Pyrrhonism and Paranoia:
Recognizing a Fake
When You See One

Jack Lynch

Delivered 18 March 2000
at the Princeton Eighteenth-Century Conference

Months of reading about the Popish Plot, George Psalmanazar, Thomas Chatterton, and William Henry Ireland prompted a naive question: what is a fake? A little algebraic manipulation will show how complicated the seemingly innocent question is: fake equals extended lie; a lie is not the truth; divide both sides of the equation by negative one, and you're left with: What is truth? Jesting Pilate didn't stay for an answer, and my fifteen-minute paper won't solve to anyone's satisfaction a twenty-five-hundred-year-old ontological puzzle.

A comprehensive theory of fakery is beyond our grasp, but an exploration of the criteria actually used in the eighteenth century to distinguish the echt from the bogus is worthwhile. I take as an example the debate over the poems of Ossian -- not in the interest of contributing to the Ossianic question, but rather to raise the issues that appear in many investigations of eighteenth-century fraud. The dispute is useful for several reasons. First, it was prolonged: it started in 1760, and as late as 1825 books were still appearing on the subject. (Macpherson still has his defenders, although an answer today is necessarily more subtle than thumbs-up or thumbs-down.) Second, Macpherson raises more questions than the relatively unsophisticated forgeries of, say, William Henry Ireland.

But how hard could it be to settle the question? In fact it's often not difficult -- at least in theory -- to determine whether a particular document is a forgery. I lack the expertise to inventory all the tools in a forensic critic's toolbox, but they include studies of handwriting and watermarks, microscopic examination of fibers, and the chemical and ultraviolet analysis of ink and paper. You needn't know the details of spectrographic analysis to understand the principles behind them. And as luck would have it, most fakes have been found guilty on more than one charge: think of Vrain-Lucas's embarrassingly awful forgeries of the letters of Sappho and Caesar written in French on nineteenth-century paper.

But many of those techniques didn't exist in the eighteenth century; besides, the new ones come from materials science, and a central feature of the Ossianic debate was the lack of material evidence. As it happens, the arguments that did exist are interesting because they're part of a major epistemological and historiographical sea-change. And so my interest is the grounds on which these disputes actually took place -- not whether specific charges were right or wrong, but what kinds of evidence were admissible before the bar.

Not all of them were dignified: many battles were fought ad hominem.1 The disputants never forgot that Macpherson wasn't the most sympathetic character. Even his friends found him hard to take: David Hume, once a supporter, said he never knew "a man more perverse and unamiable."2 Johnson famously defied the surly and petulant Scot, whose threats of bodily harm in a "foolish and impudent letter" prompted him to carry "an oak-plant of a tremendous size."3 In 1805, the Committee of the Highland Society worked hard "to conduct its inquiries, and to frame its report, in a manner as impersonal as possible," but even they felt obliged to "produc[e] some documents relative to the character and disposition . . . of Mr James Macpherson."4 On the other side, critics aimed the most scathing vituperation at "the Doctor": Johnson's "favourite and leading prejudice" against the Scots was supposed to explain everything.5

Name-calling is tiresome, but some charges were more substantive. The most common eighteenth-century criterion of authenticity was consistency -- a fine principle, at least in theory. Consistency can take two forms, internal and external, but both pose serious problems. Tests for internal consistency are the easiest to perform: they require no specialized knowledge and no equipment, only careful attention to contradictions within a work. Edward Davies, for instance, caught some curious shifts in diction in the preface to the Ossianic poems -- "He styles himself, indifferently, the author, the writer, and the translator"6 -- evidence that Macpherson hadn't got his story straight. More damning still, he spotted hundreds of variants between the 1778 and 1807 editions of the supposed Gaelic "originals," suggesting the editor-cum-author was tinkering with his material: he found such changes "not the work of the file, but of the forge."7

