University of Virginia Library

I

The two 1788 editions of the Triumph are listed in Jacob Blanck's Bibliography of American Literature, II (1957), 520, item 5041, arbitrarily as A and B.[1] In the same order they may be described more fully as follows:

A. THE | TRIUMPH | OF | INFIDELITY: | A | POEM. | PRINTED IN THE WORLD, | M,DCC,LXXXVIII.

• Collation: 8°: a4 B-E4; 20 leaves, pp. [i]-[iii] iv [5] 6-40.
• Contents: p. [i], title (verso blank); pp. [iii]-iv, dedication, beginning "[thick-thin rule] | To Monſ. de Voltaire."; pp. [5]-40, the poem complete, beginning "[thick-thin rule] | THE | TRIUMPH | OF | INFIDELITY. | [flat-diamond rule]".
• Copies examined: Houghton Library; American Antiquarian Society.

B. THE | TRIUMPH | OF | INFIDELITY, | A | POEM. | [flatdiamond rule] | Printed in the World, | M.DCC.LXXXVIII.

• Collation: 12°: i 4 A2 B4 [C]2 (signing A1 as "A2"); 12 leaves, pp. [1]-[5] 6-24.
• Contents: p. [1], title (verso blank); p. [3], dedication, beginning "[thin-thick rule] | To Monſ. de VOLTAIRE," (verso blank); pp. [5]-24, the poem complete, beginning "[line of type ornaments] | THE | TRIUMPH | OF | INFIDELITY. | [flat-diamond rule]".
• Copies examined: Massachusetts Historical Society; American Antiquarian Society.

For each edition the format is in doubt. Neither contains watermarks, and while the vertical chain lines of A suggest an octavo format, one has only the horizontal chain lines and smaller size of B (very roughly two-thirds the size of A) on which to guess duodecimo. For B, Blanck provides "<A>4, A22, B4, <B2>2" — and adds, "Collation in doubt." My own collation is based on the signing ("A2" on the fifth leaf, "B" on the seventh) and on the evidence of conjugate leaves in an imperfect copy, consisting of the first ten leaves, in the Connecticut Historical Society. After snipping some of the stitching, Mr. Thompson R. Harlow, Director of the Society, was able to report the following leaves conjugate: 1 and 4, 2 and 3; 5 and 6; 7 and 10 (in doubt), 8 and 9. The absence of leaves 11 and 12 (only) suggests that they also were a conjugate pair.

The poem has been reprinted twice.[2] A British edition of 1791 follows the text of B, making changes and corrections, and adds Dwight's name for the first time to the title-page.[3] The other reprint, the only recent printing

of the poem, is that in Vernon L. Parrington's The Connecticut Wits (1926), pp. 248-272. Parrington (or James E. Ernst, "who is largely responsible for the text" of the anthology) omits lines 159-176, and claims (p. xlix) to have used "the edition of 1797." Undoubtedly he erred in the date, for he prints in quasi-facsimile one of the 1788 titles at the head of the poem, and his text is a modernized version of A. An edition of 1797 is otherwise unknown. Perhaps Parrington's date represents a momentary confusion of the poem with Dwight's two sermons on The Nature, and Danger, of Infidel Philosophy, delivered in September 1797 and published in the following year.

Were it known that the two editions of 1788 were issued by the same printer, it would be relatively easy to infer a sequence. The fact that the compositor of A ran out of apostrophes just above the middle of page 23 (where five are missing, and spaces left for them, in five consecutive lines), but received a new supply for the lower half of the page, may suggest that the type of A was distributed immediately after each forme was printed; and the smaller format, compression of text, and poorer quality of paper in B suggest that it is a cheap reprint, possibly set up when, unable to reissue A as more copies were demanded, the printer was called upon for a second edition. But nothing about the ornaments, footnote symbols, methods of paging, signing, dating, and the like can be used to show that the editions came from the same house.

One may reasonably infer an authorized manuscript source for one of the editions from the words of an anonymous correspondent to the Hartford American Mercury, 7 April 1803, who, writing in the midst of a newspaper controversy over Dwight's authorship of the poem, asked the editors: "Do you suggest doubts whether Dr. Dwight wrote the Triumph of Infidelity? I know he was the author of it. The man who copied it for the press is not deceased — you may digest these things at your leisure."[4] The correspondent himself may have been "the man who copied it for the press"; at least he had inside information about Dwight's authorization, or he would not have offered for "digestion" the fact of a "copied" text.

