University of Virginia Library

Search this document 


expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
collapse section 
Quadruple Imposition: An Account of Goldsmith's Traveller by William B. Todd
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 

expand section 


Page 103

Quadruple Imposition: An Account of Goldsmith's Traveller
William B. Todd

THE EARLIEST version of Goldsmith's first poem, The Traveller, appears in a series of quarto half-sheets discovered and reproduced by Bertram Dobell in 1902 and now deposited under press C.58.g.7 at the British Museum. Three of these unique sheets (B--D) are leaded between lines, spaced between paragraphs, numbered throughout, and headed A PROSPECT OF SOCIETY, a reference later used as the sub-title of the published edition. The fourth (E) contains only text, signature, and catchwords, all set solid at the top of each page, but of an appropriate number of lines, in each instance, so that the letter when spaced and headed would exactly fill the page.

Whereas the peculiar form of E is not without significance, it has been generally disregarded in the light of an even more remarkable aberration occurring throughout the text. Instead of running forward, this regresses in nine sections, commencing at B1r-2r with lines 351-392 and concluding on E2v with lines 73-90.[1] The orderly derangement of the sections was first noticed by A. T. Quiller-Couch who, in his review of Dobell's reprint, thereupon offered a plausible explanation.

As Goldsmith finishing writing out each page of his poem for press he laid it aside, on top of the page preceding . . . and, when all was done, he forgot . . . to sort


Page 104
back the pages in reverse order. That is all; given a good stolid compositor with no desire but to do his duty with the manuscript as it reached him, you have . . . an immortal poem printed wrong-end-foremost page by page. And I call the result delightful, and, when you come to think of it, just the blunder so natural to Goldsmith as to be almost postulable.[2]

Though I agree with Quiller-Couch that his hypothesis is "ridiculously simple" much simpler than any I may propose-further reflection leads me to doubt Williams' belief that it offers "the true explanation" and to deny Dobell's assertion that it "accounts for all the peculiarities of the first version."[3] One considerable oversight involves lines 1-72, 91-100, and sixteen lines after 392, all present in The Traveller, all omitted here. Probably the last sixteen were added in revision, but the earlier lines, or something equivalent to them, must have been prepared for the original printing. Another matter casually dismissed as "fairly regular" is the unequal length of the sections in this version, a length ranging (in order of writing) from 18 lines in the first group, to 28 in the next, to an average of 36 in the next five, to a maximum of 42 in the last two. If each of these represents a page of copy, then we must suppose that Goldsmith wrote (1) in ever smaller script on paper of equal size, (2) in even script on larger and larger paper, or (3) in even script on paper of uniform size but more and more on one side of the page and then more and more on the other. None of these inconsistencies is observed in Goldsmith's practice.[4] A third problem left unresolved concerns the extraordinary imposition of sheet E. As this apparently anticipates the later arrangement evident in B-D it must reflect an


Page 105
earlier operation in press-work and may not, therefore, be envisaged in any "simple" explanation. Whatever the status of the one sheet, it differs from that of the others.

Another and more promising theory is the one originating with A. E. Case and briefly noted by R. S. Crane.[5] This suggests that the disarrangement results from an error in make-up-a hypothesis which, as Professor Crane has since explained to me, presumes the use of galleys. Until recently it was believed that long galleys were introduced not before the nineteenth century,[6] but Mr. Philip Gaskell has now shown that as early as 1773 William Mason was reading "Folio Proofs" for the quarto volume of Gray's Poems published two years later at York.[7] If this practice was common among provincial printers in 1773 it probably was not unknown to London printers in 1764. Like Mason, then, Goldsmith may also have read "Folio Proofs"-or something like them-revised these for press and, conceivably, approved the ordered "Folio" version before it was disordered in quarto make-up.

Before venturing too far with this last conjecture, we should first deter- mine the kind and number of proofs submitted to the author. Since the layout of sheet E indicates, as already mentioned, an earlier stage in the printing, it is likely that the earliest proofs were, like this, set solid, ordinarily contained 36 lines each (the average length of the sections as later imposed), and therefore measured, at 6mm. a solid line, a total of 216mm. Such proofs, though comprising almost twice the letter of a spaced quarto page, would extend only 6mm. beyond the total measure of that page (i.e., letter plus heading and direction) and still allow a 79mm. margin for any additions supplied in revision.[8] From their limited extent these proofs would appear to be, not "slips" or "folios" in the strict sense of these terms, but short quarto galleys set in frames sufficiently large to accommodate single page work and sufficiently thin to be machined without any special adjustment of the press.[9]


