a critical position in the new print-driven economy. To use a contemporary metaphor, writers created the programming that ran the information and entertainment networks of print culture. Before film, radio, television, recordings, telephones, and computers, the writer exercised a near monopoly on information. Since fame was a function of print, successful authors could achieve vast celebrity; since ideas traveled most quickly and farthest in print, they could also exercise immense influence.

The poet commanded a unique and privileged position in nineteenth century print culture. Embodying the still powerful link between the new age and earlier oral traditions, the poet practiced the most venerable literary art–rooted in spoken language but easily transmitted in print–and the typographic bard still played an almost priestly role. No one was shocked when a novelist or playwright went to the bad; those professions had at best dubious reputations. But a famously dissolute poet like Byron or the young Swinburne exercised the evil fascination of a fallen angel. Fiction now dominated commercial publishing, but successful poets like Tennyson or Longfellow outsold novelists. Coventry Patmore's The Angel in the House (1858), for example, sold over a quarter of a million copies. Children still memorized and recited famous poems as a central part of the curriculum. Newspapers printed or reprinted poems as a regular feature. Famous poets like Byron, Longfellow, or the Brownings enjoyed international celebrity. When Edison wanted to record the most famous voices in the English-speaking world, he approached Queen Victoria, Prime Minister Gladstone, and Tennyson. Meanwhile Longfellow lived to see his birthday become a national holiday. Novelists commanded immense popularity, but only poets received such signal honors.

If technology had transformed the social identity of the poet, mass-production also changed the cultural status of his or her manuscripts. The uniformity of machine-printed books slowly imbued the handwritten page with a unique personal aura. As mechanical typography visually standardized written language, the reader was less likely to view an autograph copy of a poem purely as a piece of verbal communication; it now became a unique artifact that invited a different sort of attention. Impersonal communication was the function of print; the manuscript suggested a more individual and direct relation between reader and author–not through its text but its medium, which resembled a private communication. The omnipresence of mechanical print made the manuscript's medium its most important message.

The special value of a literary manuscript, therefore, does not come solely from the words it contains; the text can usually be found elsewhere in printed form. In purely functional terms–legibility, endurance, portability, and perhaps even accuracy–a book is usually more useful than an autograph. The manuscript's superior value originates, to use Walter Benjamin's term from "The Work of Art in an Age of Mechanical Reproduction" in its "aura"–"its presence in time and space, its unique existence at the place where it happens to be." A modern book is a standardized manufactured object; a copy in Chicago does not differ from those in Santa Rosa, Gainesville, or Johannesburg. A manuscript, however, not only exists as a unique entity whose total essence cannot be fully reproduced; it also survives as a historical artifact that can be traced back to its creator. A modern edition of Byron's Don Juan, for instance, provides an authentic verbal and intellectual link between the past and present, but it does not possess the unreproducible and unsummarizable aura of total authenticity that the original manuscript does.

Benjamin had worried that the mass reproduction of art works led to their desacralization by destroying the aura of the original. His theory may or may not accurately pertain to the visual arts like painting where there is an identifiable original, but it proves problematic in literature. If someone asks a group of art historians where the original of Piero della Francesca's "The Flagellation" is located, they can confidently agree that it is to be found in the Galleria at Urbino. But if a group of literary scholars were asked where the original of Hamlet is found, they would not only be unable to say where it is, they would probably be unable to agree on what the original is–a modern scholarly text, a First Folio, a great performance, a lost acting version used by Shakespeare's company, a collective idea of Hamlet among all theatergoers? The mode of existence for literary works is a complex and perhaps insoluble problem. Mechanical reproduction, however, paradoxically provided an answer at least for the common reader. As developments of commercial printing in the eighteenth century turned books from luxury goods to commonplace objects, mass production may have indeed desacralized the book, but it did not destroy the magical aura of the literary work, only transferred it. The proliferation of printed books gradually created a new type of original–the author's manuscript, endowed with an aura of authenticity and authority by the hand of its creator.

The manuscript of a literary work became more than words; it represented a direct and unmediated physical link between viewer and author–a holy relic or shamanistic fetish. When Larkin, the most level-headed of poets, exclaims, "This is the paper he wrote on, these are the words as he wrote them, emerging for the first time in this particular miraculous combination," he reveals the long-term results of the new attitude–the frisson of devotional ecstasy, a sense of the sacred. This quality of untransferable authenticity explains why even the fragment of a great work (like the Morgan's fragment of Paradise Lost) possesses magical value far in excess of its scholarly contents. Philosophers may deride this subjective notion of communication with some justification; as an analytical method, it is overtly flawed. Yet who can deny that one learns something in such transactions–often something essential? Why else do scholars and biographers assiduously study manuscripts? Any reflective person recognizes how much learning happens outside the realm of analytical deduction. Most of what one knows comes from sensory, intuitive, and imaginative faculties. Reason may later examine and organize this learning, but one first assimilated it holistically.

