The Review of English Studies, May 1994 v45 n178 p185(19)

'Hebdomada Mortium': the structure of Donne's last sermon. Bevan, Jonquil.

Abstract: The structure of John Donne's 'Last Sermon' is analyzed to show that its page layout is crucial to its correct interpretation. It is already known that spellings, punctuation and other typographical features may alter, and even reverse, the original sense of written work. Since it was meant to be read, the 'Last Sermon' should be published using the original layout as created by John Donne.

Full Text: COPYRIGHT 1994 Oxford University Press

BEFORE the days of the New Bibliography, editors of Renaissance literature who chose to present their texts in old spelling sedulously retained unexpanded contractions and typographical features such as long-s and consonantal 'i' and 'u', yet felt free to 'modernize' punctuation extensively and silently. Once it has been pointed out that punctuation can alter sense (indeed, reverse it), it becomes clear that it is not enough to retain old spelling alone: original punctuation must be respected too. All this is now common ground, at least to those who choose to think about such things (many literary critics still do not). What is perhaps less generally appreciated is that the original 'layout' of a text may also affect its sense, may indeed be part of it.(1) Where this is so and, equally importantly, where it is suspected that this might be so, the physical arrangement of the text on its page must be respected as much as any other original feature. I wish to argue that this is the case with Donne's last sermon.

But first, perhaps, the general point needs to be thrust home. An obvious author to illustrate the argument is George Herbert. No sensible editor is going to alter the indentation of 'The Altar' or 'Easter wings'. Nor is any editor likely to change the arrangement of the title of the 'Ana-{ MARY }gram' or the use of { ARMY } capitals in the poem 'JESU' or of italics in 'Coloss. 3.3'. But indentation and the pattern on the single page is one thing. What of the relationship from one page to another? What of the use of running-titles? What (for example) of the take-up between verso and recto in pp. 16-19 of the first edition of The Temple (1633)?

If we investigate this particular case, we will see that page 16 concludes the long poem headed 'The Church-porch. Perirrh-anterium' which has the running-title 'The Church-porch' throughout. This closing page, running-title and all, is a verso. The recto which faces it has printed on it the poem 'Superliminare' which has double rows of printers' flowers above and below it, and a rule separating its two stanzas. But this liminary page has no running-title. The two short stanzas printed on it are one of invitation ('approach') and one of warning ('Avoid'). If the leaf which bears this recto page is turned, the reader will find 'The Altar' printed on its verso, and, confronting 'The Altar', on the following recto, the beginning of another long poem, 'The Sacrifice'. These two poems are linked by the last couplet of 'The Altar':

O let thy blessed SACRIFICE be mine, And sanctifie this ALTAR to be thine.(2)

They are further linked by a new and shared running-title, 'The Church', which occurs on verso and recto, and continues for all subsequent poems up to p. 183.

What all this means is that the volume is physically arranged so that the poem 'Superliminare' acts as the door between the Church porch and the Church itself. The reader must make an act of will and decide whether or not he will push open the door (or turn the leaf) and so enter the Church and contemplate its altar. But this experience of reading 'Superliminare' will not be reproduced for the user of Martz's Oxford Authors edition of Herbert and Vaughan (1986), since in that edition 'Superliminare' shares the running-title 'THE CHURCH-PORCH' with 'THE CHURCH-PORCH Perirrhanterium', and there is therefore no separation of the door from the Church porch; but at least the poem is on a recto, with 'The Altar' following on the verso under the new head-title 'THE CHURCH' (this becomes the running-title on subsequent rectos, the running-title for the versos being 'GEORGE HERBERT'). In Patrides' otherwise excellent Everyman edition (1974) 'Superliminare' is on a verso with 'The Altar' facing it: no page turning is involved as we enter 'The Church' and we are thus moved another stage further away from the experience offered to the reader by the first edition. F. E. Hutchinson (Oxford, 1941) retained the original layout and shows his sense of its importance in a note about the placing of 'SUPERLIMINARE' in the only two authoritative manuscripts. It could be argued, on the grounds of layout alone, that of these three editors only Hutchinson presents the full text of the four related poems, since here layout is actually part of the text.

It would not be difficult to think of other authors in whose writings bibliographical or typographical features form part of the sense of the text. What is more problematic (as with numerology) is to decide which are the authors for whom such characteristics are likely to be the results of conscious decision. But until or unless such problems can be solved, a conservative attitude towards the physical arrangement of a text on its page is wise. For example, in the case of Donne's last sermon (published under the non-authorial title Deaths Duell), it is unwise to strip away its margin notes, as John Hayward did in the Nonsuch Donne (1955) and as Neil Rhodes has done in the Penguin John Donne: Selected Prose (1978), despite the fact that Rhodes takes his text from the great Potter and Simpson edition of the Sermons (1953-62), which retains them. The economics of publishing being what they are, one sees that an editor might come under irresistible pressure to do away with margin notes; in that case, though, margin notes might be transferred to the page foot, as in the Ann Arbor edition of Donne's Devotions... together with Death's Duel (1959).

But, unfortunately, the Ann Arbor edition does not reprint all the original margin notes: it reprints biblical references but omits the far more important notes which indicate division of the sermon's thought. The same is the case with the John Carey Oxford Authors Donne (1990), which transfers the biblical references into the text, in square brackets, but ignores the division notes. These division notes are all the more important because one could supply the biblical references for oneself, at need, with the aid of Cruden's Concordance; but most modern readers could hardly supply divisions such as the threefold 'Exitus a morte vteri', 'Exitus a mortibus mundi', 'Exitus a morte Incinerationis' and indications such as are supplied by the extremely interesting margin note near the end (1. 579), 'Conformitas'. In what sense such margin notes are text, or what Gerard Genette has termed paratexte, may be debatable; but it must be a pity that the reader has to go to the ten-volume Potter and Simpson edition before he can find them.(3) I shall argue that there are other characteristics of the first edition, absent even from Potter and Simpson, that ought also to be reinstated for a full reading of Donne's work.