Page 106

If the proof was originally of this kind it probably represented, in more equal portions, more lines of text than now appear in the quarto set-up, either of the E or B-D impositions. These printed sheets exhibit eight full sections and part of another. Of these, the last full section extends from line 101 at the middle of E1v to line 128 at the foot of E2r. Then on E2v, the last page of the series, the part section begins with line 73 and ends with line 90. Below 90 is the catchword "Each", an obvious indication that the remainder of this section, 91-100 (corresponding to 93-102 in The Traveller), was in galley and intended for imposition in 1r of the unprinted sheet F. The complete section, then, like the one that immediately precedes it, consists of 28 lines. Besides 91-100 the extant printing also omits lines 1-72, an exact multiple of 36, and therefore probably of two standard sections designated for the remainder of F. In inverse order of printing, and with due allowance for spacing, the first of these would range from line 37 at the middle of F1r to 72 at the top of 2r, the second from line 1 at the top of 2r to 36 at the foot of 2v.

With F reassembled on the pattern of E the sequence is complete, and all sections may now be reversed to show the original arrangement in galleys. As illustrated in the accompanying table, there were apparently eleven of these, all correctly ordered and presenting a consecutive text. Of the eleven four depart from average, two as a pair of 28 lines each, two as another pair of 42. Very probably the shorter pair was originally composed



Page 107


Page 108
as a single galley of 36 lines to which the author then added 20 more, far too many to be held in one frame and therefore evenly divided in two. Possibly the larger pair represents a similar distribution; or it may be that the compositor, realizing that he had only 84 lines to go, set his letter in two equal sections instead of running it out to three. In either event, though longer (folio?) frames would be required for this pair, each of the proofs could still be accommodated-with 33mm. to spare-on a separate quarto leaf.

So far, the presumption that all type was originally set in galleys has resolved two of the three questions undecided in the Quiller-Couch hypothesis. For one I surmise that the unprinted lines appeared in the initial proof, and there according to an evident pattern. For the second I argue that occasional deviations from pattern imply some mechanical division easily devised in composition but hardly to be imagined in manuscript. For the third question, on the relation of E to the preceding sheets, I defer comment until I describe certain other unnoticed peculiarities. All of these are best considered in a chronological review of the work in the printing shop.

Goldsmith's copy, when delivered to the printer, was split at the end of the paragraph at line 202, approximately midway in the text, and assigned to two compositors, A and B, who then set galleys 1-6 and 7-10 respectively. From the evidence of the spelling it would seem that It, the last galley, though of that portion which normally would have gone to B, was partially set by A or eventually finished by a third compositor C.[10]

'A' lines 1-202   'B' lines 203-350   'C'? lines 351-392  
chear [13], 181, 189   cheer 217, 271  
splendour [45], 86   splendor 373  
honor 86   honour 252, 253, 304, 342 honor 352  
'Till [93]   Till 220, 260, 347  
This tabulation also shows that the unprinted lines-now evident only as finally imposed in The Traveller-were originally set by the compositor to whom they properly belong. Hence we discover still another indication of the previous existence of these lines in the original imposition.

As the type was set the galleys were printed, rinsed, and then stored in a rack to await Goldsmith's emendation; and when this was received they were again processed and again stored, we may conjecture, to await his final approval of the text. Not on the first round but apparently on the last, the galleys were accidentally reversed so that, as they lay in the rack pending


Page 109
the second imposition, the eleventh galley was in the position of the first.[11]

If the author read the original series of proofs, the second sequence in quarto make-up must have been prepared for the corrector. Of this sequence only E is extant, the one sheet set solid, like the earlier proof, but now disposed in pages so that the corrector might check layout and punctuation. Both in E and in B-D (now evident only in the third imposition) the corrector had much to do, for as the galleys were inversely disposed he soon discovered that the final line of one occasionally bore no grammatical relation to the initial line of the next.[12] First in galley 9 and thereafter in 7 and 4 he accordingly adjusted the pointing to join the disparate parts. As determined from the correct pointing again introduced in the reimposed Traveller, each of these galleys originally began with a new paragraph (at lines 275, 203, and 101 respectively), each following in a preceding galley another logically complete in itself. In inverted order, however, these paragraphs interrupted others uncompleted in the preceding sections, and were therefore joined to them.

  • Page / Galley / Line
  • C1r 10 350 And statesmen toil, and poets pant for fame;
    9 27 To men of other minds my fancy flies,
    276 Embosom'd in the deep where Holland lies,
  • D1r 8 274 The solid transports of internal praise.
    7 203 These are the charms to barren states assign'd;
    204 Their wants are few, their wishes all confin'd.
  • E1v 5 166 But man and steel, the soldier and his sword.
    4 101 Far to the right, where Appenine ascends,
    102 Bright as the summer, Italy extends;
Besides the three I have cited there are many other alterations,[13] but as all of these occur within a section it is impossible to decide whether the previous


Page 110
readings originated in the author's manuscript, or later in the compositor's galley, or later still in the corrector's page proof. Only the three, then, directly implicate the corrector at points where only he would intervene* where he must intervene to construct a continuous text; and as these three are conveniently dispersed one to a sheet, they allow us the inference not only that the text was originally approved but, in this second imposition, that it was eventually "corrected" for press.