The scholarly alibi of libraries that they acquire literary manuscripts for intellectual reasons, therefore, is inadequate at best. Those needs could be better served by microfilm or photocopies at a negligible fraction of the expense. Textual information can usually be found both more easily and completely in a variorum edition. Like the intricately rational web of theology woven around the irrational mysteries of faith, the sober explanations of institutions for hoarding literary relics seem like elegant post-factum justifications for what is essentially a sense of sacred awe. An institution of learning seeks significant manuscripts because they possess qualities that scholarship cannot entirely reproduce–an authentic, holistic connection with the great writers of the past. It is not the intellectual content of the manuscript that is important but its material presence–ink spots, tobacco stains, pinworm holes, and foxing included.

The magical nature of the author's hand can be seen perhaps most clearly in signed books–a special case in which handwriting and mechanical typography meet. An autographed and an unsigned copy of the same title have no significant difference in content except for a bit of ink across one page, but the autographed volume will usually sell at many times the price of the other copy. Bearing the mark that it once rested in the author's hands, the autographed volume is transformed from a mere book into a minor relic. Before the nineteenth century owners signed their books mostly for two reasons–to prevent theft or to inscribe them as a gift. Alexander Pope's copy of Milton's Poems in the Berg, for example, bears its former owner's signature, not Milton's, and the presentation copy of Pope's The Rape of the Lock (1714) was inscribed by the poet only because he gave it to a patroness. The notion of a stranger soliciting Pope to sign a copy of The Rape of the Lock would have seemed peculiar to the Augustan poet. Midway in the following century, however, it was so common for strangers to ask writers to autograph books that it proved a great nuisance to established poets. The autograph had become print culture's means of individuating mass-produced books by linking them with the author's person. Literary autograph collecting reached its peak at the turn-of-the-century when Rudyard Kipling discovered that shopkeepers would rather keep a small check with his signature than cash it–true fame.

The autographed book exists as a mid-point between the anonymous nature of the manufactured book and the unique personal essence of the manuscript. If a book contains sufficient authorial additions, however, it gains the unique aura of the manuscript. In Coleridge's copy of Lyrical Ballads (1798) now in the Berg, the poet has revised the printed text of his poems. "The Rime of the Ancyent Marinere" is covered with many significant additions, deletions, and revisions. The book has become a manuscript. In the Berg's copy of Tennyson's The Princess (1847) the poet has added several new short lyrics to his long didactic poem. These 1850 additions, which include "The Splendor Falls" and "Sweet and Low," now rank among his most popular poems. Sometimes one man's book became another man's manuscript. Coventry Patmore covered the margins of his copy of The Princess with a running commentary on Tennyson's poem–half praise, half criticism. Patmore's marginalia provides the modern reader with a vivid sense of how the best Victorian poets read one another's work. Consulting with Siegfried Sassoon, poet Edmund Blunden covered his copy of Robert Graves's memoir, Goodbye to All That (1929), with their copious disagreements concerning the author's account of events; their commentary effectively recomposes the printed monologue into an argumentative dialogue in manuscript. Dissatisfied with Winifred Roy's translation of his novel Despair, Nabokov reworked her version phrase by phrase writing his more forceful and accurate English between the lines of the printed text. Now in the Berg, this book-cum-manuscript provides a vivid demonstration of Nabokov's practice both as novelist and translator. Perhaps the most extravagant case of "recomposition" on record appears in Andrew Motion's biography of Philip Larkin. The librarian poet and his longtime girlfriend, Monica Sinclair, spent years "systematically defacing" and comically rewriting a copy of Iris Murdoch's 1956 novel, The Flight from the Enchanter, until they had transformed word by word nearly 300 pages into a pornographic farce. Fate has a sense of humor: this singular literary artifact now rests in the special collections of the Brynmor Jones, Larkin's own library.

