John Eccles’s Collection of Songs (1704) and John Walsh’s Publishing Model in the Early Eighteenth Century

Eccles’s A Collection of Songs was first advertised on November 14, 1704:

This makes it clear that the book was not a self-publication like Robert King’s Songs (ca. 1692) or Blow’s Amphion Anglicus. Instead, the volume was printed for the regular partnership of Walsh and Hare, although the vast majority of known copies actually bear Walsh’s imprint alone. David Hunter documents a total of four states of the 1704 Songs, primarily reflecting small variants in the preliminary materials: of these only one, represented by a single surviving copy at Boston Public Library, actually names John Hare—and even then only on the title page, not the frontispiece.¹¹ The equally rare fourth variant state was in fact sold by John Young, whose imprint appears on a paste-down in the copy at the Vaughan Williams Memorial Library, Cecil Sharp House.¹² All this suggests that Hare’s involvement in the project may have been relatively short-lived. That said, he had evidently been on board from the start, since his name also appeared in the proposal issued in October 1703:

Although the 1704 Songs ultimately bore Walsh’s imprint, and regardless of the precise nature of Hare’s role, this proposal reveals that there was at least one additional prominent partner in the endeavor. As with Blow’s self-published Amphion Anglicus, the composer’s own house is listed first among the venues for the contracting of subscriptions.¹⁴ Further parallels with Blow’s publication extend even to its price of 12 shillings, rising to 18 once published (identical also to the post-subscription price for Orpheus Britannicus).¹⁵

Eccles’s involvement in the solicitation of subscriptions, in addition to the evidence from the title page and Eccles’s dedication to Queen Anne, suggests that he was very much part of the venture from its conception. This is important when evaluating the editorial interventions discussed below, since his direct involvement in the production of the 1704 Songs would obviously lend these revisions greater authority. Given this observation, it is worthwhile to consider Walsh’s connections with the composers whose music he published, Eccles in particular, in a little more detail before examining the contents of the publication.

Although, as Carter notes, Walsh had few connections with professional musicians in the last years of the seventeenth century, there is evidence that he began to cultivate such links even at this early stage.¹⁶ In his first major single-composer song collection, Matteis’s Collection of New Songs (1696), Walsh’s signed dedication to Lord Byron¹⁷ recounts that he had “prevail’d upon [Matteis]…to let [the songs] be publish’d.” In his second Matteis collection (1699), however, Walsh addresses his Preface directly to “Mr Nicola” (Matteis). Although Walsh had clearly planned to publish these new songs regardless, after acquiring them through “good Fortune,” Walsh makes clear that Matteis had not only assented to their publication but even provided “favourable assistance in perusing and Correcting the sheets.”¹⁸ Thus while clearly not above obtaining materials from third-party sources, Walsh seems to have seized the opportunity to build relationships with composers—unlike, say, John May and John Hudgebut, whose notoriously unauthorized The Songs in The Indian Queen (1695) was explicitly published “even without acquainting [Henry Purcell] with our Design.”¹⁹ By the early eighteenth century Walsh could also name John Weldon among his collaborators. Weldon’s first two books of Songs (1702) were apparently self-publications, advertised only as “Sould by” Walsh.²⁰ His Third Book of Songs (1703), however, was “Printed for and sould by” Walsh and “Carefully Corrected by ye Author.” The same collection was soon reissued as A Collection of New Songs (1703—not to be confused with a similar 1702 collection), bearing a fulsome dedication to Lord Conway signed by Weldon.

Although Eccles’s songs and instrumental pieces had been published by several London sellers during the late 1690s, there is evidence that he worked particularly closely with Thomas Cross early on: a copy of “By those pignseys” from Thomas D’Urfey’s The Richmond Heiress (1693), “Printed and sold by Tho. Cross,” is subscribed “set by Mr. John Eccles and corrected by him.”²¹ When Eccles later published “Wine does wonders” from John Crowne’s Justice Busy (1699), he again entrusted the engraving to Cross.²² “Printed for the Author” is an unusual formula for Cross’s songsheets: this three-part piece was unlike the straightforward solo songs—with correspondingly simple layouts and broad commercial appeal—that Cross usually published himself. Regardless, the fact that Eccles engaged Cross again suggests that their relationship, even if predominantly professional, remained healthy. It also provides an exception to Rebecca Herissone’s observation that “Cross worked outside the musical establishment…, and it is unlikely that he acquired his music copy via composers or other authorized means.”²³

By 1702, however, Eccles’s relationship with Cross had apparently soured, perhaps because of an episode concerning Eccles’s setting of The Judgment of Paris, his entry in the famous “Musick Prize” of 1701.²⁴ Walsh had already published Daniel Purcell’s rival setting in June 1702.²⁵ Despite some misgivings about the commercial viability of such operatic full scores, dating back to Purcell’s The Prophetess (1690),²⁶ Walsh also advertised Eccles’s version in the Post Man on October 31, 1702—but not before Cross could print, exactly as with The Prophetess, a set of unauthorized sheets containing some of its most popular songs.²⁷ Walsh moved quickly to discredit Cross’s editions, publishing the same four songs in his own competing sheets. Echoing Purcell’s complaints in relation to The Prophetess,²⁸ two of them bear the disclaimer “There being Severall songs out of Mr Eccles Prize Musick printed very falce to the Prejudice of the Master being done without his leave or knowledge, Note these are correct from the Authers Score which is Now a printing and will be Publish’d very Spidily.”²⁹

