[In the following essay, Taylor urges that the poem "Shall I Die? " must be accepted as Shakespeare's until evidence can be brought forth against this claim. The author cites verbal parallels between the poem and Shakespearean canon as supportive of the claim for Shakespeare's authorship.]

On the evening of Nov. 14, while I was routinely checking references in the Bodleian Library at Oxford, I came across an item I did not recognize, a poem attributed to William Shakespeare, with the first line "Shall I die? Shall I fly?" I asked for the manuscript to be fetched and late the next morning I went back to check it. I found the literary equivalent of Sleeping Beauty, a nameless poem awakening from the ancient sheets in which it had lain undisturbed for centuries, a poem without a critical history.

This sleeping beauty had a rude awakening. By Nov. 22 (with the help and advice of my senior colleague on the Oxford project to publish a "Complete Works" of Shakespeare, Stanley Wells), I had subjected the poem to every accepted test of authenticity; the results were all positive and we could think of nothing else to check. Within hours of my trying to find a publisher for a scholarly article on the poem, the story had been picked up by journalists, and the poem was published on Nov. 24 by The New York Times. Public reaction to the discovery has been generous and enthusiastic; I have been overwhelmed by a tidal wave of curiosity. Academic reaction to the poem has been mixed. Its critical history over the last three weeks has reiterated in miniature the history of the critical reception of Shakespeare's early work over the last three centuries. Many of the early plays were long dismissed as spurious, because they did not seem worthy of his genius. His poems—the most popular of his printed works in his own period—were almost universally disdained for a century and a half after the beginning of the English Civil War in 1640. Even now, by comparison with his plays Shakespeare's poems (a few famous sonnets excepted) remain little read, little taught, little appreciated.

The neglect of his nondramatic verse partly originates in an accident of literary history. Shakespeare was canonized—the canon of his works was first defined—by the publication of the First Folio of 1623. That volume, compiled by two senior members of the theatrical company to which Shakespeare devoted most of his professional career, contains only his "Comedies, Histories, and Tragedies," as its title page says. No comparably authoritative collection of his poetry was ever printed. Three books of nondramatic verse were separately published during Shakespeare's lifetime—two long narrative poems ("Venus and Adonis" and "The Rape of Lucrece") and a long sonnet sequence with a narrative coda ("A Lover's Complaint"). But most poets also write shorter poems. Shakespeare did too, at least once: "The Phoenix and the Turtle" was published in an anthology for which it may have been especially written.

If Shakespeare had written other occasional poems of this type how where would we find them? Usually such works only surface in some kind of collection. In 1599 a gathering of 20 poems, entitled "The Passionate Pilgrim," was published, attributed to "W. Shakespeare." It contains two of his sonnets and three extracts from "Love's Labor's Lost," four poems known to be by other authors (but nevertheless included in all editions of Shakespeare's works) and 11 poems of unknown authorship. Although some of these 11 may indeed be Shakespeare's, the collection as a whole was clearly a disreputable and unauthorized attempt to capitalize on the enormous early popularity of Shakespeare's two volumes of narrative verse, published in 1593 and 1594.

Likewise, in 1640 a notorious edition of Shakespeare's "Poems" appeared, purporting to do for his nondramatic verse what the 1623 Folio had done for his plays; in fact, the volume simply reprinted pieces already published, rearranged the sonnets (and gave them bogus individual titles); it also added a few poems demonstrably not by Shakespeare. Both "The Passionate Pilgrim" and the 1640 "Poems" tried to exploit the market value of Shakespeare's name; their attributions are sometimes clearly wrong and always dubiously motivated. But both volumes testify to a common unspoken assumption: the belief of pre-Restoration book buyers (exploited by booksellers) that Shakespeare did write occasional nondramatic poems, some of which had not yet seen print.

If their conviction was justified, then the poems they expected could only survive in one place, or rather in many versions of a single place: in manuscript, probably in one of the thousands of surviving manuscript miscellanies from the first half of the 17th century. Patient scholarly excavation of this previously uncharted territory has revolutionized the postwar editing of Donne, Sidney, and other, lesser poets. But Shakespeare's editors, obsessed with increasingly sophisticated technologies for analyzing the transmission and manufacture of printed documents, have simply assumed that the manuscripts contain nothing of interest. We have not bothered to look, because we are sure there is nothing to find; after all, if there were anything to find, our predecessors would already have found it. Unfortunately, our predecessors were hardly in a position to study the manuscripts systematically and, for a variety of historical reasons, not disposed to study them sympathetically.