At this point it is necessary to consider what authority the physical characteristics of seventeenth-century printed texts can possess. Circumstances of publication differ widely. Donne's last sermon (like Herbert's The Temple) was published posthumously. Is it possible, then, that its physical details could be authoritative? Ought we not rather to attribute them solely to Donne's editor or to his printer? And we are on less secure ground with Donne's last sermon than with Herbert's Temple. The Bodleian manuscript of 'The Temple', a Little Gidding transcript produced for the Cambridge University licensers and, although not the printer's copy for the first edition, almost certainly closely related to it, confirms the arrangement of the first printed edition. It is probable that both MS Tanner 307 and the printed edition accord with the manuscript which Herbert, according to Walton, sent from his death-bed to Nicholas Ferrar at Little Gidding.(4) How closely any seventeenth-century edition of Donne's last sermon conformed to Donne's own manuscript is less easily decided: the manuscript is not known to survive. Nevertheless, one can assess probabilities, and I share the belief of Evelyn Simpson and Helen Gardner that Donne prepared his last sermon for the press, and that he intended the sermon to be published posthumously.(5)

Donne had published six sermons during his lifetime, and Evelyn Simpson shows that there are features of typography, spelling, and punctuation common to these sermons and to Deaths Duell: I shall discuss these later.(6) Further, the peculiar circumstances of the composition of Deaths Duell invites a comparison with the preparation and publication of Donne's Devotions, that highly individual combination of formal meditation with an autobiographical account of his own nearly mortal sickness.(7) If Simpson and Gardner are right, as I believe, the publication of the first edition of Deaths Duell in 1632 represents the fulfilment of Donne's own wish. There is some circumstantial evidence to suggest that this is so.

Donne preached his last sermon on the first Friday of Lent in 1630/1 (25 February).(8) It was preached at Whitehall, before the King and the Court, to a congregation, therefore, accustomed to the sophisticated style of preaching established by Lancelot Andrewes.(9) Donne died on 31 March and his will was proved on 5 April. The will, dated 13 December 1630, names Henry King and John Montfort as executors. Mrs Simpson assumed that as an executor Henry King was responsible for the publication of the last sermon. Considerably later, in an open letter to Izaak Walton, King claimed that Donne 'but three days before his death delivered into my hands those excellent Sermons of his now made publick: professing before Dr. Winniff, Dr. Monford, and, I think, your self then present at his bed side, that it was by my restless importunity, that he had prepared them for the Press'.(10) King was therefore what we would describe today as a literary executor. His probable responsibility for publishing Deaths Duell is further corroborated by the fact that in its first edition the sermon is followed by two anonymous elegies, both reprinted in the first edition of Donne's Poems (1633) and there identified as the work of (respectively) 'H.K.' and 'Edw. Hyde'. The poem by 'H.K.' ('To have liv'd eminent, in a degree') appears in the three manuscript collections of the poems of Henry King and in the 1657 edition of his Poems, reissued in 1664; King's editor, Margaret Crum, does not question its authorship. If the elegy is indeed by King (and all the evidence points that way), its presence in the 1632 edition of Donne's last sermon seems to confirm the involvement of Donne's literary executor in the sermon's publication. Finally, as Grierson pointed out in 1912, 'It was to King also that Redmer [one of the stationers responsible for publication of the last sermon] was indebted for the frontispiece to Deaths Duell, the picture of Donne in his shroud... [This engraving is based on a drawing made of Donne on his death-bed.] "It [the portrait] was given", Walton says, "to his dearest friend and Executor Dr King, who caused him to be thus carved in one entire piece of white Marble, as it now stands in the Cathedral Church of St. Pauls."'(11)

We are now less ready than in Grierson's day to give unhesitating credence to Izaak Walton's factual statements. However, one of his statements here can be confirmed: King's involvement in the commission of the St Paul's monument may be demonstrated because of the survival of the accounts book of its sculptor, Nicholas Stone. In 1979 Helen Gardner made a thorough investigation of the circumstances surrounding the commissioning of the Donne statue in an article 'Dean Donne's Monument in St. Paul's'.(12) There she repeated and augmented arguments previously advanced by herself and Evelyn Simpson. First, the engraved portrait by Droeshout of Donne in his shroud which serves as a frontispiece to Deaths Duell is a genuine likeness, and must therefore indeed be based on the drawing of Donne in his shroud, made on his own authority, which is described in Walton's Life of Donne, and which is also the basis of the marble monument in St Paul's. Secondly, of the inscription below the engraved portrait, 'Corporis haec Animae sit Syndon, Syndon Jesu', Dame Helen argues 'Nobody but Donne could have written this motto'. That view (although she does not say so) is presumably based in part on John Sparrow's perception that the word syndon (shroud or winding sheet) is a pun on 'sin done'.(13) This is a pun which may be set beside the pun in 'Hymn to God the Father' ('Wilt thou forgive that sinne? . . . When thou hast done, thou hast not done') and the pun described by Walton which Donne made when he realized that his marriage had caused the ruin of his fortune (John Donne, Ann Donne, Un-done).(14) If we are willing to accept, on these grounds, that only Donne could have written this motto, we may accept Gardner's further argument that the motto, with its play on 'shroud', 'can only have been written with one purpose: to stand below this picture'.

