University of Virginia Library

3. Jarvis and the Question of Plagiarism.

The third charge against Smollett's Don Quixote--namely, that the work is a plagiarism of Jarvis's version--brings its place in the canon into question no less surely than if it had been written by committee, as Linsalata supposed. The charge originated in 1791 with Lord Woodhouselee, who called

the work "little else than an improved edition" of Jarvis, whose own version, though faithful to the literal sense of the original, he found "heavy and aukward."[52] Confirmed, as we earlier remarked, by Duffield and Ormsby in the next century, and more recently by Samuel Putnam,[53] this remains the orthodox verdict on both authors.

No one before Linsalata, however, was willing to undertake the drudgery of attempting to prove by collation the extent of Smollett's dependence on Jarvis[54]--or, indeed, of his dependence on other possible models in English and in French. By comparing a dozen selected passages from Smollett with the corresponding passages in the translations by Thomas Shelton (1612-20), John Philips (1687), Peter Motteux (1700), and John Stevens (1700)--and by similarly comparing twenty-eight short excerpts from Smollett and the French of Oudin-Rosset--Linsalata was satisfied he had sufficiently demonstrated the improbability that Smollett had followed any of these versions in producing his own.[55] What these collations actually reveal, however, is the dependence of Stevens on Shelton (whose work he revised) and the closeness of Smollett's phraseology to that of Stevens--the reason for this being that Jarvis, who served as Smollett's guide through the difficulties of Cervantes' Castillian, wrote with his own eye on Shelton. In this period, threading labyrinths to the source of a translation is, generally speaking, no simple task. The version called Motteux's, for example, was not written, but published, by him--the work itself being, as the title-page declares, a pastiche executed "by several Hands." There is reason, moreover, to doubt that the version published under the name of Charles "Jarvis" was in fact wholly written by Pope's friend, the portrait painter Charles Jervas (the name itself being garbled on the title-page): Jervas died in 1739, three years before the work was published in 1742; he was not, except for this one ambitious work, an author at all; and Pope, who knew him well, declared to Warburton that he had no Spanish.[56]

What concerns us, however, is not the authorship of the translation published under Jarvis's name, but the extent to which Smollett depended on it for his own. For the answer to this question, we are chiefly indebted to Linsalata, who alone has performed the task of comparing, page by page, the two versions of Cervantes' novel--a work of more than 400,000 words. The result of his collation, however, is more ambiguous than it appeared to either Linsalata or his champion, Professor Knowles, who believed that Linsalata had "demonstrate[d] beyond any reasonable doubt that the bulk of the Smollett translation was a poorly disguised theft from that of Jarvis."[57] The documentation to which Knowles refers consists of 472 parallel passages (ranging in size from two to sixty-two typewritten lines) which Linsalata divides into four distinct categories of literary theft: namely, to use his own terms, "plagiarism," "paraphrasing," "rewriting," and "inversion." In Smollett's Hoax (pp. 14-15) he illustrates these techniques with the following examples comparing the original with the versions of Jarvis and Smollett:

In making his case against the integrity of Smollett's translation, Linsalata presumably considered these seven parallel passages among the most definitive and damning of the 472 he culled in the course of collation. For this reason--and recalling the axiom that one can gauge the full measure of Hercules from the dimensions of his foot--I have reprinted them at length. What do they in fact reveal about Smollett's habits of literary appropriation--to borrow that useful term from Roger Lund's essay on the practice of Pope?[59] We can agree, I believe, that the claim that Smollett subjected Jarvis's text to four distinct "techniques" of appropriation is specious: at what point does plagiarism shade into paraphrase, and when does paraphrase become rewriting, or rewriting, inversion? Besides this unhelpful confusion of terms, not all the examples above will be seen to serve Linsalata's purpose: the last two passages show nothing more than Smollett freely rendering the sense of a line or two of the original--an aspect of the theory of translation in the period to which he explicitly subscribes in his preface. These same two passages, moreover, remind us that, whatever liberties a translator may allow himself in order to convey the spirit of the original, he is also concerned to render as faithfully as possible the sense of the work; and the range of synonymous locutions in his own language is limited. Is it surprising, therefore, that both Jarvis and Smollett should use identical words in translating, for example, "Heroica resolución del gran Filipo Tercero"? Except for Shelton, who was first to translate Don Quixote into any language, this simple constraint, inherent in the nature of the genre, has meant that every subsequent translator of the work will be heard to echo, from time to time, one or more of his predecessors. This being so, Smollett's decision to turn when in doubt to Jarvis, author of the most literally exact translation available to him,[60] may be seen as a virtue; it imparted an essential accuracy to his translation as he set about meeting the greater challenge of capturing the color and vitality of Cervantes' masterpiece.

Linsalata has adduced enough unquestionable evidence to demonstrate the extent of Smollett's dependence on Jarvis. Even so, far from justifying the claim that Smollett's translation was a hoax, that evidence just as surely points to Smollett's having written the great bulk of the work with his eye on Cervantes, not on Jarvis. Even if we grant the relevance to Linsalata's case of all 472 examples of Smollett's alleged borrowing from Jarvis--and besides that figure being inflated, as we have seen, it includes at least one example in which Smollett's rendering of the original has been shown to be more accurate than that of Jarvis[61]--it will appear that 86 per cent of the text did not yield passages to his purpose.[62] Indeed, the number of readers who have compared

the two versions without suspecting Smollett of plagiarism supports the view that the offending passages are relatively few and far between. Windham--though, to be sure, he had examined only the specimen of Smollett's translation and not the whole work--was so little inclined to make this particular complaint that he could "scarcely suppose [Smollett] ever could have seen" Jarvis's more accurate version; and Griffiths, the first actually to compare the rival versions of the same passage from the original, was in no doubt that Smollett's "genius (notwithstanding some things that appear to be rather inaccuracies than defects in judgment) comes nearest the great original."[63] Far more impressive is the verdict of Charles Duff, an able linguist and translator of numerous Spanish authors, including Quevedo. At the request of Smollett's biographer, Lewis Melville, Duff compared Smollett's translation with other "standard" translations and came to the following conclusion:

Sixty years would pass before the Noonday Press took this hint, publishing in 1986 a reprint of the first edition with an Introduction by Carlos Fuentes. In his Foreword, Fuentes declared his preference for Smollett's Don Quixote over all other English versions. In this, he was not alone among readers whose native language is the language of Cervantes. Francisco Rodríguez Marín, perhaps the preeminent Cervantist of our century (whose edition of Don Quixote Linslata himself regards as authoritative), was of the same opinion. When told by Cordasco of David Hannay's opinion that Smollett was sufficiently in sympathy with his author to have produced a translation having "an original literary value of its own," Rodríguez Marín replied that "he heartily approved" that judgment: "of all English translations, he entertained a particular fondness for that of Smollett."[65]