For all one imagines Eccles feeling particularly aggrieved at Cross’s behavior given their previous association, Eccles’s closeness to Walsh may also reflect Cross’s apparent lack of interest in the larger-scale publications that Walsh had begun to issue in the later 1690s.³⁰ As he had with Matteis and Weldon, Walsh may deliberately have cultivated a relationship with Eccles, starting around 1698. The connection began tentatively: Walsh is listed (with John Hare) among the sellers of Single Songs, and Dialogues, in the Musical Play of Mars & Venus (1697), printed by John Heptinstall for Eccles and Gottfried Finger. Movements by Eccles then featured in the first and third volumes of Walsh’s Theater Musick (1698, 1700); and both his Act Music for The Mad Lover and his pieces for Queen Anne’s coronation appeared in Harmonia Anglicana (1701 and 1702, respectively).³¹ Walsh also featured Eccles’s music prominently in his Collection of Lessons and Aires for the Harpsichord or Spinett (1702). After Eccles was appointed Master of the King’s Musick in 1700, Walsh must have quickly recognized the potential commercial benefits of a closer relationship. The publication of The Judgment of Paris late in 1702—presumably further encouraged by the publicity surrounding the 1701 “Musick Prize”—was thus promptly followed by The Songs and Symphonys Perform’d Before Her Majesty…on New-Years Day (February 1703) and then The Songs and Symphonys…on Her Birth Day (November 1703). A Collection of Songs followed one year later.³²

The two New-Years and Birth Day ode publications of 1703 were rather experimental. The only other odes to have been published complete were the Cecilian odes of 1683 (Purcell’s Welcome to all the Pleasures, published 1684) and 1684 (Blow’s Begin the Song, published 1685),³³ while movements from several court odes appeared in Orpheus Britannicus and Amphion Anglicus. Walsh and Eccles proceeded cautiously with their ode publications, including just the solo and ensemble songs, which were suitable for domestic performance. Instrumental introductions and ritornelli were retained, justifying the title-page reference to “symphonys,” but not the overtures or indeed any concerted choral movements. The novelty of publishing ode music may also account for the inclusion of the New-Years ode as the February 1703 installment of The Monthly Mask of Vocal Musick—Walsh’s monthly periodical, begun in November 1702, offering a handful of the latest songs—which explains the title-page wording, “Published for February. 1703.”³⁴ The later publication of the Birth Day ode in the same format, but outside the periodical series, may indicate that the earlier edition had attracted sufficient interest to demonstrate the viability of a similar, standalone ode publication. Nevertheless, publication of complete odes, the ultimate “occasional” works, remained extremely unusual. William Croft’s Musicus Apparatus Academicus (1720) and Handel’s Alexander’s Feast (1736) provide exceptions; even Handel’s other Dryden setting, the Ode on St Cecilia’s Day (1739), was issued without overture or choruses.

By the time of the proposal for the 1704 Songs in October 1703, then, Walsh had published Eccles’s solo songs in The Judgment of Paris and in the two odes. In addition, five more songs had appeared in the Monthly Mask.³⁵ Strikingly, Eccles’s music accounts for a quarter of the material published in the Monthly Mask between November 1702 and October 1703, when the proposal for the 1704 Songs was released, but then does not appear again in Walsh’s periodical until the publication of the 1704 Songs. Walsh’s avoidance of Eccles’s music in the year following the proposal for the collection seems calculated to protect the demand for the new publication. In the month that the 1704 Songs was published, a single song by Eccles appeared in the Monthly Mask, followed by another three (the whole issue) the following month; thereafter Eccles’s songs crop up sporadically until July 1710.³⁶

The Monthly Mask comprised individual engraved songsheets with separate headings, such that each song could also be sold separately, and the plates kept and used again (a major advantage of engraved sheets over Playford’s typeset songs).³⁷ As such it resembled another format that Walsh made extensive use of at this time, the composite volume.³⁸ Most significant is the Collection of the Choicest Songs & Dialogues first issued in 1703 (hereafter referred to as the 1703 Collection), by far Walsh’s largest single volume to date, with some copies containing well over 200 songs.³⁹ This included more than twenty by Eccles, some already published in one of the forms described above—and nearly all were later included in the 1704 Songs. These collections allowed Walsh to reuse individual sheets (designed to exploit the market for fast, cheap publication of the latest songs) to produce prestige volumes trading on notions of quality and comprehensiveness (designed to appeal to the buyers of publications like Orpheus Britannicus). Later, less ambitious volumes of this type would group songs according to theme—for example, Comical Songs (1706) or Bottle Companions (1709)—to create smaller, cheaper collections that catered to particular musical interests.

The songs by Eccles published in The Judgment of Paris and the two odes, together with those in the Monthly Mask and the 1703 Collection, account for nearly half of the ninety-six songs included in the 1704 Songs (see the appendix). By comparing the versions in these earlier prints with the copies published in 1704, we can learn more about Walsh’s approach to the volume in general and about the editorial hand exercised over the music in particular.