If, as we have now demonstrated, proof sheet E is of an intermediate form used by the corrector, then sheets B, C, and D, all obviously of a third imposition, must represent finished work at press. In every respect these three sheets appear to be in their final state: in leading and spacing, in the headtitle and initial letter provided for the first page, and in the numbers and headlines supplied for all succeeding pages.[14] What is so apparent I therefore accept as the last and most evident fact that all sheets, in one form or another, were destined for issue. From every indication A Prospect of Society should thus be described, not as a random essay, nor as an uncorrected proof, but as an abortive edition, well conceived, badly bungled, and finally suppressed. All that remains of it now are these several sheets, in several stages of preparation.

Once this blunder was discovered the letter was transposed, revised, sup- plemented by several couplets and a new concluding paragraph-this last largely from the pen of Samuel Johnson[15] -and the whole again impressed, in its fourth imposition, for issue as The Traveller, Or a Prospect of Society. Though the text of this remains invariant, in the first published edition, the preliminaries appear in four different states, two without and two with half-titles. As these also have been erroneously described, our account may properly conclude with a record of the last of many extraordinary efforts to bring forth Goldsmith's earliest poem.

Variants in First Edition

1st state. No half-title.

THE | TRAVELLER, | OR A | PROSPECT OF SOCIETY. | A | POEM. | [double rule] | LONDON: | Printed for J. NEWBERY, in St. Paul's Church-yard. | MDCCLXIV.


Page 111
Medium quarto in half-sheets. A2 B-F2 G1. 13 leaves: [1] title, [2] blank, [3] dedication 'THIS | POEM | IS INSCRIBED TO THE | REV. HENRY GOLDSMITH, M.A. | BY | HIS MOST AFFECTIONATE BROTHER, | OLIVER GOLDSMITH', [4] blank; 1 2-22 text.

Contrary to England's supposition (op. cit., p. 331, n.1) this state does not present a text intermediate between that of the 1764 Prospect and 1765 Traveller versions. Text type is identical in all Traveller states and retains, at 5.16, the single misprint 'patrioric'.

Copy examined [photostat]: CSmH (Locker Lampson-Van Antwerp-Halsey). One other, the Rowfant copy, is said to be of this state. 2d state. No half-title.

THE | TRAVELLER, | [&c., as in 1st state, except] POEM. | INSCRIBED TO THE | REV. HENRY GOLDSMITH. | BY | OLIVER GOLDSMITH, M.B. | [double rule] .... π1(=G2?) a2 B-F2 G2 (-G2). 14 leaves: [1] title, [2] blank; i ii-iv dedication in six paragraphs; 1 2-22 text.

Copy examined: NN (Berg Collection). Apparently a unique exemplar of this state.

3d state. Half-title: THE | TRAVELLER, | OR, A | PROSPECT OF SOCIETY. | Price One Shilling and Six-pence. THE | TRAVELLER, | [&c., as in 2d state, except] REV. MR. HENRY GOLDSMITH. | ... | MDCCLXV. π2 a2 B-G2. 26 leaves: [1] half-title, [2] blank, [3] title, [4] blank; i ii-iv dedication; 1 2-22 text, 23-24 advertisements for 11 books.

Copies examined: NjP, NN (Berg-lacks final leaf). 4th state. Half-title: THE | TRAVELLER, | OR A | PROSPECT OF SOCIETY. | [Price One Shilling and Six Pence.]

Description and collation as for 3d state. Half-title is corrected to agree with reading of title page.

Copies examined: MH, NN (Berg-3 copies), ViU



All line references are to the original version as reconstructed in my edition of a Prospect of Society, Water Lane Press, Cambridge, 1954. In this lines l-80 correspond to those in The Traveller, lines 81-370 to 83-372, 371-374 to 375-378, 375-392 to 383-400. Traveller lines 81-82, 373-374, 379-382 and, presumably, 40l-416 were added in revision. Readers using Dobell's facsimile edition of the two texts should correct the following misprints.