That evidence is spot-on, but some shafts missed the target. When Davies drifts from verbal to other sorts of internal evidence, his accusations become, at least for us, less convincing. He complains, for instance, that Macpherson's "characters . . . do not appear to have been immediately drawn from real life," and that "He has not informed our judgments, with the colouring of history, but amused our fancy with that of poetry. It follows that the work cannot be ascribed to Ossian."8 Such accusations claim to invoke no specialized knowledge of the historical or linguistic context -- just good old, reliable human nature -- but in fact the supposedly internal judgments are riddled with implicit assumptions about what an age was capable of. And although they often make for stirring rhetoric, they're liable to embarrassing blunders. Though not a case of fakery per se, the dispute over the Epistles of Phalaris provides a telling example of seemingly internal evidence used improperly. Sir William Temple's defense of their antiquity and authenticity -- "He must have little skill in Painting that cannot find out this to be an Original; such diversity of passions . . . could never be represented but by him that possest them"9 -- is simply wrong, as Richard Bentley delighted to point out.

Even when free of such subjectivity, internal tests aren't always reliable: they sometimes condemn the innocent. In the early days of the Homeric problem, a favorite sport was attributing portions of the epics to interlopers on the basis of inconsistencies. But Parry and Lord demonstrated convincingly that such seeming contradictions are simply inherent in oral composition. And who among us hasn't contradicted him- or herself? If on page 5 of this paper I contradict page 3, it doesn't follow that I didn't write it; it means only that I wrote too hastily or that I'm not entirely in control of my meaning. When confronted with such inconsistencies, our documents may well give a Whitmanesque reply: "Do I contradict myself? Very well then I contradict myself, (I am large, I contain multitudes)."

We escape many of these problems when we turn to external consistency, but that too is a troublesome standard. It depends on objective and reliable evidence about the real world -- linguistic, archaeological, historical, numismatic, even zoological -- against which the doubtful specimen is measured. The most famous example lies well outside our period: in 1440, Lorenzo Valla debunked the Donation of Constantine by spotting references to historical events and linguistic forms unattested in the fourth century. What now seems obvious to us was a breakthrough in the fifteenth century: it depended on a nascent sense of historicism, a hallmark of the new humanism. And eighteenth-century forgery-detectives learned Valla's historicist lesson well. Davies, for example, notes "the glaring appearance of anachronism" in the epics: Oscar died in a battle conducted in A.D. 296, though his sons fought enemies known to have arrived in the eleventh century.10 He appeals to our knowledge of Britain's history, and considers the poems' departure from it evidence of fraud: "Mr. Macpherson overturns the long established account of the colonization of Britain and Ireland."11 William Shaw, author of an Earse dictionary and grammar, used his linguistic knowledge to pick holes in Macpherson's argument. Others pointed out that the lack of wolves in Ossian's Scotland doesn't jibe with what we know of ancient Britain. The arguments come down to this: if the conditions under which a work or artifact was supposedly produced are inconsistent with the conditions under which it was actually produced, we reject it.

But what counts as an inconsistency? Einstein taught us there's no fixed point in the universe; if you say object A is moving with respect to stationary object B, I can say B is moving with respect to A. That's not a bad metaphor. Even if we spot a disagreement between the doubtful specimen and the known body of knowledge, on what grounds do we award the prize of veracity to one side and dismiss the other as worthless? -- what's stationary, what's moving?

Paul Maas, in a technical discussion of textual transmission and corruption, pauses to caution would-be textual critics: "We must distinguish sharply between anomaly and singularity. What is unique is not for that reason alone to be regarded with suspicion."12 This is a salutary lesson even for those with no interest in Lachmannian stemmata and lectiones difficiliores. For if we judge the plausibility of a phenomenon solely on its conformity with what we already know, how can we be receptive to new knowledge? The textual critic rejects anomalies by drawing on his or her knowledge of language, style, the history of the transmission, and so on: Homer never used the genitive here; the line is therefore corrupt. But what if we're confronted with a single instance of the genitive there and have no reason to suspect corruption? Is it actually authorial, a unique usage -- a hapax legomenon -- or is it the product of textual corruption? And what about metaphorical hapax legomena? The rara avis is no less a bird for its rarity. If we admit as true only what is plausible, and draw our standard of what is plausible only from what we have already experienced, we foreclose the possibility of encountering anything new, anything outside our experience.