That the magic value of manuscripts surpasses their meaningful value can be further attested by the passion both private collectors and public institutions exercise in obtaining the personal effects of famous writers. In addition to Mrs. Browning's slippers, Dickens's desk, and Thackeray's pen, the Berg Collection possesses such modern relics as W. H. Auden's suitcase, Randall Jarrell's library card, E. E. Cummings's death mask, and Howard Moss's pencil sharpener. The auction prices for literary artifacts may be lower than for Marilyn Monroe or Elvis Presley memorabilia, but–the prestige of possession aside–the underlying motivations to acquire them are equally irrational and superstitious. The new owners hope mysteriously to gain some part of the original celebrity's power or allure.

Medieval towns once competed for the bodies of saints to protect their citizens from plagues and natural disasters. (One still sees the waxy cadavers of the blessed in the churches of Southern Europe–displayed behind glass like rare manuscripts in a library exhibition.) If the physical remains of saints were unavailable, then their personal effects would suffice. Cloaks, missals, rosaries, belts, and shoe leather were all to be cherished for their talismanic power. While the Protestant North ridiculed this Southern Catholic custom, they did not escape its primal attraction; they merely translated it into secular terms. The remains of great poets were buried in Westminster Abbey; their personal effects and manuscripts were collected by libraries. Although there seems to be no commercial traffic in the bones of poets, Tennyson once claimed that dealers would have sold his toenail parings had they gotten hold of them. None of the esteemed Laureate's clippings have been preserved for posterity, but locks of his hair still exist. There are, in fact, many private and institutional collectors of literary hair. The Berg, for example, owns two locks of Whitman's famous gray mane.

Manuscripts also represent the imagination's passport; they allow the viewer to travel from the public and impersonal world of mechanical typography into the private, human world of the author–from literature as an institution to literature as friendship. A book is a public object collectively produced by many hands and designed for many readers; in contrast, even a typewritten draft seems intimate and individual. The manuscript–handwritten or typed–invites the viewer to step from the faceless crowd of readers and become an individual. It is tempting to portray manuscripts as objects of aesthetic contemplation, and the elegant fair copies of poets like Poe, Longfellow, or Keats reward such attention, but the intrinsic beauty of a charm is no measure of its magic. Reliquaria must be beautiful, not the relics. For this reason foul copies often provide as much or more pleasure than fair copies; in the corrections, deletions, fragments, dead ends, and doodling, one discovers the author's humanity. Kipling and Cummings sometimes playfully illustrated their manuscripts and letters. Anne Sexton's triumphant "Sold to New Yorker" scrawled at the top of a typescript poem lets the viewer share the author's pride and pleasure. If the magnificence of the finished work excites our aesthetic imagination, the imperfections of the working drafts allow our participation.

One suspects that Shakespeare's status has only been enhanced by the absence of manuscripts; he remains Olympian and behind his masterpieces. Among the well-documented moderns, Stevens has deepened the mystery surrounding his inner life by leaving no working drafts of mature poems. Perhaps the special appeal of final manuscripts and fair copies is that they provide the most direct comparisons between the private and public identities of the poet; they present more or less the same words one already knows from the printed page. If the words are identical between a manuscript and a printed text, then what one notices are the non-verbal aspects like the handwriting, paper, number of deletions or additions, even the stains–all the clues of the author's physical presence. Final manuscripts reveal the juncture between the secret realm of poetic inspiration and the external existence of the printed text.

The study of literary manuscripts also suggests the complex but vital connections between poetry and technology. In the visual arts the impact of technical innovation like lost-wax casting, oil-based paints, color lithography, or daguerreotype constitute an essential part of a medium's history. In literary studies, however, poetry is mostly seen as pure language generated by the writer's imagination in relation to the tradition. Scholars may study the economics and technology of publishing to understand the social identity of the poet, but the most intimate relation between technology and poetic composition remains largely unexplored, except by general theorists like Marshall McLuhan, Walter Ong, or Eric Havelock. Does anyone know, for example, the first significant poet to compose on a typewriter?

Only six years separate the composition of W. B. Yeats's 1917 volume, The Wild Swans at Coole, and William Carlos Williams's 1923 collection Spring and All, which contains his famous "The Red Wheelbarrow," but the impact of technological change is already apparent. The Berg's copy of Yeats's manuscript exists only in the author's swift, strong hand. In 1917 Yeats still did not use a typewriter. His method of composition remained oral, aural, and manual–a scribal method not appreciably different from the procedures of Virgil, Dante, or Donne. Yeats never wrote a poem that did not rhyme. The forceful meter, symmetrical stanzas, and overt musicality of his verse not only demonstrate his mastery of traditional prosody but also his commitment to shaping poems in sonic terms. Perhaps Yeats's only technological advantage over Virgil is inexpensive paper, which allows him to take each poem through numerous drafts.