It might, I think, be argued against this view that although the drawing of Donne in his shroud which underlies both the Droeshout engraving and the marble monument must have existed (they clearly derive from a common source), it need not have been made with Donne's authorization or even during his lifetime; Walton may be wrong on both counts. The drawing could be a true portrait taken after death. And an admirer of Donne's style (such as Henry King) might have fudged up the 'Corporis haec Animae' motto in the sincerest form of flattery. But in that case, the reference to sin (even sin that is done) might seem a little out of place. And equally tellingly, the commissioning of a portrait is highly characteristic of Donne. Donne shared with his friend Sir Henry Wotton an interest in the visual arts (as may be illustrated by the pictures listed in his will), and that he was particularly interested in portraits of himself is evidenced by the number of separate authentic likenesses of him, at different stages of his life, that have come down to us. If one considers the known likenesses of, say, Shakespeare, Ben Jonson, or George Herbert, it becomes remarkable that we have five separate images of Donne, taken by some of the foremost artists of his day, and the more remarkable when we remember that Donne was short of money for much of his life.(l5) Since we know of this strong attachment to having his likeness taken, the impulse to have this final portrayal made seems the more likely to have come from Donne himself. Nor is it safe to assume that because a fact is recorded by Walton in his Life of Donne it must necessarily be untrue: Walton must be a better witness for the end of Donne's life than for its earlier stages, since he had unquestionably come to know Donne well during the latter's last years.(16)

Against the view that the publication of Donne's last sermon was authorized by Donne himself (and therefore prepared by him for the press) Sir Geoffrey Keynes advanced the delay between Donne's death (which, as we have seen, was on 31 March 1631) and the publication of the sermon in 1632.(17) However, an article published by Sir Geoffrey himself in the Times Literary Supplement can be put forward to weaken his case, and suggest external circumstances which would have enforced postponement.(18)

In 1938, Sir Geoffrey records, Franklin B. Williams drew to his attention the existence of a rare book entitled La Dance Machabre or Death's Duell by W.C., with the imprint 'LONDON Printed by William Stansby. [1632?] '.(19) W.C. signed his dedicatory poem to the Queen 'W. Colman' (sig. [A2.sup.v]), and has been identified as Walter Colman, an observant Franciscan friar educated at Douai.(20) What, about this book, is of interest to our present purpose is the addition to it Colman made which occurs on sig. [G1.sup.r], and which consists of two verses, addressed by Colman to 'Roger Muchill'. Of these it is sufficient to quote the first:

DEath in a furie hath the Fellon tooke That stole my Title, Donne, to grace thy booke. To wrong the living and commit a rape Upon the dead, how could he thinke to scape? I am but too much honord to be stil'd Th'unwilling Gossip to thy unknowne child. But he that sought so basely my disgrace Behind my backe; hath wrong'd thee to thy face.


I would revenge thy quarrell but that he That deales with dirt shall but defiled be. Live in thy living fame; and let this serve Not thine, but mine owne honor to preserve.

From this we learn that the stationer Roger Michell (Colman's spelling enables him to make play with 'Much-ill' in his second verse) was associated with Donne's last sermon before it came into the hands of the stationers whose names appear on its title-page. Its imprint declares that it was 'Printed by THOMAS HARPER, for Richard Redmer / and Beniamin Fisher, and are to be sold at the signe / of the Talbot in Alders-gate street. / M.DC.XXXII.', and it was entered to Redmer and Fisher in the Stationers' Register on 30 September 1631.(21)

In addition to the fact that Colman's protest identifies the sermon's title as non-authorial (and that is not surprising: none of the sermons published in Donne's lifetime have titles), it is also clear that we have now a reason for a delay in publication. As Colman's less than charitable verses make clear, Roger Michell had died before the publication of La Dance Machabre. When exactly he died we cannot be sure. La Dance Machabre was entered in the Stationers' Register on 13 June 1631, but Michell might have died after that date, since Colman's verses about him could well have been added to the book as an afterthought: the concluding leaf on which they appear is a singleton.(22) The last copyright Michell entered in the Stationers' Register (STC(2) 15463) was entered on 29 April, so he was still alive then, and he was presumably dead by the time that Donne's sermon was entered in the Stationers' Register to Redmer and Fisher on 30 September 1631.(23) But the exact date of Michell's death hardly affects the issue. It seems clear that the responsibility for publishing Donne's last sermon was first entrusted to a stationer who died before it could be published. It had then to be retrieved, and agreement reached with fresh stationers. The date 1632 on its title-page does not necessarily mean that it did not appear until that calendar year; it could have appeared late in 1631.(24) Indeed, the fact that the book had not been entered to Michell but was entered to the men whose names appear on its title-page may mean that the book was registered, not from manuscript, but from a printed copy.(25) The gap between Donne's death and the book's appearance is probably not very great, and Michell's death sufficiently accounts for it.

And finally, in addition to the circumstantial evidence relating to the question of whether or not Donne prepared his sermon for the press himself, there is the important textual evidence put forward by Evelyn Simpson. She explains that the italicization of the 1632 edition resembles that of the six sermons published during Donne's lifetime: 'throughout Deaths Duell "God" and "Christ" are italicized as they are in our Vol. IV, No. 7, lines 1, 6, 10, 14, 15, 18, 26, et passim, while in the Folio they are consistently printed in roman. Again in the Quarto of Deaths Duell we frequently find italics used for any words on which Donne desired to lay a slight stress, as similarly in the sermons published in Donne's lifetime.'(26) She makes a similar point about spelling. Her reason for putting forward these arguments is to defend her decision to take the 1632 Quarto of Deaths Duell as her copytext rather than the version published in 1660/1 by John Donne the younger in XXVI Sermons.

As Mrs Simpson says, the text of Deaths Duell in XXVI Sermons was not printed from the Quarto of 1632 (as the Second Quarto of 1633 had been), but has independent authority: it contains passages which had been omitted from the 1632 Quarto through homoeo-teleuton.(27) But although the manuscript which underlies the 1660/1 Folio may have equal authority with the manuscript which underlies the 1632 Quarto, it had been prepared for the press by John Donne the younger, not by his father. What intervention may have been made by John Donne the younger, or what 'corrections' he may have permitted to the printer which his father might not have licensed, it is impossible to know: it is evident that the spelling is that of 1660, not 1631, and there may be other alterations too. Donne the younger makes it plain in his epistle 'To the Reader' that he is publishing for money: 'Fac hoc & vives, was often repeated to me when I undertook this work, and a fair reward was promised me; but at last I am constrained hoc facere ut vivam, to publish them at my own cost that I may sell them.' John Sparrow suggests that Donne the younger profited from his father's sermons not only by publishing them but by preaching them: 'That he preached them, or intended to preach them, is proved by a letter from him which accompanied a presentation copy of Biathanatos,(28) in which he says: "I was encouraged to undertake [the printing of LXXX Sermons] by the learnedest men in the king-dome, of all professions, and was often told, that I should deserue better by doinge soe, then by keeping them to my owne vse, for by this meanes, I did not only preach to the present adge, but to our childrens children."'(29) If Sparrow is right, and Donne the younger did preach his father's sermons, it is possible that he may also have revised them for his own purposes.