The usual possibilities that one edition is a straight (uncorrected) reprint of the other, or that both editions derive from the same manuscript, are ruled out by a consideration of the variants between them — more than 350 in all, mainly accidentals, but also a number of substantives (e.g., "the Lord" / "our Lord" 157 n.; "Cherburg's" / "Herbert's" 200 n.; "or" / "and" 296; "blest" / "new" 405; "lie" / "curse" 464 n.; "even him" / "him e'en" 487; "principle" / "scale" 574 n.).[5] Where accidental variants occur, the text of B, from the reader's point of view, is the better version roughly three times out of four; that is, its spelling is preferable, its punctuation is more reasonable (and more often grammatically correct), and it more frequently

(and more nearly consistently) uses apocopated forms (e.g., "plung'd" 31, "heav'n" 90) to emphasize its metrical regularity. Where substantives occur, the preference for B is somewhat stronger, and in addition B offers clearer hints to the identities of "blank" names in lines 355 ("* * * * * * *" / "E - - - - - -" for "Edwards"), 381 ("* * * * * * *" / "L - - - - - -" for "Ledyard"), and 734 ("* * * * *" / "S - - - -" for "Smith"). Given such variants, if one edition came from an authorized manuscript, the other can have taken its text only from another manuscript, whether or not also authorized, or, if it is the later edition, from an altered copy of the earlier edition. Since it is unlikely that Dwight would have provided an inferior text for a second edition, we must suppose that B, whatever its source of text, is the later of the two editions.

Further evidence may lie in the reasons for suggesting that B was printed from a corrected copy of A. Typographical similarities — the same disposition of words on the title-pages, the use in both editions of the double rule above the heading of the dedicatory epistle to Voltaire, their similarities in title printed at the top of each first page of poetical text (a double rule in one vs. a line of type ornaments in the other; the same disposition of words, followed by flat-diamond rules beneath), their use of italics in lines 345, 573, and 716 — may not in themselves constitute proof that B was printed from a (corrected) copy of A, but one has to suppose a number of coincidences and fairly well detailed manuscript sources to explain them otherwise. More to the point is the fact that while A generally uses asterisks for "blank" names and B generally uses spaced hyphens, they both depart from their systems to print alike "-----" (23 n., for "Christ"), "A---n's" (391 n., for "Allen's"), "C-----'s" (492, for "Chauncy's"), and "J-----n's" (601, for "Johnson's"). The best evidence, however, comes from the variants.

There are, first of all, omissions from B of four words that appear in A: "very" (157 n.), "other" (335 n.), "&c." (411 n.), and "Scriblerus." (671 n.). None is important; but if we apply the general rule that a compositor reprinting an earlier edition may naturally omit words but is not likely to add any, the omissions may suggest, or at least support other evidence, that the printer of B worked from a copy of A rather than from a manuscript. Secondly, there are two uncorrected errors common to both editions: "salute" (402, for the singular verb form "salutes") and "mens" (555, for

"men's").[6] These, like the omissions, may be explainable in other ways, but they are most readily seen, since a number of similar errors in A are corrected in B, as resulting in B from the printer's too faithful following of the text of A. Finally, there are two variants in B that are best explained as new errors resulting from corrections to the text of A. The first is "pow'rs" (44, for the possessive singular form "power's" as in A, or the hypothetical "pow'r's" — since the word is to be read as a single syllable), which clearly came about from the attempt, evident through much of B, to print apocopated forms of two-syllable words pronounced and scanned as one syllable; the possessive ending was lost in the alteration. The second, "were" (278, for "wore"), apparently came about through the misreading (or misprinting) of an alteration of A's "wears" to the past tense.

None of the evidence is conclusive, but taken together it allows the assumption that the supposed later edition B was printed from a corrected copy of A; and it follows, then, that A represents the text of the inferred authorized manuscript. There the matter should rest until new evidence turns up. According to the principles of Greg and Bowers,[7] a proper text of the poem should be based on A rather than B, and should incorporate such variants from B as in each case can best be attributed to the author. One of these variants, though a mere matter of typography, is discussed at the end of the next section.