Six years after Yeats's book, however, what constitutes a poem in technical terms has broadened enormously. Williams's "The Red Wheelbarrow" exists primarily not as sounds moving through time but as words visually fixed in space. Like "The Wild Swans at Coole," Williams's typographic poem displays strict order and symmetry. There is no rhyme or meter, but each stanza contains an identical pattern of four words arranged in combinations of three and one.

Critics often celebrate Williams as the poet of natural American speech, but the visual technique of "The Red Wheelbarrow" deliberately subverts normal speech rhythms. By breaking the poem's single short sentence into eight shaped lines surrounded by the white space of a blank page, Williams forces the reader to slow down normal speech into the static visual rhythm of the poem. It is no coincidence that Williams kept a typewriter in his office to write between patients when a paper and pen would have been more convenient. Cummings also systematically explored visual features of typography that have no exact equivalents in speech–lower and upper case letters, visual symbols (like &, $, and %), arabic and roman numbers, punctuation, spacing and abbreviation. For good reason the Berg preserves his Royal typewriter along with his papers. The keyboard was integral to his creative process.

A literary manuscript allows the viewer to probe–however subjectively or inadequately–the mystery of artistic genius. If it is a great manuscript, it invites one to observe an author at a moment when he or she performed at the limits of human possibility. The printed text may allow one to view that performance from the stadium stand, but the manuscript puts one at the author's side. Here, as so often in the arts, there is a powerful element of voyeurism.

The urge to see the author face to face is not merely fandom; it is a deep-rooted, primitive human desire. The physical separation of the poet from the audience is a relatively recent phenomena. For most of human history the audience heard the poet's physical voice; a direct physical relationship was unavoidable in pre-literate societies. Even after the introduction of the phonetic alphabet, which allowed writing to preserve the text of poems, the links with oral culture remained. In Roman times a poet "published" his work by reading it aloud to an invited audience. Later troubadours and Minnesingers sang or recited their poems. Handmade books were rare and expensive; and the connection between poetry and oral presentation was an unbroken tradition from time immemorial Only the introduction of printed books fully separated the living poet from his or her audience. The age of typography amounts to a small portion of total human existence–three or four centuries out of perhaps a million years.

The popularity of poetry readings is a reminder of the strong aural and tribal roots of poetry. Readings bring an audience into a direct physical relationship with the author–and momentarily form a tribe of like-minded listeners versus isolated readers. This longing for personal–and non-verbal–knowledge of the poet is also found in the now common practice of dust-jacket photos. Handwritten manuscripts satisfy the same desire in readers. The author's hand provides both a direct human link and a suggestive invitation to the viewer's imagination. Graphology may be only a pseudo-science, but it rests on a genuine insight: there is some essential connection between handwriting and character. Just because the connections cannot be standardized does not invalidate them. Different aspects of the writer's personality emerge in different hands, and the penmanship of the same person may shift noticeably according to mood. Biographer Edmund Morris has described handwriting's special ability to communicate the inner state of its author:

Script's primary power is to convey the cursive flow of human thought, from brain to hand to pen to ink to eye–every waver, every loop, every character trembling with expression. Type has no comparable warmth.

Examining the manuscripts of famous poets, one is often struck by how clearly some aspect of their character emerges in the handwriting. Poe's conspicuously beautiful hand reveals the same sensibility that articulated his dreamy aestheticism. Dickinson's eccentric script enlarges the letters so that a short poem sprawls across an entire page–one or two words stretching from margin to margin. Can there be any doubt the author is insisting on the outsize importance of her words. "Don't mistake my short poem for a small poem," the handwriting shouts. Is anyone surprised that Longfellow has neater handwriting than Whitman? Or that Keats wrote with more deliberate care than Byron whose quick, vertical hand dashes dramatically across the page?

The value our culture has placed on literary manuscripts reflects an admirable and ineradicable human impulse–the desire for a direct and authentic relation between art and its audience. Literature may be an institution, an imaginary library too vast and labyrinthine for any single reader to explore entirely, but the experience of studying great manuscripts reminds the viewer that each individual work is also a conversation, an imaginative and emotional transaction from one person to another, a bridge across time and place. Reading is never more intimate than with script. The hand of the poet reaches out to greet the reader.