For various reasons, then, it seems that Evelyn Simpson was absolutely correct in taking 1632 as her copytext and not 1660/1. Indeed, I would go further than she does, for she adopts punctuation from 1660, despite her own statement that the light punctuation of the 1632 Quarto is characteristic of sermons published during Donne's lifetime. Where 1632 is obviously incorrect (as with the omitted phrases) it is plainly right to correct from 1660/1.(30) Nevertheless, for example, I would not follow her in reading 'purchas'd' from 1660/1 against 'prepared' from 1632 in the sermon's last sentence. 1632 reads: '... that Kingdome, which hee hath prepared for you, with the inestimable price of his incorruptible blood. AMEN.' 1660/1 reads: '... that Kingdome which he hath purchas'd for you, with the inestimable price of his incorruptible blood. Amen.' Both phrases can be represented as being scriptural, and it is not obvious that the 1660/1 reading is preferable.(31)

But while I agree with Evelyn Simpson's choice of 1632 for her text, one of my principal points in the present paper is to say that I do not think Mrs Simpson was wise to reorganize the paragraph divisions so that they are not those either of 1660/1 or of 1632. For, I hope to demonstrate, paragraph division in this sermon constitutes an important part of the sermon's sense. I shall argue that its paragraph divisions are textual, in precisely the same sense that the layout of Herbert's 'Easter wings' or 'The Altar' are textual. And I now come at last to speak of the sermon itself.(32)

In setting out his text at the sermon's start, Donne makes use of a threefold division. His text, he explains, is open to three distinct interpretations, each one of which he introduces in turn. The text is from Psalm 68, verse 20, and Donne gives it at his sermon's head: 'And unto God the (Lord) belong the issues of death. i.e. From death.' Following his usual custom, Donne also quotes his text from the Vulgate, using the phrase 'exitus morris' (1. 18). This, he tells us, may be read in three possible ways: it can be understood as 'liberatio el morte, a deliuerance from death' (1. 19), it can be understood as 'liberatio in morte, A deliuerance in death' (11.33-4), or, following St Augustine, it can be understood as 'liberatio per mortem, a deliverance by death, by the death of this God our Lord Christ Jesus' (ll. 45-6).(33)

Each of these three interpretations may be associated with a different Person of the Trinity: the first, deliverance from death, is through the power of the Father; our preservation during the second, deliverance in death, is through the presence of the Spirit; and, thirdly, as St Augustine has shown, our salvation is by means of the death of Christ. In addition, the threefold interpretation of the text offers a threefold support of the immediately preceding verse of the Psalm. The sermon opens:

BVILDINGS stand by the benefit of their foundations that susteine and support them, & of their butteresses that comprehend and embrace them, and of their contignations that knit and unite them: The foundations suffer them not to sinke, the butteresses suffer them not to swerue, and the contignation & knitting suffers them not to cleaue; The body of our building is in the former part of this verse: It is this, hee that is our God is the God ofsaluation and salutes; of saluation in the plurali, so it is in the originall; the God that gives vs spirituall and temporall saluation too. (ll. 1-9)

The threefold interpretation of the text has a threefold application (foundations, el morte, the Father; buttresses, in morte, the Holy Spirit; contignations, per mortem, Christ), and, as we shall see, this ninefold pattern of the sermon's opening is repeated, in a yet grander form, at the sermon's close.

The sermon's opening metaphor, that of the building, remains, throughout the sermon, of central importance. Helen Gardner is surely correct in repudiating Stanley Fish's facile association of the Building with a House of Memory.(34) As she says, Donne is not concerned with rooms and their furniture: he is concerned with the idea that his 'building' is the statement 'hee that is our God is the God of salvation'. As the sermon develops, we discover that this building of salvation is a container which may be equated with Noah's ark (a familiar type of the Church), with Moses's ark or cradle, with the womb, with the tomb, with the prison of this life, with the many sepulchres and graves of earth where we have no continuing city and no abiding stay, with the many mansions or staying places of 'the days of the years of my pilgrimage', and with the many mansions which we will finally find in our Father's house (John 14: 2). This imagery is centred in the third paragraph of the sermon (11.99-364), although it is pervasive. In this third paragraph we are reminded that though our body, the Temple of the Holy Ghost, comes to ruin, rubbish, and dust (11. 328-9), yet Christ lay three days in the tomb uncorrupted and rose on the third day (11. 216-22). Later we learn that death is a deliverance from prison: the gates of our prison may be (1) opened by an oiled key, or (2) hewn down, or (3) burnt, according to the manner of death we may die (11. 435-8). The building metaphor, then, is of continuing significance: but something of equal importance is established in the opening paragraph, and that is the ninefold patterning, which is echoed throughout the sermon by various groups of three.