It's a fine example of the hermeneutical circle: we recognize anomalies because they differ from the norm, but we come to know the norm only by rejecting the anomalies. Hermeneutical circles aren't necessarily inescapable. We can look for an abundance of apparent anomalies, and let the laws of probability take over -- laws first formulated, by the way, in our period. A single inconsistency in a long document will likely be called either a legitimate hapax or a minor textual corruption. Two hundred fifty such forms, however, will probably make us reject the entire document as a fraud. Ditto outrageous departures: one of Woody Allen's stories mentions a newly discovered set of Dead Sea Scrolls whose authenticity is in doubt because they contain the word "Oldsmobile." But even when such judgments are straightforward in practice, making a sound theoretical case for rejecting Nag Hammadi Oldsmobiles but admitting Homeric genitives is complicated. We're told that the word genocide never appeared in the sixteenth century: if we're a text purporting to be from that period does contain the word, is it evidence of the form's early existence, or evidence of fraud?

We're thinking dangerous thoughts, because there's no easy way out. And it just gets worse: you can't prove a negative. Even if the orthography, watermarks, and chemistry of a document are all consistent with its supposed origin, even if the battles were fought in the right century and the wolves are where they belong, you still can't say for certain that it's genuine; perhaps it was a really good forger, better than you at forensics. So how do you know when to apply these techniques? Our stories of perfidious fakers brought to justice by the linguists, the numismatists, and the boys in the lab have something unsettling in common. If we consider the fake as a literary kind, we know only the inferior exemplars of the genre: the best are by definition unknown to us. Some may even be undetectable. What's worse, even detectable forgeries are not always detected, because we don't think to put the question or examine the evidence critically. Some fakes are so outrageous that they force themselves on our attention: an orange dollar bill, for instance, would never make it into our wallets. But can you say for certain that you aren't carrying a dollar bill on which George Washington sports Groucho glasses or a novelty-store arrow-through-the-head? We could spot such a fake in an instant if we thought to look for it, but who among us is so conscientious as to check every dollar bill with an eye for forgeries? Why, then, should we have any confidence that our canon isn't filled with literary Groucho glasses and arrows-through-the-head we never thought to look for?

This thought tormented many eighteenth-century thinkers, who became obsessed with questions of authenticity. This evidence from the ESTC shouldn't be accepted without qualification, but I present it to show a trend: in the first decade of the eighteenth century, only one English book included the word "authentic" in its title. The next decade there were six; in the twenties, seven; in the thirties, fifteen; in the forties, seventy-seven; and so on, increasing each decade, to fully 324 titles asserting their authenticity in the nineties. Eighteenth-century readers were haunted by the prospect of fakes. It's no surprise; like us, they're terrified of seeming stupid. Fakes hit us academics particularly hard because they threaten to reveal we're not as smart as we think we are. Get it wrong once, and history will never forgive you: Hugh Trevor-Roper never lived down his too-hasty thumbs-up on Kujau's Hitler diaries. It terrifies us to imagine sharing a legacy with Pope's dunces.

