Icebergs and Icecubes: the Editing of Pratt's "Complete" Poems

E.J. Pratt: Complete Poems, Parts 1 and 2. Edited by Sandra Djwa and R.G. Moyles. Toronto: University of Toronto Press, 1989. Part 1, xi-lvii + 413 pp. Part 2, 504 pp.

If the editors of major collections of poetry remain oblivious to the body of scholarship that relates to their task, or are unwilling to draw upon it, then their editions may be as flawed, incorrect and misleading as E.J. Pratt: Complete Poems, Parts 1 and 2. More specifically, unless editors steer away from the vague, unjustified inconsistencies and imprecisions characterizing this new edition, readers will have every reason to mistrust future editions published by the Univesity of Toronto Press. Regrettably, the "Complete" poems, as edited by Sandra Djwa and R.G. Moyles, is the only Pratt collection currently in print, and readers who must depend on this latest edition for teaching and research deserve to know its failings. The presentation of Pratt's poems in this "Collected" edition, as well as the sections on "Textual Variants and Emendations," suggest that both the editors and the publishers had little interest in, little commitment to, and little experience with editing complex material.

     Before proceeding to more detailed matters, some initial comments are warranted on the title chosen for Djwa and Moyles' edition and on the accuracy, consistency, and validity of the "theory" and practices claimed to govern the editing of this two-volume work. Since no single section of this edition offers a thorough and reasoned argument outlining the editorial theory and methodology applied, a reader in search of such matters must fumble through various portions of Part 1: the "Introduction," page xi, the "Notes on the Text," pages xlix-lii, the short paragraph at the head of the "Textual Variations and Emendations" section, page 339, and the occasional comment jammed into the "Introduction." Only by thus flipping to and fro in the volume can a reader begin to disentangle the editors' claims from their achievements.

     Many of the editors' unjustified statements seem to be made to convince readers that this new two-volume collection is, somehow, the "definitive reading text" (Part 1, p. xi). "We have lacked a complete text" (Part 1, p. xi), they tell us, and this two volume work, they imply, now fills that void. Although they correctly point out that all earlier "collected poems" are merely "selected poems" (Part 1, p. xi), Djwa and Moyles compound the traditional error by offering in their title the misleading phrase: "Complete Poems." A more appropriate title might be "Complete Published Poems and Selected Unpublished Poems." On page "li," (forty pages into the text) the editors themselves confess that only a "selection of Pratt's unpublished poetry" is included.

     The erroneous emphasis on "completeness" is especially puzzling because the editors do not offer even a hint of how many poems have been omitted, or why. Moreover, without the rather thesis-ridden "Introduction" and the overly intrusive "Annotations" sections, there would have been more than enough room to include the poems they have left out. My point is simply that a collected poems should use every page for the writer's own material, not for the editors' preferred interpretations of the author's achievement, nor for their overly long lists of annotations. Admittedly, the section of "annotations" might have had value, had they been Pratt's own — and Pratt's were the ones a reader might well have expected to find — but his were ignored and omitted. Even the few annotations that Pratt included as "footnotes" in many first editions and in the Collected Poems of 1958, which is the copy-text for much of the material in the "Complete" poems, the editors silently delete. Djwa and Moyles, then, a) neither offer explanations nor justifications for leaving the edition in-complete; b) nor do they explain why their annotations, rather than Pratt's, should be preferred. In a text which purports to offer the "author's final intentions" (p. xlix) — would it not be logical to include, clearly and unambiguously, Pratt's annotations rather than those of Djwa, Robert Gibbs and David Savage?

     The issue of whose annotations should dominate the book is only one important consideration. Another is the omission of some of Pratt's unpublished and occasional poems. Djwa and Moyles assert that poems excluded by Pratt from earlier editions cannot be considered a "part of his serious corpus" (Part 2, p. 293). Some clarification or elucidation of the adjective "serious" might assist readers in deciding whether Pratt's work has been fully and judiciously presented. For the Christmas season, for example, Pratt — the same Pratt who asked his wife to translate "Hegel 'on the idiocy of believing in Christianity'" (Part 1, p. xvi) — wrote religious poems that he had privately printed on cards to send to friends. Here is an example of such an untitled Christmas poem (I cite my own copy, but one is also available in the Douglas Archives at Queen's University).

            Grant us, Lord, amidst the Carols
            And the chimes that hail Thy birth,
            Deeper anthems from the nations
For Thy "Peace on earth."