In the final paragraph of the sermon (the seventh), we are brought, as the side note reminds us, to '3. Part. Liberatio per morte', that is, our liberation through the death of Christ. After much consideration of Christ's death, Donne comes to consider the Transfiguration, when (according to Luke's Gospel) Moses and Elias talked with Christ about his death, and Peter suggested that the disciples should then build three tabernacles, one for Christ, one for Moses, and one for Elias (Luke--the third Gospel--9: 30-33) (11. 567-72). Nothing, of course, came of that suggestion, but Donne picks up St Peter's 'Master, it is good for us to be here' and continues: 'It is good to dwell here, in this consideration of his death, and therefore transferre wee our tabernacle (our devotions) through some of those steps which God the Lord made to his issue of death that day.' (11. 575-8)

The tabernacle of our devotions turns out to be a ninefold meditation on the events of Christ's last day of life:

Take in the whole day from the houre that Christ receiued the passe-ouer vpon Thursday, vnto the houre in which hhe dyed the next day. Make this present day that day in thy deuotion, and consider what hee did, and remember what you haue done. Before hee instituted and celebrated the Sacrament, (which was after the eating of the passeouer) hee proceeded to that act of humility, to wash his disciples feete, euen Peters, who for a while resisted him; In thy preparation to the holy and blessed Sacrament, hast thou with a sincere humility sought a reconciliation with all the world, euen with those that haue beene auerse from it, and refused that reconciliation from thee? If so and not els thou hast spent that first part of his last day, in a conformity with him. After the Sacrament hee spent the time till night in prayer, in preaching, in Psalmes; Hast thou considered that a worthy receaving of the Sacrament consists in a continuation of holiness after, aswell as in a preparation before. If so, thou hast therein also conformed thy selfe to him, so Christ spent his time till night; At night hee went into the garden to pray, and he prayed prolixious he spent much time in prayer, how much? Because it is literally expressed, that he prayed there three severall times, & that returning to his Disciples after his first prayer, and finding them a sleepe sayd, could ye not watch with me one houre, it is collected that he spent three houres in prayer. I dare scarce aske thee whither thou wentest, or how thou disposedst of thy self, when it grew darke & after last night: If that time were spent in a holy recommendation of thy selfe to God, and a submission of thy will to his, It was spent in a conformity to him. (ll. 579-604)

Here we have the first three meditations: (1) on the washing of the disciples' feet; (2) on Christ's prayer, preaching, and Psalms until night; (3) on his three hours of prayer in the garden. Six more follow: (4) on the agony and bloody sweat; (5) the taking with a kiss and binding at midnight; (6) the return to Jerusalem, questioning, buffeting, and denial by Peter; (7) the appearance, at dawn, before Pilate for judgement; (8) about eight of the clock, the transference by Pilate to Herod and Herod's referral back to Pilate; (9) the scourging, crowning with thorns, and call for crucifixion. Finally, a tenth recollection of the events of Good Friday brings us, not to a 'conformity' question, but to a statement in the historical present:

Towards noone Pilat gaue iudgement, and they made such hast to execution, as that by noone hee was vpon the Crosse. There now hangs that sacred Body vpon the Crosse, . . . There wee leaue you in that blessed dependancy, to hang vpon him that hangs vpon the Crosse, . . . (ll. 651-69)

This conclusion in the historic present, and the actual present of the sermon-preaching (the first Friday of Lent in 1630/1, 25 February, as we have seen),(35) brings us out of the sequence of time with which this sermon is otherwise preoccupied, and allows a glimpse of the eternal. Now which, it may be believed, will be the soul's experience after death, when, just as it will be delivered from the series of containers--the womb, the cradle, the prison, the tomb--through which, chrysalis-like, it has proceeded during life, so it will be delivered from the series of deaths from one state to another, by the culminating death-birth into eternity. This construction had been in part anticipated in The Devotions, which ends with a twenty-third section of vigilance: the last hour before death.

Each of the preceding nine meditations has within itself three parts: recollection of what Christ was doing at a particular hour between the Passover and the Crucifixion; questioning what the listener (or reader) was doing at the equivalent hour; meditation as to whether or not this was in conformity to the experience of Christ.(36) This patterning is marked from its outset ('Take in the whole day from the houre that Christ receiued the passe-ouer . . .') with the margin note 'Conformitas' (l. 579), a word which evokes the Spiritual Exercises of St Ignatius Loyola.

Indeed, Donne's set of nine Maundy Thursday to Good Friday meditations may be compared to the first nine meditations in the Third Week of the Spiritual Exercises: (1) the Last Supper, including the washing of the disciples' feet; (2) the Agony in the Garden, including Christ's prayers and the agony and bloody sweat; (3) the events after leaving the Garden until the end of what happened at the house of Annas; (4) from the house of Annas to the events in the house of Caiphas; (5) from Caiphas to Pilate; (6) from Pilate to Herod; (7) from Herod to Pilate; (8) the events at Pilate's house; (9) from Pilate's residence to the nailing on the cross; (10) from the nailing on the Cross to Christ's death. Ignatius' choice of division of the events from the Last Supper to the Crucifixion into nine meditations (ending half-way through his fifth day, leaving him the rest of his Week for the Deposition, the Entombment, and a meditation on the whole of the Passion) is not the same as Donne's; but that he has a ninefold division of these events probably had its influence on Donne.

It may further be said of Donne's ninefold Passion meditation that it contains internal numerical ordering as well. Thus Donne's third meditation contains an unscriptural assumption that Christ spent three hours in prayer (ll. 595-600); his median meditation, the fifth, occurs at midnight (l. 609); his sixth meditation contains six subdivisions ('irrisions, and violences, the covering of his face, the spitting vpon his face, the blasphemies of words, & the smartnes of blowes'), followed by a separate and seventh division, 'that Gallicinium, that crowing of the Cock which called vp Peter to his repentance' (ll. 612-19); his eighth meditation occurs at eight of the clock (ll. 635-6). As I remarked earlier, it is not always easy to predict which authors are or are not likely to make use of numerology, and I would not venture here to suggest more than that Donne is casting his ninefold meditation into a highly patterned form. However, it is doubtless relevant that the source of the third interpretation of the sermon's text, Augustine's City of God, is also an important early source of numerological theory: chapters 30 and 31 of Book 11 are devoted respectively to 'the perfect number six', and to 'the seventh day, whereon completeness and rest are introduced'.(37)