Of the forty-two songs in the 1704 Songs previously published by Walsh, some thirty were demonstrably printed from earlier plates. A typical example is [28] “Fie Amaryllis,” from The Fickle Shepherdess (1703).⁴⁰ Figure 1 shows the song as published in the Monthly Mask in June 1703 and again in the 1703 Collection, while figure 2 shows the relevant page from the 1704 Songs.

The sheets shown in figures 1 and 2 were clearly printed from the same plates: not only do both share an identical layout, down to the precise placement of line breaks, but they also share idiosyncrasies such as the poor spacing of stave-lines in the third stave up from the bottom of the page. Nevertheless, there are obvious differences.⁴¹ Whereas the 1703 sheet has no thoroughbass figures, the 1704 Songs adds a fairly extensive set. This is a common feature of the revisions made in 1704: in every case in which an existing plate was reused, an unfigured or sparsely figured original is updated with what would have been considered a comprehensive set of figures. Before discussing these and other musical changes in detail, however, it is important to examine the treatment of peritextual elements such as titles and ascriptions, page numbers, and other features of the page layout.

In these aspects too, “Fie Amaryllis” illustrates a set of changes that are applied uniformly to the plates that were reused in the 1704 Songs: continuous page numbers are added centrally at the top of each page, and the original ascriptions to Eccles removed, being redundant in a single-composer collection. In figure 2 this necessitated the deletion of the first half of the page title, with the words “A SONG” inserted afresh, slightly larger, and in the middle of the resulting space. The retention of the name of the performer, (Mary) Hodgson, is common; other titles name the play from which the song came and/or the text author, and such details might be retained or indeed inserted to fill space originally occupied by an ascription in the earlier state of the plate.⁴²

Such commonplace revisions may seem comparatively trivial, but they acquire greater significance when considered alongside other features of the book and in the context of its composite nature. As we have seen, many of Walsh’s editions were composite: the 1703 Collection, Comical Songs (1706), and Bottle Companions (1709) are the most obvious examples, but several single-composer songbooks of the early eighteenth century also collect otherwise separate sheets behind passe-partout title pages.⁴³ As David Hunter and N. Frederick Nash observe, modern commentators are frequently indignant at the apparent dishonesty of “new” publications comprising existing sheets, not to mention the use of identical title pages for sometimes wildly varying contents.⁴⁴ A pertinent example is Richard Luckett’s apparently approving quotation of William Smith and Charles Humphries’s disparaging description of Walsh’s 1745 Orpheus Britannicus: “This unimpressive work ‘is merely a collection of single sheet songs engraved at different periods about the beginning of the eighteenth century.’”⁴⁵ In fact both the 1745 Orpheus Britannicus and Walsh’s earlier Mr Henr. Purcell’s Favourite Songs out of…Orpheus Britannicus (1724) contained numerous sheets that had long been available, some also from earlier collections including the 1703 Collection.

Given the prominence of composite publications in Walsh’s everyday modus operandi, Luckett’s criticism of the 1745 Orpheus Britannicus may seem anachronistic, but in truth the situation is more complicated. Walsh and Eccles—or whoever took responsibility for the design of the 1704 Songs—went to considerable lengths to present it as a coherent whole, quite unlike the impression given by the 1703 Collection or even some of Walsh’s smaller single-composer collections, which, for example, frequently lack contents pages and retain redundant musical ascriptions. While composite volumes were tolerated for some publications, they were evidently considered unsuitable in other contexts. Clearly Walsh conceived of Eccles’s collection as a unified whole designed to compete with (or rather, aspiring to stand alongside) the similarly prestigious collections of Purcell and Blow. The handling of peritextual matter is just one feature among many designed to convey the high status and structural integrity of the 1704 Songs:

Frontispiece: The 1704 Songs was the first Walsh publication to use a new passe-partout (fig. 3) designed by French-born painter Pierre Berchet (1659–1719/20), engraved by Henry (Hendrick) Hulsbergh (d. 1729).⁴⁶ Rather than simply providing an elaborate frame like the John Collins example mentioned above (see note 16), Berchet’s more sophisticated design integrates a smaller cartouche containing the title within the architectural background, off-center, behind a viol-playing putto who provides an illusion of depth. This preserves the convenience of the passe-partout while resembling a bespoke engraving, its novelty further enhancing its cachet.⁴⁷ The classical iconography confers further prestige, functioning in lieu of the Latin titles of Purcell’s and Blow’s songbooks. Presiding over the scene is Apollo, with his lyre—perhaps representing Eccles, and thereby echoing the “British” Orpheus (Purcell) and “English” Amphion (Blow). Below this is a reclining Venus(?) singing to the harpsichord, further accompanied by a group of putti: one appearing to sing from a book, from which the aforementioned viol-player also reads, while a third holds a recorder. Finally, by naming the contents a “General Collection of Songs” and not merely a “Collection of Songs” as on the title page, the frontispiece gestures toward comprehensive coverage, as well as emphasizing the collection’s independence from any specific event or production.

Title page: While consciously imitating the wording of Orpheus Britannicus, Amphion Anglicus, and similar collections, the title page nevertheless advertises the composer’s professional status (“Master of HER MAJESTY’s Musick”) as a guarantee of quality. It also draws attention to the presence of instrumental music (“such Symphonys…As were by the Author design’d for any of Them”) and continuo figures, selling points that distinguished the collection from many of the familiar single-sheet engravings.⁴⁸

Dedication: This draws together the contents of the book as a single offering to the Queen, while once more advertising the composer’s exalted status by royal association—albeit in an unusually succinct and matter-of-fact tone.