[Dobell line, reading]  [text line, original copy]  
306 honour   86 honor  
265 ascends   101 ascends,  
232 querry   134 quarry  
235 late,   137 late  
205 Cheerful   181 Chearful  
224 breast,   200 breast;  
173 coarsely   221 coarsly  
85 skies:   233 skies,  
125 days,   245 days  
99 profession   289 profusion  
106 brings.   296 brings,  
45 wing   311 wing,  
66 annoy.   332 annoy:  
14 home,   364 home;  
17 tyrant   367 coward  
42 mine,   392 mine.  
[Traveller readings] 
61 rest   rest,  
82 patriotic   patrioric  
157 And   As  
223 flow;   flow:  
224 low,   low.  
354 unhonour'd   unhonor'd  


Daily News, March 31, 1902, as cited in appendix of later issues of Dobell's edition. The same quotation, with supporting testimony, is offered by George England in The Library, n.s., III (1902), 327-332.


Dobell, op. cit., appendix; I. A. Williams, Seven XVIIIth Century Bibliographies (1924), pp. 133-134.


For manuscripts of any length (Goldsmith ordinarily wrote on foolscap paper in a hand averaging (with spacing) 8mm. to a line. One contemporaneous with the Prospect (CSmH, Preface and Introduction to the History of the Seven Years' War c.1761) is a folio in twos, written throughout on both sides of the leaf, and containing about 37 lines to the page. Another of the same period (NNP, The Captivity, c.1764) is in quarto half-sheets, written throughout on one side only, and averaging 25 lines to the page. For his correspondence, on the other hand (as judged by 1 letter at NNP, 4 at MH), Goldsmith consistently used post quarto half-sheets.


CBEL, II, 641-42.


McKerrow (p. 66) cites as the earliest reference Savage's Dictionary of Printing, l841; but earlier reports, in this century, appear in (1) Johnson's Typographia, 1824, II, 193, 587; (2) the Inventory of the Glasgow University Printing Office, 1826 (reproduced by the Water Lane Press, Cambridge, 1953), which accounts for "51 Mahogany Galleys-32 Fir Galleys"; the 1832 report of the Union Society of Compositors (Ellic Howe, The London Compositor [1947], p. 195), which refers to the pulling and correcting of "slips and galleys".


The First Editions of William Mason, Cambridge, 195l, p. 18. Even earlier than this-but possibly of little significance for English printers-is a reference to galley proof for the 12 "Neaulme, Amsterdam" [Duchesne, Paris] edition of Rousseau's Emile, 1762. See P. H. Muir's account in The Book Collector, I (1952), 68-69.


An uncut copy of The Traveller in the Berg Collection measures 295 x 228mm.


"Slips", as defined by Savage (cited in Howe, op. cit., p. 177), are of "the length and half the breadth of a demy leaf of paper," i.e., of half-sheets divided lengthwise in strips measuring 451 x 279 mm. These presumably are the "folios" to which Mason refers, a series of proofs necessarily longer than those required for the Goldsmith quarto (since Mason's work remained unleaded), and struck from long, thick wooden galleys (cf. fn. 6) on a press especially reserved for this purpose. As Mr. Gaskell reminds me, the use of such galleys in a press set for ordinary printing would be a great nuisance, requiring first a considerable adjustment of the platen cheek mortices and headbolts, and then a readjustment of all these before normal work could proceed.


Brackets enclose readings unprinted in the Prospect but evident in The Traveller. To simplify this account I enter only the stem of the word, ignoring prefixes and the like.


For other probable reversals, not of galleys but of formes, see my account in Studies in Bibliography, III (1950), 184-191. In these examples, of course, the text remains unaffected, whatever the order of printing.


If as Dr. Walker has shown (Studies in Bibliography, VI [1953], 58), a corrector was occasionally concerned only that letters should make a word and not that words should make sense, we may excuse (indeed, applaud!) the example of this corrector in leaving to the author the meaning of his lines and drawing to himself only the responsibility of pointing the text. Even in this distorted form the lines make sense, of a kind-as much certainly as the Shakespearian motley constructed by Harvey Breit or the modern mess compounded by Lord Edward John Moreton Drax Plunkett Dunsany. See NYTBR, January 24, 1954, p. 8.


Some 22 in fact. See notes to my edition.


Unfortunately press figures and skeletons, the two most obvious signs of successive impression, are not used either in this printing or in the first published edition.


In a copy of the poem which Boswell tendered him in 1783 (a pirated '1770' edition: see my note under 'C', Studies in Bibliography, VI [1953], 41), Johnson marked nine lines as all that he could recollect as being his. These, in the first edition, are 398, 407-l2, 415-116. Nonetheless, Johnson told Reynolds on an earlier occasion (Boswell, Life of Johnson, ed. Hill-Powell, II, 6, n.3) that he might have revised as many as eighteen; and from the collation given in my edition of the Prospect it would seem that even this number may be an underestimate.


Page 112