If St Ignatius' Spiritual Exercises are one probable influence on Donne's ninefold Passion meditation, it is perhaps natural to wonder if it has also been influenced by the Stations of the Cross, a form of liturgical devotion which was still undergoing development and change at the time at which Donne was writing. Although the number of Stations is now fixed at fourteen, their number was originally variable: Herbert Thurston gives a list of different orders recorded between 1294 and 1744.(38) Certainly Donne contemplates the theme of Stations, or resting places, in Deaths Duell, as we have seen:

Whatsoeuer moued Saint Ierome to call the iournies of the Israelites, in the wildernes, mansions; The word (the word is Nasang) signifies but a iourney, but a peregrination. Euen the Israel of God hath no mansions; but iournies, pilgrimages in this life. By what measure did Iacob measure his life to Pharoah; the dayes of the years of my pilgrimage. (ll. 152-7)

It is evident that although some reflection on the Stations of the Cross may underlie Donne's thought in the construction of his last paragraph (as, too, it may underlie St Ignatius' thought in the construction of the Third Week of his Spiritual Exercises), Donne's last paragraph shares the emphasis of the whole sermon, which, by this stage, has moved from consideration of place to consideration of time. Meditation about the womb, the tomb, the prison of this life, has been centred in paragraph three; paragraph six (that immediately preceding the final paragraph, with its ninefold meditation) refocuses our thoughts on time.

This sixth paragraph considers a wide variety of units of time. Starting with the reflection that we cannot predict the outcome of the hour of death (ll. 376-9), it speaks of our life of three score years and ten (l. 392) (a life, that is, of seventy years, in contrast to the life of Christ of thirty-three years, although Donne himself does not make this comparison); it speaks of the day of our death, hodie, as both the first day and the last day (ll. 393-8); we are to 'Make no ill conclusions upon any mans loathnes to dye' (ll. 399-400), for 'the mercies of God worke momentarily in minutes' (ll. 379-80); the critical day is not the day of our death, but of the whole course of life (ll. 411-31); but the unlocking key of our prison is the hour of our death (l. 435).

An earlier paragraph, paragraph three, had introduced the idea of the span of life, not as a day, but as a week:

That which we call life, is but Hebdomada mortium, a weeke of death, seauen dayes, seauen periods of our life spent in dying, a dying seauen times ouer; and there is an end. Our birth dyes in infancy, and our infancy dyes in youth, and youth and the rest dye in age, and age also dyes, and determines all. Nor doe all these, youth out of infancy, or age out of youth arise so, as a Phoenix out of the ashes of another Phoenix formerly dead, but as a waspe or a serpent out of a caryon, or as a Snake out of dung. (ll. 164-71)

Here Donne seems to be indebted to the Ages of Man, which, as John Burrow has shown us, can, like the Stations of the Cross, be variable in number at this date.(39)

Donne appears to offer us two schemes: the seven Ages of Man, which divide life up into a heptameron pattern, and the simpler threefold division of Infancy, Youth, and Age. The sevenfold division obviously relates to three score years and ten. It can also be linked to the Week of Creation, as in Du Bartas's Semaines, widely familiar in England at this date as The Divine Weeks and Works, translated by Josuah Sylvester.(40) But Donne is offering us a deliberate twist to this convention, by giving, not a Week of Creation, but 'Hebdomada mortium, a weeke of death, seauen dayes, seauen periods of our life spent in dying, a dying seauen times ouer'. The threefold definition of the week of death, 'seauen dayes, . . . seauen periods . . . seauen times ouer', makes way for the threefold division that follows. And Donne's sermon culminates, not in the seventh day of rest, Sunday, but in the day of Christ's death, Good Friday. More exactly, Donne's sermon culminates, not with a day, but with a night. The ninefold meditation of his concluding paragraph covers not a twenty-four hour period, but the period between the Passover and noon the following day. It seems, indeed, appropriate that a Week of Death should be measured by the hours of darkness rather than by those of light. Early in paragraph seven Donne says: 'this part of our Sermon must needes be a passion Sermon; since all his life was a continuall passion, all our Lent may well bee a continuall good Fryday' (ll. 474-7). But if it is 'this part' of the sermon (the seventh paragraph) that is a Good Friday (and its vigil), what of the rest of the sermon?

The curious paragraph structure affords, I believe, a clue to the answer. And the paragraph structure is indeed curious, which is doubtless why so many of Donne's editors have succumbed to the temptation to tidy it up.(41) There are, as has been said, seven paragraphs. Two of these are inordinately long, two inordinately short. The long paragraphs are paragraphs three and seven (numbers which we have seen to be important throughout the sermon), and the two short paragraphs are paragraphs four and five. Paragraphs one, two, and six, although varying considerably in length, may be regarded as 'normal' paragraphs. The exact word numbers are: paragraph one, 762; paragraph two, 465; paragraph three, 3,254; paragraph four, 60; paragraph five, 50; paragraph six, 853; paragraph seven, 2,787. The word-numbers of paragraphs four and five seem curiously neat, but I am not aware of any particular significance in these numbers, beyond their obvious disparity.(42)

This disparity seems so odd as to suggest a deliberate purpose. I wish to suggest that the strangely proportioned paragraph structure is intended to relate to the ninefold meditation of paragraph seven, so that, if paragraph seven contains Maundy Thursday evening to noon on Good Friday, the last day (or night) of the Week of Death, the other paragraphs are intended to represent the other days of the week. In that case, paragraph one represents Friday/Saturday, paragraph two represents Saturday/Sunday, paragraph three (especially long) represents Sunday/Monday, and so on. The two especially long paragraphs, three and seven, then, each starting the evening before, culminate at noon, on, respectively, Monday and Friday. These days are significant within the sermon's scheme, because Monday starts the Week of Creation and Friday concludes the Week of Death. Thus the whole sermon, which the opening paragraph declares to be a building in the dimension of space, is also a week in the dimension of time.