Contents page: While the frontispiece, title page, and dedication all allude to the collection’s completeness, it is the contents page that fixes the ideal copy. The 1704 Songs sits at the “full” end of Hunter and Nash’s “spectrum of reliability of ideal copy,”⁴⁹ its contents completely consistent across all twenty surviving exemplars.⁵⁰

Page numbers: The provision of continuous page numbers was an obvious way of emphasizing the fact that each page belonged to a larger whole, while the imposition of a consistent pagination format contributes to the visual uniformity of the anthology. Other Walsh composite volumes had no page numbers at all—the 1703 Collection and Comical Songs, for example—or adopted less consistent schemes.

Ordering of contents: The consistent pagination and detailed contents page were necessary because the 1704 Songs was carefully organized by key (see the appendix)—establishing a specifically musical corollary to the conceptual and visual integrity already observed. Orpheus Britannicus had announced on its title page that its contents were “placed in their several Keys according to the Order of the Gamut,” but in fact many songs ignored its key sequence, and major and minor keys were juxtaposed rather haphazardly.⁵¹ By contrast, Eccles’s collection carefully follows the sequence g–G–a–A–b–B♭–c–C–d–D–E♭–e–E–f–F.⁵² Organization is hazier within each key, however: there is some suggestion of rough alphabetization and an approximate progression from solo songs to larger ensembles.

Paper use: Sixteen of the twenty copies listed by Hunter have double-sided printing. This is unusual for composite books incorporating single-sheet songs, which are normally printed on one side (the resulting blank pages emphasizing the separability of the printed items). The integrity of the double-sided format, with single-page songs frequently bearing just one page of a multi-page song on the reverse, meant that such copies were largely indivisible. Conversely, the printing of a few copies on single sides may have been designed to facilitate selling of individual songs separately.⁵³ The planned printing of some copies double-sided and others single-sided would also explain the decision to center the page numbers, rather than placing them on the outer edge of each page.

Page titles: As noted above, these are altered to remove ascriptions made redundant by the single-composer contents. Universal conformity to this decision shows the extent to which Walsh and Eccles prioritized the integrity of the volume over any future use of its pages as single-sheet songs.

Musical text: Although most musical interventions in the reused plates were made for localized reasons, the consistency of editorial procedure—in particular, the provision of thoroughly revised and comprehensive continuo figures—contributes to the volume’s unity of presentation.

Most of these features were ultimately controlled by Walsh and his engraver, even if the premium on the concealment of the book’s composite status and its presentation as an integral whole might be traced to Eccles’s own ambitions for the publication. Glimpses of a musical mind nevertheless shine through already in the organization of the contents, and this in turn may be attributable to the reviser of the music, including the new thoroughbass figures. Having shown how carefully Eccles’s previously separate songs were curated to present the 1704 Songs as a coherent collection, I turn now to the musical revision of individual pieces.

As already noted, the figures added to the thoroughbass of [28] “Fie Amaryllis” in the 1704 Songs (see fig. 2 above) are typical of the collection’s musical interventions in songs that reused plates from earlier Walsh publications. Other songs sometimes had more figures in the original plates (especially when they were extracted from longer works), but in every case the figures are thoroughly corrected and supplemented in the 1704 publication. By early eighteenth-century standards, comprehensive figuring demanded a symbol for anything other than a third and fifth. The figures are often sparser than this might imply, however, because chromatic alterations are irregularly noted and certain commonplace sonorities could reliably be assumed. A sixth above the bass was implicit, for instance, on degrees three and seven of the major scale (see fig. 2, third system, m. 5²; sixth sys., m. 9³), over any raised note (first sys., m. 5), and as the resolution of a 4/2 (fifth sys., m. 4¹; seventh sys., m. 6³).

Complete figures are by no means a given in eighteenth-century sources, particularly for sources notated primarily for personal use (composers’ own manuscripts, for example) or when other factors such as speed of publication or association with celebrated performers were prioritized over accuracy and completeness, as was the case for many single-sheet engravings. The provision of these figures in the 1704 Songs transformed them from documents suitable for professional use (by performers in the composer’s own circle) into texts more suitable for use by musicians—including amateur continuo players—lacking ready familiarity with Eccles’s music.

Given Eccles’s proximity to Walsh around 1704 and the composer’s evident involvement in the project from the outset, collecting subscriptions at his home, Eccles would be the obvious candidate to have carried out this work. Although trained, unusually for the time, as a violinist and not a keyboard player, such a task would not have confounded anyone capable of composing the songs in the first place. Peter Holman has argued that Eccles’s vocal music (particularly the Semele autograph) suggests he was “trained as a composer through the figured bass system,” noting further that he had apparently acquired sufficient keyboard facility to apply—albeit unsuccessfully—to become organist of St. Bride’s, Fleet Street in early 1693.⁵⁴

The case for Eccles as the musical editor is significantly strengthened by close analysis of the musical editing, especially in songs that reuse earlier plates. Although the number and nature of the changes are often trivial when considered individually, when analyzed collectively they reveal a consistent approach that extends beyond the correction of errors and the provision of figures to include more substantive matters such as performance practice issues and even occasional recomposition. This sort of work is unlikely to have been undertaken by anyone except the composer,⁵⁵ so we can be reasonably confident that that the 1704 Songs presents Eccles’s own revised versions. Nevertheless, to maintain due caution, I refer only to the “editor” in the following analysis. For convenience this is divided into three main categories, although in practice (as the discussion reveals) these are closely interrelated: (1) editorial corrections, (2) compositional reworkings, and (3) performance-practice adjustments. I will return to the question of the editor’s identity in the conclusion.