Rawlinson Poetic Manuscript 160 is one of a horde of early manuscripts amassed by Richard Rawlinson (1689-1755) and donated to the Bodleian Library in 1756. In this century the manuscript has been called from the stacks by several eminent scholars, including E. K. Chambers and the editors of Ben Jonson's works, C. H. Herford and Percy and Evelyn Simpson. Margaret Crum catalogued its entire contents for her "First-Line Index of Manuscript Poetry in the Bodleian Library" (1969). She recorded that the miscellany contained an untitled poem attributed to William Shakespeare, but she drew no extraordinary attention to this fact. Why should she? The attribution had been noted long before, equally nonchalantly, in Falconer Madan's great "Summary Catalogue of Western Manuscripts in the Bodleian Library" (1895). The poem was even included in a Bodleian exhibition of "Shakespeariana" in 1927 (item 120 in the exhibition catalogue). But although the poem has been known to a few Bodleian librarians for almost a century, it has never been reprinted or discussed.

The poem, which I here reproduce in a modernized and edited text, runs from the middle of folio 108 (recto) to the middle of folio 109 (recto) in the Rawlinson manuscript. It is written in the same elegant secretary hand used throughout the volume. The miscellany was apparently put together in the 1630's; it contains poems by Raleigh, Donne, Jonson, Beaumont, Fletcher, Herrick, Carew and others. Fifty other poems in the manuscript are attributed to specific authors; none of those other attributions are demonstrably wrong, most are demonstrably right, and only two ambiguous initials are even dubious—"J. D." (John Donne?) and "G. H." (George Herbert?). The miscellany's attributions deserve our respect.

One other poem in this manuscript is also credited to "Wm. Shakespeare"—a six-line "Epitaph" on Elias James, first noted by Edmond Malone in the 18th century. So short a poem in so convention-bound a genre could never be proved to be by any one author on the basis of style alone. But it is never attributed to anyone else. The British scholar Leslie Hotson has just shown that Shakespeare may have known a Londoner named Elias James, and John Pitcher, who is editing Shakespeare's poems for the Oxford English Texts edition, had already persuaded us to accept it into the canon, even before we found this new poem in the same manuscript. Nevertheless, for my purposes in assessing the manuscript's reliability I classified the attribution of this epitaph as "not demonstrably wrong."

The compiler of a private miscellany has no motive for lying about the authorship of a poem; he may make mistakes, but they will be honest. In fact, we do not know of a single verified instance of a poem attributed to Shakespeare in an early manuscript which can be proved to have been written by someone else. Nor is there any reason for falsely attributing this untitled lyric to Shakespeare. A sonnet, or a poem on a prominent Stratford citizen, or one about Venus and Adonis, or by one William Strode (who shares with Shakespeare the initials "W. S."), might be credited to Shakespeare by guesswork; but nothing about this lyric invites such a mistaken association.

The scribe of the Rawlinson manuscript testifies to Shakespeare's authorship of the poem, and although we do not know exactly who this witness was we know approximately when he wrote, that he had no reason to lie, and that his other attributions are reliable. Unless other evidence emerges that decisively contradicts this attribution, such external evidence itself establishes a prima facie case for Shakespeare's authorship. This evidence could only be overturned by the discovery of some other reliable document that more convincingly attributed the poem to a different author. Like Margaret Crum, I am unaware of the poem's existence in any printed collection from this period; nor does it survive in any other copy in the major manuscript collections at the British Library, the Bodleian, Folger, Huntington, Rosenbach, Yale or Harvard libraries. Although the poem may surface in some other collection, pending any such discovery the Rawlinson manuscript is our only evidence of its text and authorship. (And of course even if another manuscript surfaced it might confirm the Shakespeare attribution, or not attribute the poem at all.)

An anonymous Caroline scribe says Shakespeare wrote this poem; less explicitly, but no less forcefully, the poem itself says so. Its vocabulary, imagery, style—everything scholarly jargon lumps together as internal evidence—are at least compatible with Shakespeare's authorship, and, if one gives them the most weight they will bear, they suggest that it could hardly have been written by any other known poet. This is not the place for a full commentary on the poem, but it is at least worth drawing attention to the more interesting verbal parallels (see list below).

These parallels vary widely in quality and importance. No one will suppose Shakespeare is the only author to have called lips "red" (line 61), but it is worth knowing that he did use so conventional an adjective so often—and that three of the other four examples occur, as does this one, in the midst of a catalogue of physical attractions. (Spenser, by contrast, although he refers to lips 53 times, never calls them "red.") Even the clichés of the poem are clichés Shakespeare couldn't resist. The cumulative force of the verbal similarities between the poem and Shakespeare's acknowledged works could only be weakened by the identification of another poet whose works provided more and better parallels; skepticism may busy itself surveying the works of all possible candidates in the half-century before 1630. But the example of Spenser—the only Elizabethan poet with a canon comparable in size to Shakespeare's—suggests that any such search is likely to be fruitless. The Shakespeare canon supplies 107 quoted parallels for 52 phrases in the poem; the Spenser canon yields only 47 parallels for 18 phrases. (Shakespeare's poems alone—which add up to a mere fraction of Spenser's—provide parallels for 14 phrases.) None of the Spenser parallels are as striking as the best Shakespearean ones.