A final point may be made. Whether or not my association of the seven paragraph structure with the seven days of the week be accepted, the tripartite structure of the sermon is clearly Donne's own, since it is signalled in margin notes: '1. Part', '2. Part. Liberatio in morte', '3. Part. Liberatio per morte'. The three Part notes (of which we may conjecture that the first should be expanded in accordance with the other two, since the text at that point reads 'First, then, we consider this exitus mortis, to bee liberatio a morte') divide the sermon into three very unequal parts. Again, it seems, the very disparity invites our attention. The first of these margin notes lies alongside the opening of the second paragraph; the second, although alongside the opening of the sixth paragraph, ought to be alongside the opening of the fifth paragraph ('And so wee passe vnto our second accommodation of these words (vnto God the Lord belong the issues of death'); and the third of these side notes comes alongside the opening of the seventh paragraph. The tripartite division of the text therefore (a morte, in morte, per mortem) divides the sermon in the following way: the first paragraph is introductory, and explains the tripartite division which follows, adding the building metaphor, and the associations with the three persons of the Trinity; then the first section, associated with God the Father, occupies three paragraphs; the section associated with God the Holy Ghost occupies two paragraphs; and the section associated with God the Son occupies a single paragraph. Here we have, in fact, 3 + 2 + 1: the reason why St Augustine believed six to be a perfect number: 'It is because six is a perfect number that it is told that the creation was completed in six days, the same day being repeated six times, . . . Six is the first number that is the sum of its factors, that is, its sixth part, its third part and its half, which are one and two and three; these added together make six.'(43) I quote from The City of God (although St Augustine makes the same point elsewhere), which continues in its next chapter to consider the number seven, adding the day of rest to turn the six days of creation into the Week of Creation.

It is evident that this sermon is very highly structured, and structured in ways that would not be wholly available to a congregation who heard the sermon preached. Deaths Duell, so often anthologized in selections of Donne's writings, is in fact very untypical of his sermons, having more in common, in many ways, with the Devotions. The superbly baroque closing sentence shows us that even in his last months Donne was continuing to develop his amazing skills as a writer of prose, which, here, powerful as it was when preached, is also meant to be read; and to be read with careful attention to its layout on the page.

1 This point was made forcefully by Professor D. F. McKenzie in his Lyell Lectures in Oxford, 1988, which have, unfortunately, been published in part only: 'Speech-Manuscript-Print', New Directions in Textual Studies, ed. D. Oliphant and R. Bradford (Austin, Tex., 1990), 86-109; also published as vol. 20, nos 1/2 of The Library Chronicle of the University of Texas at Austin (1990).

2 Here, and in all quotations and transcriptions that follow, I have normalized long-s.

3 G. Genette, Seuils (Paris, 1987).

4 MS Tanner 307: 'except that the two-line "Anagram" is placed earlier, between "Church-Musique" and "Church-lock & key"'. The Works of George Herbert, ed. F. E. Hutchinson (Oxford, 1941), p. li.

5 E. Simpson in The Sermons of John Donne, ed. G. R. Potter and E. M. Simpson (Berkeley and Los Angeles, 1953---62), x. 275-6; H. Gardner in Evidence in literary scholarship: Essays in memory of James Marshall Osborn, ed. R. Wellek and A. Ribeiro (Oxford, 1979), 34.

6 Although Deaths Duell is not an authorial title it is convenient to continue to use it.

7 There is an excellent recent study of Donne's Devotions by K. G. Frost, Holy Delight: Typology, Numerology, and Autobiography in Donne's Devotions Upon Emergent Occasions (Princeton, NJ, 1990).

8 R. C. Bald, John Donne: A Life (Oxford, 1970), 526; Edmund Gosse incorrectly stated that the first Friday in Lent in 1630/1 fell on 12 Feb.: it fell on 25 Feb., as Evelyn Simpson pointed out to Geoffrey Keynes (Keynes, A Bibliography of Dr. John Donne Dean of Saint Paul's (Oxford, 1973), 52). (It had fallen on 12 Feb. in 1629/30.) This error has caused some writers (Bald, John Donne, 525-6, and H. Gardner, In Defence of the Imagination (Oxford, 1982), 98, doubtless following the usually reliable Bald) to suppose, mistakenly, that Donne was to have preached on 12 Feb. (in fact a Saturday) and that illness caused him to delay to 25 Feb.

9 I am indebted to Miss Eluned M. Brown for stressing this point to me.

10 I. Walton, Lives (1670), sig. B[1.sup.v].

11 The Poems of John Donne, ed. H. J. C. Grierson (Oxford, 1912), ii. 255.

12 In Evidence in literary scholarship, 29-44.

13 Times Literary Supplement, 13 Mar. 1953.

14 J. Donne, The Divine Poems, ed. H. Gardner (Oxford, 1952; 1959), 51; Donne's pun about his marriage appears in the 4th edition of Walton's Life of Donne--Lives (1675), 18.

15 The Donne portraits are: (1) the lost original (perhaps by Hilliard) upon which the Marshall engraving of Donne as a youth in 1591 is based; (2) the Lothian portrait, probably painted about 1595, described in Donne's will as 'that Picture of mine [] is taken in Shaddowes and was made very many years before I was of this profession', and which is now at Monteviot in the Scottish Borders; (3) the 1616 Isaac Oliver miniature in the Royal Collection (upon which the National Portrait Gallery half-length is based, and which is also the basis for the Merian engraving in LXXX Sermons); (4) the portrait at the St Paul's Deanery (1620); (5) the lost original (1631) which provides the basis, both for the Droeshout engraving of Donne in his shroud prefixed to Deaths Duell, and for the Nicholas Stone white marble effigy in St Paul's cathedral.