The most straightforward interventions in the reused plates are the corrections of obvious mistakes. In [87] “I gently touched her hand,” for example, a bass e♯ (= e♮) at measure 53, the third of a C-major chord, was originally printed a tone too low in the 1703 Collection. The resulting d♯ is an obvious error, making no sense whether interpreted as a sharp or a natural. When the editor corrected this in 1704 he also took the opportunity to insert some missing accidentals elsewhere in the song, although he did not catch every mistake: at measure 22³ f′ was allowed to stand in the melody despite the newly added figure “6” beneath the bass B♭.

The editor also paid careful attention to the presentation of other details, such as repeats. In [37] “Fear not mortal,” for example, he added first- and second-time endings in measure 32 to ensure that both strains of the verse were repeated. In [89] “As Cupid roguishly one day” the repeats indicated by the original sheet were so confused that a new plate was required to correct them. The 1703 sheet included a written-out petite reprise of the last seven measures, concluding with a first- and second-time mechanism that implied a further grande reprise of the whole duple-time section (more than fifty measures, or about three-quarters of the song). In the new 1704 plate the written-out petite reprise is eliminated in favor of a slightly longer repeat of the last fifteen measures that did not imply a grande reprise. Although the grande reprise may already have been an error in the 1703 engraving, its removal in the 1704 plate may equally reflect a decision made in connection with the original stage performances, or the editor’s assumption that domestic performers would not appreciate such a long repeat. Other corrections and clarifications in the revisions include matters of instrumentation (for example [8] “From this happy day”: “violin” becomes “Hoboy”), tempo ([5] “Stay, ah stay”: “Not to[o] fast” inserted at the $46$ section; discussed further below), and punctuation ([56] “Firm as a rock”: commas replaced with semicolons on third page).

Some of the 1704 corrections reveal important details about Walsh’s earlier sources. A revealing example is [69] “Come, come ye nymphs,” where a particularly idiosyncratic leap up a major sixth to f♯″ in measure 1 of the 1703 state becomes a much more idiomatic downward fourth to e′ in the 1704 Songs (see figs. 4a and 4b). Readers familiar with the pitfalls of typeset music will recognize that these two pitches (under the prevailing G2 clef) exemplify what Alon Schab calls a “mirror-relation,” the result of an inversion of type during setting.⁵⁶ Since both the 1703 and 1704 copies are engraved, this implies that Walsh’s original source must have been a typeset sheet (or at least copied from one). Revealingly, Henry Playford’s print of the same song in Mercurius Musicus (January 1701), typeset in Pearson’s “new London character,” contains precisely the same error (see fig. 4c).⁵⁷

Another sheet demonstrably copied from Mercurius Musicus is the 1703 state of [87] “I gently touched her hand”: in a more straightforward error here, the note printed a tone too low discussed above is identically incorrect in Mercurius Musicus (March 1699). Confirmation that this was Walsh’s source comes from the idiosyncratic placement of slurs in his engraving, which erroneously preserves details originally occasioned by Pearson’s inability to slur together more than a few typeset notes. Given his obvious reliance on Mercurius Musicus in these cases, many other Walsh songsheets may also have been copied from there (barring contrary evidence). His earliest sheet featuring [52] “Must then a faithful lover go” is another likely example, preserving a near-identical text—even down to the sparse bass figures—to Mercurius Musicus (January 1701).

Walsh may have also copied from engravings by Thomas Cross. Walsh clearly had access to and indeed sold some Cross sheets, even including a few in copies of the 1703 Collection; examples include [23] “Oh, the mighty power of love” and [24] “So well Corinna likes the joy,” both apparently copied from these same Cross sheets when re-engraved for the 1704 Songs.⁵⁸ Another song likely to have been copied from Cross is [33] “Wine does wonders,” engraved by Cross “for the Author” (see above); Walsh’s 1704 engraving follows this A major version rather than those in B flat or C major found elsewhere.⁵⁹

Walsh’s early reliance on Cross editions occasionally left more specific textual markers. In [71] “Belinda’s pretty, pleasing form,” for example, Walsh’s 1703 version preserves an odd reading of measure 34¹: an unexpected vocal c♭′ (= c♮′) amidst a passage otherwise in A major, which is immediately cancelled on beat three with a sharp (fig. 5a). This song had also appeared in Mercurius Musicus (August 1699), but this specific corruption appears to originate from Cross’s engraving.⁶⁰ There, measure 34 is split across two systems (fig. 5b) and the quarter-note on the third beat immediately follows the key-signature, the rightmost sharp of which is easily mistaken for a (redundant) accidental. Apparently, the engraver of this song in the 1703 Collection inferred from this that there must have been a chromatic alteration earlier in the measure, inserting the flat on beat one.