            Give us now the real fulfilment
            Of the Father's message when
            The Gloria rang out in Starlight
His "Good-will to men."

            Visit those who are in hunger
            Where the tables are not spread,
            Millions wait beside Thy cupboards
For their daily bread.

Such "unknown" poems are important for establishing that Pratt was capable of writing in different modes and from different perspectives. By excluding such poems from the "serious corpus," the editors prevent readers from gaining a full awareness of the differences between Pratt's public and private writings. But to return to the main point: had the editors of this latest collection at least explained why only "some" unpublished poems were included, and why Pratt's annotations were excluded, a reader would have a basis for forming an opinion about the soundness of the editorial decisions.

     The misleading title of the Complete Poems and the curious, unexplained selection of "Annotations" may be considered mere icecubes that hint at the icebergs to come. These may be addressed in point form:

1) Unlike Lila Laakso, who in skilfully editing the "Descriptive Bibliography" in volume two of the "Complete" Pratt, cites Fredson Bowers (Part 2, p. 373) as the basis for her methodology, and sticks to these guidelines, Djwa and Moyles state only that they are following "modern editorial theory" (Part 1, p. xlix). They offer no further explanations. Possibly one could blame the blunders, errors, mistakes and omissions on the accurate application of inadequate theory but to do so, one has to know which theory so misguided them.

2) Djwa and Moyles claim that "all . . . authoritative" texts have been "collated" (Part 1, p. xliv) and that "all substantive variants" have been noted, yet they leave out major "substantive" variants, and further complicate matters by offering the wrong "variant" as the "copy" text. For example, although a number of variant passages from Brébeuf and His Brethren are utterly overlooked by Djwa and Moyles, one omission indicates that something is seriously wrong with the version they offer. Have the editors actually used the claimed copy text? In the "Complete" poems, as in the first edition, lines 1928-1929 of Brébeuf and His Brethren read:

                                          Grant that
            I may so bear it as to merit Thy grace.1

Pratt, as Carl Klinck long ago pointed out, learned that a "Catholic" would never use the word "merit" in this context, and so changed the word to "win" in later editions.2 The copy-text that Djwa and Moyles claim to use, Collected Poems, 1958 (see p. 255, lines 1928-1929), uses the word "win," not "merit":

                                          Grant that
  I may so bear it as to win Thy grace.
                                          (11. 1928-1929, p. 292)

In the new "Complete" poems, however, the word "merit" (line 1929) has been inexplicably restored. Djwa and Moles either did not use the 1958 Collected Poems as their "copy text," despite their claims (p. 1), or they silently made editorial changes without acknowledgement, or they somehow "mixed" texts.

     The "merit"/"win" example is also disturbing because it points to the fact that not once in the "Complete" poems is Klinck's work referred to, nor, indeed, is any other scholar's listing of variants. Books, articles and theses that have noted numerous variations in Pratt's poems, listed them, and, at times, explained them correctly are, inexplicably, never used. "Modern editorial theory," it seems, entails the blanket dismissal of all previous scholarship in the field.

3) Djwa and Moyles claim that "all substantive variants in all other versions" of Pratt's poems (p. 339) have been listed, yet, they miss variants both obvious and subtle. Here are only a few serious omissions from poems initially included in Newfoundland Verse and later reprinted in Collected Poems 1958.)

     a) In "The Flood-Tide," in both Newfoundland Verse3 and Collected Poems, 1958, every second line is indented. There are no indentations in the "Complete" poems, Part 1, p. 53. (In fact, every narrative poem in these "Complete" poems has faulty indentation, false stanza breaks, and missing stanza breaks. Furthermore, numerous lyrics lack proper line indentation and some have joined stanzas because of omitted stanza breaks.)

     b) "The Ground-Swell" has a hyphen in Newfoundland Verse, one which is dropped in the Collected Poems, 1958. There is no comment about this in the "Complete" poems.

     c) The last line of "In Lantern Light," in both Newfoundland Verse and Collected Poems , 1958, reads: "It marble white." In the "Complete" poems, the line incorrectly reads "Its marble white" (Part 1, p. 66).

     d) In "On the Shore" of Newfoundland Verse and Collected Poems, 1958 the repeated lines 4 and 11 read: "Come home! he will not hear your call." In the "Complete" poems, the lines suddenly read: "Come home! he will not hear you [sic] call" (Part 1, p. 109).

     e) "Overheard in a Cove," Newfoundland Verse, lines 88-89, read:

With every age, the race. Take medicine,
And note its triumphs. How shall I begin

The "Complete" poems omit to note that Pratt revised these lines for the Collected Poems, 1958:

As knowledge grows from more to more, and stronger,
With every age, the race. How shall I begin

     f) Line 21 of "The Sea Cathedral," in both Newfoundland Verse and Collected Poems, 1958, reads:

It passed on southwards slowly to its fate

Without explanation, the "Complete" poems offers:

It passed on southwards to its fate

(These examples are sufficient for present purposes, but it must be stated that I have 80 pages of commentary on errors, omissions, and oversights in the "Complete" poems in my computer, and that these are based only on my own comparisons of texts. Any publisher interested in presenting a fuller version of "Textual Variants and Emendations" is welcome to contact me.)

4) Djwa and Moyles state that if a title is not included in the "Textual Variants and Emendations" sections of the "Complete" poems, "the reader is to understand that no variants exist" (Part 1, p. 339). Even a cursory glance at the case of Brébeuf and His Brethren, where several variants are omitted (in fact, only the ending is listed as a variant [Part 2, 255-256]) reveals that Djwa and Moyles claim a completeness which simply is not there. To save space, it should be noted that examples "a", "c", and "e" above are not included in the sections on "Textual Variants and Emendations." The list of all omissions is long. Furthermore, even the entries that Djwa and Moyles include are at times incorrect or misleading. For example, the full title of the section of "Rachel" included in Newfoundland Verse is "The Conclusion of 'Rachel' (A Sea Story): In Memory of R. S. Le D"; the variant line 420 listed on page 342, "Complete" poems, Part 1, should read "Broke the dead silence . . ." not "Broke the thick silence . . ."; the listed lines of 420-427 include lines never revised (line 420, alone, should have been cited).

5) Djwa and Moyles claim that the "Complete" poems is suitable for "the general reader" and the "specialist." The approach does a disservice to both audiences because the differences between them are insurmountable. "General readers," I assume, are those readers who must rely on the "Complete" poems for accurate transcriptions. But such readers are ill- served by the fact that in the "Complete" poems numerous poems are transcribed incorrectly. "Comrades" has false wording in line 12; "A Prairie Sunset" does not follow the "copy text"; The Roosevelt and the Antinoe and "The Big Fellow" offer incorrect stanza breaks; The Last Spike omits the subtitle from the main body of the text; and finally, The Iron Door contains three obvious proofreading errors. Djwa and Moles often not only repeat the errors of the Collected Poems, 1958 but also add to them.

     A "general" reader can only be confused with the result. The example of The Iron Door illustrates the problems. The well known-lines

                  A giant hand
Had wrought it cruciform
(lines 7-8, Collected Poems, 1958)

are mistranscribed as

                   A giant hand
Had wrought its cruciform.
(lines 7-8, "Complete" poems)

Line 218, which should correctly read "And all the music left upon thy waters" is altered to ". . . the waters"; line 265, which should read "With an insatiate hunger, to discover" loses the word "an" in the "Complete" poems. Clearly the "general" reader does not fare well when even a major poem such as The Iron Door is not accurately presented.

     What of the specialist? I assume that Djwa and Moyles are referring to their "Textual Variants and Emendations Section" as one of the features that might earn the praise of this group. My conjecture may be justified, particularly since Djwa and Moyles emphasize that Laaska's "Descriptive Bibliography" (included at the end of the second volume) makes available to them the "multiplicity of versions" which demands skill and thereby "delights and challenges a modern editor" (p. 1). They accentuate their point by stressing that "The Sea-Gull" [sic.] (Part 1, p. 1) itself has as many as "eleven versions" (Part 1, p. 1). I insert the "sic" here because I believe the editors are referring to "Sea-Gulls." (Possibly "The Sea-Gull" is a variant title, but the editors fail to explain that in their "Textual Variants and Emendations.") More significantly, the "eleven versions" of "Sea Gulls" that delight the editors, as their own section on variants and emendations confirms, has only one change: the indefinite article "the" is changed to "a". The specialist can only wonder if these editors have confused reprintings with versions.