16 See J. Bevan, 'Henry Valentine, John Donne, and Izaak Walton', RES NS 40 (1989), 179-201.

17 G. Keynes, A Bibliography of John Donne (Oxford; 4th edition, 1973), 53.

18 Times Literary Supplement, 24 Sept. 1938.

19 ST[C.sup.2] 5569; there is a copy of this book in the Bodleian, press-mark Malone 404.

20 See ST[C.sup.2] 5569 and DNB.

21 The entry, taken from a microfilm of the SR, is as follows (I have normalized long-s):

30 Sept 31

Beniamin ffisher & Richard Redmer Entred for their Copy under the handes / of [m.sup.r]. Buckner & the Wardens a / Sermon preached by [D.sup.r]. John Dunn / upon the ^ [[interlineated:] [] verse of [y.sup.e]] [] Psalme at whitehall / before the king

22 The book's collation is 80, [[Pi].sup.2] [A.sup.4] [[Chi].sup.1] B-[E.sup.4] [F.sup.4] [G.sup.1].

23 Unfortunately it appears from D. F. McKenzie, Stationers' Company Apprentices 1605-1640 (Charlottesville, 1961) that Michell had no apprentices, so evidence for the date of his death cannot be derived from the transference of apprentices to another master; nor have I found record of his death in the parish registers of St Gregory by St Paul's, St Martin Ludgate, St Andrew by the Wardrobe, St Ann Blackfriars, St Bride, St Benet, St Mary Magdalen Fish Street, or St Dunstan-in-the-West.

24 This seems to have been general practice later. P. Gaskell, A New Introduction to Bibliography (Oxford, 1972), 317-18, writes: '"The Rule in general observed among Printers," wrote Nichols concerning eighteenth-century practice, "is, that when a Book happens not to be ready for publication before November, the date of the ensuing year is used."' [J. Nichols, Literary Anecdotes (London, 1812), iii. 249 n.]

25 I am indebted to Professor Desmond Neill for drawing my attention to the distinction between SR entries from manuscript and those from print. The title-page reads as follows (I have normalized long-s): [Within a double rule] DEATHS / DVELL, / OR, / A Consolation to the Soule, against / the dying Life, and liuing / Death of the Body. / Deliuered in a Sermon at White Hall, before the / KINGS MAIESTY, in the beginning / of Lent, 1630. / By that late learned and Reuerend Diuine, / IOHN DONNE, [D.sup.r]. in Diuinity, / & Deane of S. Pauls, London. / Being his last Sermon, and called by his Majesties houshold / THE DOCTORS OWNE FVNERALL SERMON: / [orn.] / LONDON, / Printed by THOMAS HARPER, for Richard Redmer / and Beniamin Fisher, and are to be sold at the signe / of the Talbot in Alders-gate street. / M.DC.XXXII. / (This may be compared with the SR entry, n. 21 above.)

26 Sermons, ed. Potter and Simpson, x. 274.

27 Ibid. 275-6.

28 This copy of Biathanatos is in Cambridge University Library, press-mark G. 11. 8. The letter is printed in full in the 4th edition of Keynes's Bibliography of Donne, 114; where Sparrow's transcript (1930) differs from Keynes, I have followed Keynes.

29 J. Sparrow, 'John Donne and Contemporary Preachers: Their Preparation of Sermons for Delivery and Publication', Essays and Studies by Members of the English Association, xvi (1930), 167-8.

30 I am indebted to Professor John Richardson and to Dr David Mealand for investigating the word 'hestae' which occurs twice in 1632 and is replaced by 'flesh' in 1660/1; they report that they can think of no word in New Testament Greek or in Latin which would make sense in the context; E. Simpson must be right to adopt the later reading (ll. 130 and 297).

31 Acts 20: 28: '. . . the church of God, which he hath purchased with his own blood'; John 14: 2-3: 'I go to prepare a place for you . . .'

32 In what follows I shall quote the 1632 text, but give Potter and Simpson line references; I have normalized long-s.

33 St Augustine offers this interpretation of Psalm 68: 20 (which he numbers Psalm 67, following the Vulgate) in The City of God, XVII, xviii.

34 Gardner, In Defence of the Imagination, 96.

35 See n. 8 above.

36 Donne follows the Geneva Bible gloss in distinguishing the Passover from the Last Supper; not all writers did, e.g. Johan Hiud, The Storie of Stories (1632).

37 See J. MacQueen, Numerology: Theory and Outline History of a Literary Mode (Edinburgh, 1985), 47-9.

38 H. Thurston, SJ, The Stations of the Cross: An Account of their History and Devotional Purpose (London, 1906). The best known early representation of the Stations is the eight carvings of Nuremberg, probably completed before 1490; two English 16th-century lists give seven or fourteen Stations. Several Stations appear to have remained variable into the 18th century: e.g., the Swooning of Mary, the meeting with Simon, with the women of Jerusalem, and with Veronica. An extremely interesting use of a form of the Stations of the Cross for privately constructed (though publicly performed) devotion will be found in Henry Suso's autobiographical The Life of the Servant, trans. J. M. Clark (Cambridge, 1952), ch. 13. I am indebted to Dr Jan Rhodes for telling me that Suso's writings were widely known in recusant circles in England in the early 17th century. David Novarr in The Disinterred Muse (Ithaca and London, 1980) supposed (pp. 163-7) that the Stations of the Cross played a part in the structure of Donne's Devotions.

39 J. A. Burrow, The Ages of Man: A Study in Medieval Writing and Thought (Oxford, 1986).

40 Ed. S. Snyder (Oxford, 1979).

41 John Carey's Oxford Authors edition, 1990, provides an honourable exception.

42 I am indebted to Mr Christopher Shaddock for the information that none of these paragraph word-totals has any particular mathematical significance.

43 'Haec autem propter senarii numeri perfectionem eodem die sexiens repetito sex diebus perfecta narrantur, . . . Numerus quippe senarius primus completur suis partibus, id est sexta sui parte et tertia et dimidia, quae sunt unum et duo et tria, quae in summam ducta sex fiunt' (City of God, XI, xxx).