Although this song does not reuse earlier plates in the 1704 Songs, a revised version of the 1703 plate apparently intended for the 1704 Songs remarkably survives in the Huntington Library copy of the 1703 Collection (RB474197 [85]). This sheet is printed from the same plates used in other copies of the Collection, but the ascription “Set by Mr. John Eccles” has been removed (although traces of it can still be discerned) and a comprehensive set of bass figures added, as with the songs that reused earlier plates in 1704. In the end, the song was re-engraved over seven systems rather than six, but the new version for the 1704 Songs incorporated the same continuo figures as the Huntington 1703 plate (with only light revisions) and finally removed the erroneous accidental from measure 34 (though not the now-redundant sharp on beat three).

The initial revision of the 1703 plates of “Belinda’s pretty, pleasing form” reveals an editorial process that took place over several stages. The editor must have first supplied the thoroughbass figures for this song to be added to the 1703 plates; only after this work was complete was it decided to start from scratch and spread the music over an additional system. This is not the only song in the Songs to show successive stages of revision. A copy of [5] “Stay, ah stay” at the Royal Academy of Music (Robert Spencer Collection, uncatalogued) preserves an impression of the original Monthly Mask and 1703 Collection plate, again with the ascription removed (but still just visible) and continuo figures added, as well as featuring a small melodic reworking (discussed below). This time the engraver had gone so far as to add the 1704 page number (8) before it was decided to replace this rather congested plate altogether and save space by omitting a written-out petite reprise at the end of the first section. In Walsh’s 1706 Comical Songs (GB-Lbl K.8.k.19), [73] “Fill all the glasses” also appears to be a revision of the 1703 plate, without ascription and with some figures added: perhaps another remnant of a preliminary revision for the 1704 Songs, run off before the complete figures and pagination were added.⁶¹

The process of altering engraved plates did not permit wide-ranging compositional changes. In theory, Eccles could have undertaken more extensive revisions to songs that were newly engraved, although no sources have yet been found to demonstrate this conclusively. Nevertheless, the changes made to the reused plates often exceed mere correction. A good example, illustrating nicely the potential crossover between correction and revision, can be found in the approach to the final cadence in [85] “If I hear Orinda swear.” The version in The Monthly Mask (January 1703) and 1703 Collection (ex. 1a) is perfectly acceptable, albeit difficult to accompany in four parts without creating consecutive octaves—which might explain why the editor intervened. The revision (ex. 1b) is much improved, removing a bland octave at measure 43¹ and proceeding in tenths to provide a more tonally driven cadential approach.⁶²

[87] “I gently touched her hand” (discussed above) also contains a small compositional reworking toward the end of its third section. Here the original bass in measure 59 preserved a modified echo of that in measure 57, underpinning a melodic sequence (ex. 2a). Eccles’s first version may even have retained an identical bassline in both measures, perhaps assuming a melodic doubling at the lower third (m. 57) or lower sixth (m. 59), connecting the measures by octave-inversion in the upper parts. If so, even before the 1703 print Eccles had modified the bass rhythm in measure 59—perhaps disliking the accented passing note on the last beat in the putative original (ex. 2b). By the time the song was revised for the 1704 Songs the connection between measures 57 and 59 was forgotten, and the weak measure 59—a [6]/4 followed by an uncharacteristically hasty, eighth-note progression on beat four—was improved by a half-note d♭′ on beat three, providing a stronger approach to the dominant (ex.2c). In the process this instead anticipated the similar figure in measure 62.

It is significant that in both of these examples the editor privileged melodic integrity, instead modifying the bass: his primary concern was to ensure the best harmonic accompaniment to the song. Indeed, I would further relate this observation directly to the general revision process that is in evidence throughout the 1704 Songs. The most significant musical intervention affecting the whole collection is the provision of comprehensive continuo figures, and the revisions to the bass line I have observed are precisely the sort that one might conceive while trying the songs at the keyboard (or indeed imagining doing so) to determine or check over the necessary figures. On occasion, the editor may even have overthought this process. Example 3 shows a revision from [35] “If wine and music,” apparently designed to remove a bass passing note at the end of measure 66. At the lively compound-duple tempo needed here the original seems entirely unproblematic; only if the harmonic rhythm is slowed down does the implied sonority present any difficulty. This song, incidentally, is another that was clearly edited in two stages. The reworking first appears in a single sheet printed from the 1703 plates with added bass figures (GB-Lbl G.425.rr [8]). Only later were the continuo figures completed, the title altered to replace the ascription with a reference to Matthew Prior as author of the text, and the 1704 page number added.⁶³

Melodic revisions are much less common in the 1704 Songs and are often also associated with revisions to the thoroughbass. Example 4 shows a revision made to the 1703 plate of [5] “Stay, ah stay” before it was discarded in favor of a new engraving for the 1704 Songs (see above). An implausibly long accented passing note at measure 25²—made worse by the rising contour—is replaced with a forefall-like half-note c″. An F in the bass could have presented a different solution, but would have forced a return to the root-position chord heard in the previous bar. Another small melodic revision can be found in the opening arioso of [52] “Must then a faithful lover go,” where the editor perhaps considered the leap down a minor seventh in measure 4 (fig. 6a) too awkward for buyers of the 1704 Songs. The solution might again have arisen from work on the continuo, since the resulting falling fourth (fig. 6b) touches briefly on the suspended seventh in the figures.