     How far do the shortcomings indicated by the handling of such poems as "Sea-Gulls" and Brébeuf and His Brethren extend into the "Complete" poems? The transcription of an extensive section of a typescript of The Witches' Brew, housed in the Douglas Library at Queen's University, offers the specialist the opportunity to answer. Not only is the section of The Witches' Brew typescript incompletely recorded in the "Complete" poems, but the editors fail to mention that there are at least two "clean" typescripts that vary in important ways (only one of the copies, for example, contains footnotes for numerous foreign words and phrases). It could also be added that the editors' terminology on page 348 of Part 1 becomes confusing. Does the phrase "holograph manuscript" simply mean typescript? I was not able to find a hand-written version of The Witches'Brew in the Queen's archives, but then the lack of information in the "Complete" poems often hinders the researchers from finding specific materials. Even if it could be pretended that this were not the case, a specialist has to point out that some of Pratt's revisions have slipped through the net of the "two" unnamed "independent collators" (Part 1, p. l).4 Here is one section omitted from the typescript housed in the Douglas Library of the Queen's Archives — the "holograph" that the editors appear to have used:

And when the zest was on thy slew
Yet every fish that swam the seas
Had most ideal obsequies;
When aged or sick both foe and friend
Would take him to his timely end,
For he by day and night was followed
By tender gulps till he was swallowed.
Even the most tempestuous quarrels
Were never prejudiced by morals.

Speaking as a "specialist," I can add that the lengthy sections of The Witches' Brew that are transcribed in the "Complete" poems are also incomplete. Two examples must suffice: directly after the "Inventory of Hades" (Part 1, p. 139) should follow the instruction "(To be read, as items of stock, with clerical monotony)"; and immediately preceding the section listed in relation to lines 281-2 (Part 1, page 348) should follow the heading "The Disaffection of Hades I with reasons for the Exodus."

6) The editors simply state that they selected some "interesting variants from manuscript sources" (Part 1, p. 338). But on what grounds they elected one variant as more "interesting" than another is never explained. Moreover, why only a few poems receive such extra attention is never clarified. To a "specialist," such a lackadasical attitude or editorial policy in relation to original typescripts is disturbing, to say the least. Arbitrarily dipping into some typescripts while ignoring others may mislead readers into believing that only the typescripts considered have enough substantial material to make the effort worthwhile. The fact is that Pratt's "clean" typescripts — typescripts that he sent to friends or editors with only a few minor pencil corrections — are utterly fascinating.

     Again The Iron Door furnishes an example. Pratt sent a typed copy of the poem to Lorne Pierce, and in an accompanying letter explained why and how certain lines of the poem were linked directly to some of Pierce's own views and comments. Pratt's illuminating comments could easily have been incorporated explicitly into the annotations of the complete "Complete" poems. More importantly, an examination of the typescript sent to Pierce reveals that Pratt had not mailed the "final version" of the poem used for the first edition. The following section of the typescript (roughly the lines of 195-210 of the final published text) reveal an interesting and extended variant:

Why from her love should issue hate?
And every claim to mercy be denied,
With justice humbled for its pride
Before the Eternal gate;
And why her only-born
Should have by some ancestral strain
Joined to the death-writ of the law been torn
So fiercely from her side,
Were quierries rained upon the iron plate;
Not that her gift to life should be returned
To death, but that the Fates should so conspire
To have the one devoted offering burned
At such an altar, and such a fire. . . .
But what availed
Those finger raps against the arrest
Of life when every rubric paled
Before the Theban mockery of the crest.

Had the Djwa and Moyles' rationale been to look at the "clean" typescripts submitted to friends and publishers, and had that policy been consistently applied, then the "specialist" might be content with the "Collected" poems. Or, had these editors clearly called for a need to produce an independent book to deal with such material, then their case might be solid. However, the arbitrary manner of padding the variants section with references to only some typescripts while overlooking others — and at the same time leaving gaps and omissions in their present task of dealing with all published versions — is, bluntly put, disappointing and unsatisfactory.

