Introduction

Literary fakes have a history as long as that of writing; non-literary faked artefacts have existed longer still.1 In specific instances, the volume of forged historical artefacts and literate works could be quite high. Thus, Anthony Grafton (1990: 24) has estimated that ‘perhaps half the legal documents we possess from Merovingian times, and perhaps two-thirds of all documents issued to ecclesiastics before A.D. 1100, are fakes,’ while Andrew G. Vaughn and Carolyn Pillers Dobler (2006) have speculated that—based on statistical considerations—even the majority of unprovenanced Hebrew seal impressions are, in fact, forgeries.

Despite the prevalence of frauds and fakes, modern scholarship of the last three hundred years exhibits methodological ambivalence on the subject. On the one hand, the necessity of providing constrained criteria for controlling the study of literary fakes has been emphasized.2 One of the earliest such pleas is present in Jean Mabillon’s 1704 supplement to his De re diplomatica (1681), which states that ‘it is necessary to devise and hand down rules for distinguishing genuine manuscripts from those that are false and interpolated.’3 For Mabillon, lack of established criteria would lead scholars to become ‘hypercritical’ of the suspected documents, leading to numerous false positives. Specifically, Alfred Hiatt (2009: 357) has noted that Mabillon feared the hypercritical attitude ‘could undermine public institutions, which based their legitimacy upon old documents: royal and aristocratic houses, the Church itself, and particularly monasteries, which were routinely accused of forgery.’ On the other hand, any and all of such proposed criteria have been lacking in their degree of constraint, nor have they become generally accepted within academia. The six rules for discerning between true and false documents Mabillon had suggested were already deemed inadequate by his earliest critics, and subsequent studies have each continued to provide a variable number of criteria of their own.4 This suggests a continued lack of general acceptance of any one of them, except on the most abstract of levels. Most recently, Tony Burke (2017a: 233) has observed that existing inquiries into forgeries could be codified into a ‘canon of evidence,’ applicable for efforts to ‘test the authenticity’ of spurious works, implying the lack of such canonicity at present.

The key concept above, ‘constraint’, refers to the qualitative aspect of such criteria. A criterion can be considered constrained when it cannot be used equally well—after satisfying its condition(s)—to argue for both A and ¬A, i.e., both for an argument and its opposite. Examples of constrained criteria within the academic study of history include the following: a) various notions of intertextuality, where, for example, psychological experimentation has suggested that in the genre of ‘prose narrative material’ the common use of at least sixteen consecutive words between two documents strongly suggests one of the two documents has been used as the source for the other (McIver and Carroll, 2004: 1251–69); b) the notion of provenance (i.e., the discovery of an artefact as part of a controlled excavation and its unbroken chain of custody afterwards), which virtually guarantees authenticity of the artefact (Joyce, 2013); and c) the use of Bayesian statistics to assess the available historical evidence (Day, 2008: 31–37). The constraint in each of these cases lies at the preciseness of the criteria themselves—either there are sixteen consecutive words or not; either the artefact has been unearthed as part of a controlled excavation or not. In other words, such criteria have a built-in, inherent control on their conclusions to avoid a methodological free-form that would otherwise follow from their application.

Instead of the constrained criteria Mabillon envisioned, scholars of literary forgeries have been compelled to cite a number of general guidelines they have found useful in distinguishing between the genuine and the faked. These include James Anson Farrer’s (1907) concentration on anachronisms, Edgar Johnson Goodspeed’s (1956) weighting of the availability of the manuscript, Grafton’s (1990) point of uncovering out-of-place subtexts, and Harold Love’s (2002: 184) notion that ‘a fake … is usually a little too good to be true’; even an assurance that ‘years of hands-on research and examination’ simply develops a connoisseurship that distinguishes the real expert from those ‘scholars and curators’ who ‘look but do not see’ (Muscarella, 2013: 41).