I began this section by observing that the process of correcting engraved plates discouraged wide-ranging recomposition, so it is fitting to conclude it with a further observation about “Must then a faithful lover go,” which offers an example of the converse: since the 1704 Songs does not in fact reuse an earlier plate, the editor was free to make some wider-ranging revisions. The reason for the new plate is obvious: the original has an unusually tall format, incompatible with the layout of the 1704 Songs (fig. 6a). The new engraving was spread over two pages, with a transposed flute version filling up the second. Apart from the slight revision to the opening arioso described just above, the primary changes that the editor made were to recast the $43$ section in $23$⁠, with doubled note-values, and to rewrite the bass in measures 14–16 to remove a tricky unprepared seventh (fig. 6a, end of system 2; fig. 6b, systems 3–4). The change of time signature and note-values may have been intended to ensure the correct tempo relationship with the music immediately before and after this section, notated under : the new version implies a simple and intuitive sesquialtera (three minims of $23$ to two of ), whereas in $43$ some performers may have attempted to take the triple-time twice as fast. This revision anticipates the final category of musical interventions in the 1704 Songs: changes relating to performance practice.

One simple change relating to performance is the extraction of individual sections from longer continuous passages. In [86] “Saturnia, wife of thund’ring Jove,” for example, the plates from the 1703 opera print began with a descending bass scale in quarter notes from b to e, thereby linking the preceding song, “Happy thou of human race,” in B minor, to the new song in E major. When these plates were revised for the 1704 Songs the editor replaced this measure with a whole-note e. He neglected the equivalent change at the end, however, retaining the second-time bar, with its change to triple time—and link to A major—designed to introduce the next movement in the opera, [32] “This way, mortal” (which appears nearly one hundred pages earlier in the 1704 Songs).

A more dramatic change can be seen in [1] “Cease of Cupid to complain.” Like “Must then a faithful lover go,” this song was re-engraved for external reasons—in the 1703 Collection it shared a sheet with “Chloe found love” (The Agreeable Disappointment) in A minor (fig. 7a),⁶⁴ whose proximity to “Cease of Cupid to complain” (in G minor) would have spoiled the key-scheme of the 1704 Songs. Freed from the constraint of reusing the earlier plate, the editor supplied this song with a moto-perpetuo walking bass in eighth notes (fig. 7b), elaborating the simpler pattern of the original. Very few changes are made to the vocal part, and the implied harmony remains largely unchanged. It seems that the original plate derived from an earlier manuscript that simply sketched out the harmony; the bass divisions were added later in the creative process and included by the editor in 1704 to enable amateur performers to achieve something equivalent to the song’s performance in the theater. By providing this version in the 1704 Songs the editor could both add value to a song already available in print in simpler form and allow readers access to a style of performance they may have been ill equipped to devise for themselves.⁶⁵ Something similar could be said in relation to the addition in 1704 of many more performance directions, dynamics, and repeat markings.⁶⁶

I began with the aim of examining the 1704 Songs to understand better how Eccles’s songs were assembled and revised for publication, to determine the composer’s degree of involvement in the publication, and ultimately to evaluate its reliability as a record of Eccles’s own preferred versions of his songs. It is clear that Walsh and Eccles went to considerable lengths with the 1704 Songs to adapt Walsh’s publishing model—in particular, his successful strategy of producing inexpensive engraved songs that could be sold separately or bound up into composite books aimed at different markets—to create a product worthy to stand alongside Orpheus Britannicus and Amphion Anglicus as a prestigious single-composer collection with visual integrity and consistent editorial control.

Although the close involvement of composers in publications of their own songs might be taken for granted today, seventeenth- and early eighteenth-century printers and publishers did not rely in the same way on direct contact with composers; as we have seen, Cross and Walsh frequently published music that they had obtained from other sources entirely. Nevertheless, the history of Walsh’s publication of Eccles’s music demonstrates that the publisher gradually cultivated a closer relationship with the composer in the period leading up to the 1704 Songs, and the book’s paratextual elements (together with the original subscription advertisement) make clear Eccles’s involvement at every stage. There is even evidence that Walsh’s relationship with the Eccles family transcended the strictly professional: Walsh was named executor in 1711, along with Eccles himself, in the will of Eccles’s father, Henry II Eccles (ca. 1646–1711).⁶⁷ Most persuasive, though, is the extent and consistency of the editorial approach to the music: the degree of compositional intervention in the print as a whole—even if many individual changes are comparatively trifling—makes it unlikely that the editor was anyone other than Eccles himself.

Starting from White’s observation that some of the 1704 Songs were reprinted from earlier plates, my analysis of that process has revealed a thorough and coherent editorial approach: the songs were carefully corrected, supplied with extensive continuo figures, and in many cases revised to improve details of harmonization, counterpoint, and occasionally the song melodies themselves. The integrity of this process and the strong likelihood that Eccles was responsible mean that we can be relatively confident that the composer approved of the published forms of his songs.