Shakespeare's style, in all his authenticated works, continually stitches old to new: old phrases and images recycled, but—always—new and unique language, too. Shakespeare commands the largest vocabulary of any writer of the period; his genuine works always contain an unusually high proportion of words he uses only once. Consequently, any work with a credible claim to Shakespeare's authorship must also contain a reasonable number of such "unique" words. Paradoxically, if a poem of any length does not contain words that Shakespeare never uses elsewhere, then that poem cannot be by Shakespeare.

This new poem contains seven unique words: "explain," "inflection," "admiring" (used as a noun), "desiring" (noun), "speck," "scanty" and "contenting" (noun). Moreover, one of these words probably, and another certainly, antedate the first occurrences recorded in the Oxford English Dictionary: "admiring" (first recorded elsewhere in 1603), and "scanty" (first recorded in 1660). Shakespeare coins more new words than any poet of the period, with the possible exception of Nashe; these neologisms therefore reinforce the claim for the poem's authenticity. And Shakespeare elsewhere makes much use of words based on the same root from which both these coinages are formed—including the first recorded use of "admiringly" ("All's Well That Ends Well," 5.3.44) and of a particular sense of "scant" ("Troilus and Cressida," 4.4.49).

Most readers would, I trust, immediately agree that if Shakespeare wrote the poem at all, he must have written it fairly early in his career. Those to whom I have shown the poem have reacted instinctively in this way; I did myself. The poem—in its subject matter, tone, obsessive rhyming, heavy verbal symmetries and the conventionality of much of its imagery, repetition of certain words and relative lightness of tone—bears every stylistic hallmark of belonging to the 16th century, and more particularly to the period between the publication of Spenser's "Shepherd's Calendar" (1579) and the growing influence of Donne. Consequently, if Shakespeare did write it, we should expect to find not only evidence of Shakespeare's style in vacuo, but more specific evidence that associates the poem strongly with a particular period of his work.

Although there remain minor disagreements about the relative dating of the canon, in general scholars have little difficulty distinguishing work from Shakespeare's early, middle and late periods; a variety of internal stylistic evidence all tends to suggest that each play and poem belongs in a certain range on a sliding chronological scale. We should expect a similar clustering of evidence here, and we should expect that clustering to suggest an early date. Few people would be willing to credit the attribution if it forced us to place the poem alongside "The Tempest" or "Antony and Cleopatra" or even "Twelfth Night."

Noticeably, most of the parallels do come from Shakespeare's early period. The four works with the most parallels are "Romeo and Juliet" (10), "Venus and Adonis" (8), "The Taming of the Shrew" and the Sonnets (5 each)—all probably earlier than 1596. If we ignore the Sonnets (which probably straddle the borderline), 58 of the 85 quoted parallels (68 percent) are from works no later than "Henry IV, Part 1 " and "The Merchant of Venice," which mark a kind of watershed in the Shakespeare chronology (1596-7). Shakespeare's authentic works manifest a similar clustering of verbal parallels in specific chronological periods.

Such evidence is confirmed by another test that has proved successful in dating the sequence of Shakespeare's work. The late Eliot Slater demonstrated a statistically significant correlation between the distribution of rare words (those which occur less than 11 times in the canon) and the demonstrable or probable order of composition, as determined by external evidence. There are 15 such rare words in this poem: "annoy" (as a noun), "bare" (noun), "besot," "exempt" (adjective), "exile" (verb), "impair," "inferior" (adjective), "mishap," "repenting," "scape" (noun), "star-like," "suspicious," "tresses," "twine," "wantonly." These words occur 73 times in the Shakespeare canon (excluding the parts of "The Two Noble Kinsmen" attributed to John Fletcher). Of these links with the rare vocabulary of Shakespeare's works, 52 (74 percent) occur in plays earlier than the 1596 watershed; if we added the Sonnets (many of which must have been written by 1596), the figure would be 57. Only 16 come from all the later works put together. The strongest links are with "Henry VI, Part 1" (10), "Venus and Adonis," the Sonnets and "Henry VI, Part 3" (5 each).