As criteria for distinguishing between the authentic and the faked, all of the above are sufficiently loose to be as much a hindrance as a boon to the task. Consider, for example, Love’s generalization ‘too good to be true’, a rule-of-thumb that is entirely dependent on the framing of the suspected object. Put another way, ‘too good to be true’ would be the perfect description of jars recovered from a Coptic graveyard containing Greek papyri leaves, caves full of ancient scrolls mostly in Hebrew, or a sealed jar containing twelve codices of mostly early Christian texts in Coptic. These chance discoveries are most commonly known as the Chester Beatty Papyri, the Dead Sea Scrolls, and the Nag Hammadi Library, respectively.5 Perhaps we should not be surprised, then, that the publications of these landmark manuscript finds generated initial suspicions of forgery. After all, apart from being ‘too good to be true’, all of them also challenged established scholarly ideas—and such challenges could always be alternatively framed as anachronisms. All of them were also interpreted in various ways according to the interpretive choices of their respective times of discovery, which always leaves room for the option of portraying an unfavored interpretation as an out-of-place subtext.6 From hindsight, to give some examples, it is easy to perceive how the early scholarship on the Dead Sea Scrolls overplays the Essene connection, while the early scholarship on the Nag Hammadi library underplays the distinctiveness of the various Christianities that it contains. The Chester Beatty Papyri eventually changed scholarly understanding regarding the early Christian usage of codices, a change in academic history writing that could also have been portrayed as a deviation, i.e., an anachronism from the received scholarly wisdom of the times.7

The most apt observation on the loose character of such criteria comes from Gilbert Bagnani (1960), who made the following comment on the characteristics of debates ‘On Fakes and Forgeries’:

The arguments used in such discussions are not often impressive for their logic. An object may be declared a fake because (a) it is much too good to be true, (b) it is much too bad to be true; because (a) it is like countless other objects, (b) it is not like any known object; because (a) it confirms an established theory, (b) it explodes an established theory; and so on and so forth. (Bagnani, 1960: 244)

What would a forger do, or the present state of forgery discourse

At present, the scholarly attitude, as demonstrated in the words of Christopher Jones (2015: 368), is a casual, untroubled acknowledgement that ‘forgeries, proved or probable, ancient or modern, abound.’ This tendency manifests itself in two troubling developments. First, the discourse of the prevalence of forgery promotes suspiciousness, consequently abating the general tendency to ascribe honesty to interpersonal messages.8 Thus, scholars working at times of reticence are more likely to interpret all kinds of features as peculiarities, solved by historical explanantia. In an analogous manner, scholars working during more hypersensitive times are more likely to ascribe similar features to indications of foul play. Frank Moore Cross, for instance, was content to judge the so-called Ivory Pomegranate, a small artefact with a Hebrew inscription possibly linking it to Solomon’s Temple, to be authentic in 1984, yet change his mind regarding the object’s ‘serious palaeographical problems’ in the early 2000s in parallel with the intensification in efforts to uncover forgeries, stemming from the high-profile cases such as the so-called James Ossuary forgery trial.9 On a societal level, an omnipresent suspiciousness does not lead us to make better judgment calls between the truthful and the falsified—we would end up toiling under a ‘weariness born of wariness,’ as Ralph Keyes (2004: 214) has worded the effects of ‘being on guard lest someone succeed in telling us a lie.’ Thus, objections that we are simply more aware of the possibilities of forgeries because their absolute numbers have increased owing to market forces, miss the point of the first troubling development. Regardless of the fact that faux antiquities exist in surplus numbers compared to the past, the tendency to subscribe to a forgery explanation leads to an increase in both positive identifications of forged artefacts, and false positive identifications of genuine artefacts as forgeries.

Second, there are consequences emerging along the lines of Mabillon’s fears of ‘hypercriticality,’ as the hunt for fakes and forgeries has become a methodological free-form because of the lack of commonly accepted constrained criteria. It is as if the 21st century scholarly attitude has morphed into a frenzied tendency to turn every spurious historical feature into an ad hoc indication of forgery. Symptomatic examples are plentiful: for example, see Stephen C. Carlson (2005) and other scholars’ conviction that the manuscript of Clement’s Letter to Theodore had been forged by its putative discoverer, who peppered the document with multiple clever allusions to his own name; Richard L. Arthur’s (2008) suggestion that the Codex Tchacos as a whole is one elaborate hoax of recent origins; the many articles and monographs of Luciano Canfora and other scholars who argue that the so-called Artemidorus Papyrus is a fake, allegedly forged by none other than the famous nineteenth-century forger Constantine Simonides (Canfora and Bossina, 2008); and the ‘instant reaction’ of extreme suspicion (or, for some, eyebrow-raising certainty) of foul play that scholars exhibited mere minutes after their laying eyes on photographs of the Gospel of Jesus’ Wife (GJW), even though these sentiments were ultimately validated by the investigative journalism of Ariel Sabar (2016).10