There's one sure way not to be duped by fakes: disbelieve everything. This was the approach of Jean Hardouin, who in 1697 suggested that virtually every surviving classical text was a medieval forgery. Out of the entire classical corpus, he admitted only Pliny's Natural History, Virgil's Georgics, Horace's Satires and Epistles, and a few of the writings of Homer, Herodotus, Plautus, and Cicero. The rest, he declared, were composed by thirteenth-century monks under the direction of one Severus Archontius.13 This creeping fear of forgers under every rock and around every corner is a paranoia worthy of Joe McCarthy. But one needn't be unbalanced to see fakes everywhere: when Varro looked at the 130 plays attributed to Plautus, he deemed only twenty-one genuine.14 Even Anthony Grafton admits that "perhaps two-thirds of all documents issued to ecclesiastics before A.D. 1100, are fakes."15

Creeping doubt is bad medicine, and eighteenth-century Britons were afflicted with it. Skepticism was the order of the day. Hume notes of Macpherson's fragments: "The first time I was shown the copies . . . I was inclined to be a little incredulous."16 Michael Lort wrote in 1778, "There is said to be a poem of Homer's of above 500 lines found lately in the Royal Library at Moscow. . . . Possibly this also may be a forgery . . . in this age of forgeries."17 Horace Walpole too worried about getting it wrong: in 1781, as Chatterton's star was rising, he lamented, "I believed in Ossian, who is now tumbled into the Apocrypha; and I doubted of Rowley, who is now to rank with Moses and the prophets! -- I doubt, I have very bad judgment."18

This sort of thinking is dangerous: it can lead us to a more-than-Humean skepticism bordering on pyrrhonism, a radical doubt that we can know anything, with no Cartesian cogito to break the stalemate. But perhaps this mise-en-abyme goes some way toward explaining our own scholarly fondness for stories about Psalmanazar, Chatterton, and Macpherson. Tales of deception are popular because, however much we pride ourselves on celebrating the marginal and the counterhegemonic, however much the fake seems to challenge authority, the detected fake is a validation of the status quo: our beloved methods have successfully warded off an assault. Though the existence of a villain like Macpherson is nervous-making, his conviction is reassuring in a get-tough-on-crime sort of way. Seeing the faker brought to justice lets us relax, and to forget for a moment how tenuous our grasp on authenticity really is.


1. Shaw suggests the tenor of the debate: "My adversaries, instead of proving the genuineness of Fingal, have contented themselves with insulting me" (An Enquiry into the Authenticity of the Poems Ascribed to Ossian, 2nd ed. [London, 1782], p. 51).

2. The Letters of David Hume, ed. J. Y. T. Greig, 2 vols. (Oxford: Clarendon Press, 1932), 1:403.

3. John Hawkins, The Life of Samuel Johnson, vol. 1 of The Works of Samuel Johnson, 11 vols. (London, 1787), p. 491.

4. Henry Mackenzie, ed., Report of the Committee of the Highland Society of Scotland, Appointed to Inquire into the Nature and Authenticity of the Poems of Ossian (Edinburgh, 1805), p. 15.

5. Report, p. 35.

6. Edward Davies, The Claims of Ossian, Examined and Appreciated: An Essay (Swansea, 1825), pp. 11-12.

7. The Claims of Ossian, pp. 302-303, 313.

8. The Claims of Ossian, pp. 24, 35.

9. The Epistles of Phalaris, Translated into English from the Original Greek by S. Whately . . . to which is added Sir W. Temple's Character of the Epistles of Phalaris (London, 1699), sig. A1r.

10. The Claims of Ossian, pp. 15, 17-18.

11. The Claims of Ossian, p. 9.

12. Paul Maas, Textual Criticism, tr. Barbara Flower (Oxford: Clarendon Press, 1958), p. 12.

13. See Anthony Grafton, Forgers and Critics: Creativity and Duplicity in Western Scholarship (Princeton: Princeton Univ. Press, 1990), pp. 72-73. Hardouin makes the challenges in his Chronologia veteris testamenti (Paris, 1697) and his Prolegomena ad censuram veterum scriptorum. Among his other singular theories was the contention that the New Testament was originally written in Latin.

14. See Grafton, Forgers and Critics, p. 13.

15. Forgers and Critics, p. 24.

16. Letters, 1:328.

17. The Correspondence of Horace Walpole, ed. W. S. Lewis, 50 vols. (New Haven: Yale Univ. Press), 16:179.

18. Correspondence, 33:317.