The poem contains 74 rhyming pairs; of these, seven involve unique words, for which we cannot expect to find parallels elsewhere in Shakespeare's rhymes. Of the remaining 67, almost all are rhymed elsewhere in the canon, and 25—over a third—yoke rhyme pairs Shakespeare uses elsewhere: die/fly, breeding/proceeding, beauty/duty, ever/never, out/doubt, love/prove, annoy/joy, blot/not, last/past, pleasure/treasure, find/wind, amazed/gazed, fair/hair, high/lie, eyes/prize, cheeks/seeks, meet/sweet, made/trade, endure/sure, asunder/wonder, blot/spot, awake/take, find/mind, delay/say. These rhyme pairs occur 115 times in the canon. It is perhaps not surprising that the great bulk of these parallels come from early work, because Shakespeare used rhyme more often in that period.

But the rhymes link the poem not only with the early period generally, but with specific early works. The strongest links are with the Sonnets (24), "Love's Labor's Lost" (12) and "Venus and Adonis" and "The Rape of Lucrece" (9 each). The last three all date from 1593-94; many of the Sonnets were probably written during the same period. "Venus and Adonis," a very short work, and the Sonnets in particular provide large numbers of parallels in all three categories of chronological evidence available: verbal parallels, rare words and rhymes. The stylistic features thus not only conspire to agree that the poem, if by Shakespeare, is by the young Shakespeare; they conspire to place the poem's likeliest date of composition in the very years when Shakespeare undoubtedly indulged in a brief burst of nondramatic verse.

I have so far dealt only with quantifiable, quotable aspects of Shakespeare's style; by those criteria the poem could certainly have been written by Shakespeare in the early to mid-1590's. But is the poem good enough to be Shakespeare's? In recent weeks a few distinguished critics—at least one on the basis of nothing more than hearing the poem read by a journalist over the telephone—have denied the poem's authenticity because it is, they say, "feeble." Anyone is entitled to an opinion of the poem's merit. But judgments of quality cannot be made the primary, or even the secondary, basis of attributions of authorship; if they were much of Shakespeare's work would be relegated to the foot of the page (as indeed it was, by Alexander Pope in his 1723 edition of Shakespeare). Nor can Shakespeare's early work be judged by the standards he himself created later in his career. And whoever wrote the poem, it must have been intended as a technical exercise, a kind of verbal obstacle course in which one of every three syllables is a rhyme. The originality and difficulty of the rhyme scheme produces a poem which is artificial, and hence as admirable to Elizabethan critical taste as its seems perverse to ours. Undoubtedly the effort to rhyme distorts the syntax and weakens the sense in places. But Shakespeare's rhymed poetry is often awkward and much of the rhyme in the plays was once dismissed as spurious because it is awkward.

By such means one can excuse the poem's weaknesses—it is the work of a young poet, who was never at his best (as Campion and Jonson were) in rhyme, engaged in a technical exercise. But one must also say, in the poem's defense, that reports of its feebleness have been greatly exaggerated. From its very first line, the poem sets up a conventional Petrarchan metaphor, and then subverts it; to "fly" in a battle would be immoral, to "fly" from an obsessive sexual entanglement would be wise. And whereas no soldier wishes to "die" (perish), a lover does wish for death (orgasm), and yet in seeking that orgasm he must, in the traditional hyperbole, "die" over and over in the inhospitable climate of his beloved's neglect or disdain. The narrator seesaws between obsessive engagement and ambiguous detachment. His love "breeds" sorrow, not children (line 3); he must "vent" his lust, verbally or physically (line 11); does "admiring" breed "desiring," or vice versa? (line 59). Are the besotting "plots" which part asunder breasts or thighs? (lines 72-3). His dream "did seem"—and even then he interrupts himself parenthetically to lament the fleeting insubstantiality of his dream (which may of course have been a reality that he recognizes only now as an emotional illusion). He ends with a mere detumescent iteration of what "some say"—implying, of course, that others say something else. But what those some say is itself ambiguous. What causes repenting: to delay the pursuit of his human prize, the commencement of his wooing, or to linger in his predicament? Given the sexual pun on "case," which Shakespeare uses so often elsewhere, the last lines may mean that he will regret it if he lingers too long in that most intimate place.

But I do not want to pre-empt interpretation of the poem or dwell on its crisp irony at the expense of its luxuriating sweetness. I have in any case already said enough, I hope, to persuade readers that the poem must be regarded as Shakespeare's until proved otherwise. An early document attributes it to him; we have no particular reason to doubt that document; the poem's style is compatible with the document's attribution. Whoever demands more proof is demanding that a poem pass the threshold of his own critical esteem before it can be admitted into Shakespeare's house. But documents, like defendants, must be presumed innocent until proved guilty; unless this document's attribution can be disproved, this poem must be included in any edition of Shakespeare's works that claims to be "complete."