     The "Index of Titles in Parts 1 and 2" indicates just how unsatisfactorily these books have been prepared. A poem such as "The Doctor in the Boat" is not listed — why? When poems are listed, there is no cross referencing of material: readers can only find out if there were emendations or variants by skimming through pages 338-364 (Part 1) and pages 253-261 (Part 2). If readers forget to keep a bookmark beside the poem considered in the body of the book, there is no way of easily going back to the body of the book no page references are provided. In a book crowded with irrelevant and distracting material, the lack of cross-referencing is particularly both ersome. Moreover, when solid information is provided, and usually only by Laaska, readers have a very difficult time finding it. For example, a check of the index of titles to find a sequence of poems by Pratt called "Flashlights and Echoes From the Years of 1914 and 1915" will only produce a blank. Why? There are three reasons. First, the editors have decided to put all the poems into chronological order. Any effort by Pratt to integrate or unify a series of poems is ignored, overlooked, and unnoted in the textual variants. Second, the authors have elected to use the last version of a poem in conjunction with the earliest date of its supposed publication. The late version receives the authority of the date of an early version which, at times, is neither printed nor correctly "emended" by Djwa and Moyles. By placing the date of the "copy text" beside the date of the first publication, the editors might have realized that any reader familiar with an original title of a poem will not be able to trace it through the index. The little poem entitled "?" is in the "Complete" poems, for example, but a readers will have to do a lot of turning of pages before they will find it. Third, although the editors pride themselves on recognizing that Pratt's editions are so rare and limited that many readers may not have access to them (Part 1, p. xi), they make no effort to consider what should be done in a "complete" poems to help readers recognize that an "Index of Titles" might not be able to list all the titles Pratt used in his various books. Only readers patient enough to dig through the final section of Part 2 will discover that "Flashlights and Echoes 1914 From the Years of 1914 and 1915" was a series of twelve poems (which only Laaska lists), or that Pratt once envisioned a series of "Monologues and Dialogues," or that he once even gave thought to "Creatures of Another Country." Djwa and Moyles do not seem to have thought about what to do with revisions that range beyond the rewriting of lines.

     Finally, the editors claim silent corrections and a hands-off editorial policy. This is, to say the least, a debatable policy. By offering incomplete and skimpy notes, while at the same time inundating the reader with excessive sections like the "Introduction" and the "Annotations" they skirt the issue of tampering with the text. An explicit claim is made that only slight changes in punctuation have been made (Part 1, p. 339), yet nearly every poem included in the "Complete" poems has had a long, complicated history of changes in punctuation. The editors claim, for example, that by relying on the privately published Rachel of 1917 as the "copy-text," the editors needed to make only "minor punctuation changes (of which there are few)" (p. 339). However, a comparison of the "copy text" and the version included in the "Complete" poems reveals that a total of 18 "comma-dashes" and "semicolon-dashes" have been altered. What is a "comma-dash" or a "semicolon-dash"? Before explaining my terms, let me remind readers that the trend of "modern" editors is to eliminate and eradicate all idiosyncratic features of "accidentals" — even if the poet used them in his own typescripts and his books. Such wholesale alterations of accidentals may be the norm, but the convention simply helps to eradicate a poet's habits of controlling his poems. (Richard J. Finneran's Editing Yeats's Poetry is an excellent study that discusses such "minor" issues, as well as a few "major" ones.) Pratt, in his typescripts and his books, often used ",—" and ";—" and ":—" and "—." At least a note explaining these "comma-dashes," and other conventions, would help general readers understand why, in the "Complete" poems, the punctuation in the body of many lines uses a ",—" or ";—" or":—" while most end-lines use "—." The ends of the lines, more often than not, have been modernized while the internal punctuation often remains as in the original. Even the moving of a question mark, from line 508 (Rachel 1917) to line 513 ("Complete" poems) might be justifiable — depending on how one "reads" these lines — but making the change without noting it suggests that the accidentals of punctuation have no effect on how to read a poem. When long poems are not in question, short ones are. The puzzled reader of "Sea-Gulls" may theorize about why the comma at the end of the first line was omitted, but the reader familiar with the Collected Poems of 1958 will remember that a comma was at the end of the first line, though it is omitted in the "Complete" poems version. In other words, the problem with silent correction is that the reader has to have confidence that the text has been proofread thoroughly and scrupulously.

     Moyles and Djwa had "intended to provide a definitive reading text" (Part 1, p. xi), but — alas, the old cliché about "good intentions" holds true. The E.J. Pratt: Complete Poems, Parts 1 and 2 is not an edition upon which the specialist or the general reader can rely.


  1. The Collected Poems of E.J. Pratt, second edition. Edited with and Introduction by Northrop Frye. Toronto: Macmillan, 1962.[back]

  2. Wells, Henry W. and Carl F. Klinck, Edwin J. Pratt: The Man and His Poetry. Toronto: Ryerson, 1947, p. 54.[back]

  3. E.J. Pratt. Newfoundland Verse. Toronto: Ryerson, 1923.[back]

  4. I would like to acknowledge the staffs of the Pratt Library, Victoria University, Toronto and the Douglas Library, Queen's University for allowing me to doublecheck my com parisons of texts and typescripts.[back]

    Ed Jewinski