My focus here has precluded a thorough examination of the songs Walsh printed for the first time in the 1704 Songs, but if Eccles was indeed responsible for editorial oversight of the reprinted songs, it makes the most sense to assume that this applied equally to the remainder (indeed, the greater part of the book). This might conceivably be tested by examining the newly engraved songs to see whether they occasioned similar treatment. Four of the five songs from Eccles’s Hymn to Harmony (his setting of Congreve’s text for St Cecilia’s Day, 1701) printed in the Songs, for example, are marked with a pencil cross in the autograph manuscript (GB-Lbl RM 24.d.6), perhaps to identify them to a copyist preparing them for use in the 1704 Songs. In [46] “Ah, sweet repose” several new continuo figures are added to the manuscript in different ink, and the same figures appear in the Songs; the same song has a small revision to the bass in measure 24 as printed, which must have been made in an intermediate copy, since it does not appear in the manuscript.⁶⁸

For a further example we might briefly examine Eccles’s most famous song, [84] “I burn, I burn,” which Walsh had included in his own engraving—heavily reliant on an older Cross sheet⁶⁹—in the 1703 Collection, but for which new plates were produced for the 1704 Songs. This may have been to provide a more spacious layout by removing the transcription for flute. The 1704 revisions clearly accord with the focus and scope of those found in the reused plates discussed above. There are obvious corrections: the addition of a missing sharp to a lower neighbor-note in measure 4 of the solo part (ex. 5a), replacement of a whole-note bass e at measure 15 with the correct c (ex. 5b),⁷⁰ and provision of the missing second beam from the four sixteenths at measure 19¹. As in many reused plates, a repeat is clarified, in this case using the customary :S: to indicate a petite reprise at the end of the central bourrée-like section (“’Twas pride hot as hell…not when I was well”).

The process of checking the bass of “I burn, I burn” in 1704 and providing continuo figures also led (as in the cases described above) to several localized compositional revisions in the bass (ex. 5c–e). Each improves the counterpoint between treble and bass, providing greater harmonic variety and momentum (especially ex. 5c) and removing some rough edges (ex. 5e). Finally, several changes help clarify details of performance practice for purchasers of the 1704 volume outside Eccles’s professional circle: tempo indications are supplied in four places, and the precise placement of a change from to C is moved to include an anacrusis. Even the alteration of the melody at measure 18^1–2 (ex. 5c)—unusual, since like throughout the 1704 Songs, the editor intervened in the bass far more often—can be understood in these terms. While the intuitive addition of this suspension could have been expected of professionals, the inclusion of the 4–3 bass figures and written-out melodic ornament suggest that Eccles thought readers of the 1704 Songs may require explicit prompting.

Given the extent of Eccles’s likely involvement, the 1704 Songs begins to resemble something like an authorized print. Such authoritative prints are of course more familiar in the nineteenth and twentieth centuries, and their status accrues in part from their position at the end of a creative process that begins with initial sketches and drafts, continues through successive revisions and fair copies, and arrives at a printed version bearing the composer’s imprimatur—figuratively and often literally. In the late seventeenth and early eighteenth centuries, however, when publishers, printers, and composers had competing (and often conflicting) reasons to print (or not), and when varied sources—manuscript and print alike—continued to circulate long after revised versions had been published, it would be rash to assume that the 1704 Songs fixed an “ideal” form of Eccles’s songs. Conversely, it is clear that Eccles and Walsh’s publication is only one part of a story of increasing collaboration between composers and publishers (witnessed also in the prints of Eccles’s close contemporary William Croft and gathering pace in the relationships between Handel and his printers, for instance) that would gradually strengthen the influence composers had over the musical texts disseminated by printers, and ultimately facilitate the later establishment of the authorized print.

For this reason, although the 1704 Songs derives some authority from its apparent proximity to Eccles, it is important to remember that like all early musical sources, this publication must be interpreted carefully in light of its original functional contexts. I have frequently remarked that Eccles may have been guided by the need to adapt texts originally designed for close colleagues in order to make them more intelligible to less experienced readers. While this often meant replacing assumed performance practices with explicit notation, it could equally entail removing or obscuring aspects of the original theatrical performances that were impractical or irrelevant in the new context. As a result, by privileging these versions modern critical editions may inadvertently drift further away from what was heard in the theater, like the 1704 version of “Saturnia, wife of thund’ring Jove,” with its partly obscured links to its original surrounding music.⁷¹ The songs may be largely the same and bear recognizable markers of original context, but some aspects of their theatrical forms may be removed or concealed.

One response to such observations may be to condemn the whole concept of the modern critical edition; indeed, there may be better ways to record and disseminate the range of surviving textual and musical information for a given piece—especially as editors, readers, and performers become increasingly familiar with a wider range of formats that can imaginatively supplement the printed book. Regardless of how modern critical editions develop, however, their utility will continue to rely upon the ability of readers and performers to take advantage of the information contained therein. This is where close analyses such as those provided in this article are most beneficial: the additional information offers potential benefits for the performance of Eccles’s music and also a way of training scholarly readers and performers in the use of sources and editions, across a much broader range of repertoire. Even—perhaps especially—the clear and reliable text presented by a good critical edition can only ever be a starting point for vibrant and eloquent performance. Studying the kinds of editorial changes made by a composer such as Eccles in preparing his music for publication can expose a range of possibilities for performance that would otherwise remain hidden.