In lieu of constrained criteria for distinguishing between fakes and authentic artefacts, contemporary scholars have informally, and most likely unconsciously, developed a powerful method for justifying their conclusions of forgery in each of these instances. As the method remains inexplicable by its proponents, I have chosen to refer to it colloquially as WWFD, an acronym for ‘What would a forger do (were I to be one in this specific situation)?’11 This refers to a thought experiment of placing oneself into the imagined shoes of a forger, and consequently constructing a small justification narrative on how a particular feature of the alleged forgery could have come into existence by the actions of the forger (as imagined by the scholar in question). For a straightforward example, consider the justification narrative offered by Gesine Schenke Robinson (2015: 387) as an objection to the claims that the characteristics of the ink within the damaged areas of the papyrus fragment of GJW could not have come into being in modern times:

A determined forger should not have too much difficulty in folding protruding papyrus fibres over just written letters, sticking a fibre on still wet letters, or even subsequently removing a fibre or two in order to purport deterioration.

The crucial problem here is the issue of validity, for without further decisions to ensure quality control, such stories are only required to contain an internal logic that remains consistent. By means of contemplating actions of an imagined forger, who inherits their intention to forge from an equally imagined motive, a straightforward WWFD requires only a small first-person narrative in which the conceivable functions as its own justification. In other words, WWFD has the potential to explain any imaginable detail as resulting from the deliberativeness of the forger, having no control over itself apart from the request of internal consistency. As such, WWFD exhibits a theoretically naïve manner with which scholars have nevertheless been tempted to attack questions of authenticity, as documented below.

In many instances scholars have also chosen to refer informally to the old popular culture adage of ‘means, motive, and opportunity’ (MMO), a framework within which the establishment of this triad according to the available skills, state of mind, and circumstances favorable to the suspect is understood to strongly point towards the suspect’s culpability. Theoretical support for this approach is offered, for example, in Grafton (1990: 37–68). What the popular MMO-formulation misses, however, especially regarding alleged forged artefacts, is the principle of corpus delicti (literally, ‘the body of the crime’). According to Schmalleger and Armstrong (1997: 57), ‘corpus delicti is the presence of all elements needed to establish that a crime has been committed, which serves as the legal foundation for charging the accused.’ In other words, a criminal act has to have been established before any consideration of MMO can be successfully pursued—yet with a disputed artefact it is the very existence of the ‘criminal act’ (of forging) that is disputed.12 In other words, there is a poor conceptual separation between the notions of whether a given artefact is forged and whether a given forger was responsible for its fabrication. Often WWFD is used to argue for the former, when it pertains only to the latter. One might, of course, point out that to establish the latter as deliberate is to establish the question of forgery at the same time. This I do not deny, but such a conclusion would require that the arguments presented for the ‘means, motive, and opportunity’ were robust enough to actually distinguish the deliberateness of the forger from the deliberateness of the scholars themselves in construing their instances of plausible ‘means, motive, and opportunity’ of the forger. If such methodological robustness has been established somewhere in scholarly debates on fakes and forgeries, I have yet to encounter it.

Unconcealed use of WWFD

Sometimes the use of WWFD is unwittingly spelled out, as in the Robinson example above. For corresponding examples from other scholars, consider the following selection of quotations:

A forger … would have taken the safest course at hand. He would have supplied our scholars with exactly the formula they expected—by inscribing a blessing for the king! (Sasson, 2016)13

But an ancient support and an ancient mode of preparing ink could both be used by a modern forger; [Simonides] was an accomplished chemist … and boasts of his knowledge of the different inks that were used at different times. (Janko, 2009: 405)14

What if there were two stages of forgery? If the recto was made by Simonides and a century later in the second half of the 20th century someone added the verso drawings trying to follow the 19th century manner? … It is rather unlikely that the forger visited every museum of antiquities and then made his drawings. Rather, he had access to diverse catalogues of antiquities or their models in gypsum. Judging by the flattened frontal images of sculptures (Rl, R21) the most probable models were photographs. (Miziur, 2012: 144)15

It was not out of the question that a forger could have repulped historical sheets to fabricate new ones on a modern mould with old fibre material. Or, on an old mould. Another option: rather than making new paper, a historical stock could have been stolen from an archive. Ledgers or notebooks will sometimes have blank sheets. A determined thief could have hunted down blanks in an archive and secretly cut them out without being detected. It is also conceivable that a ledger had already been slimmed at an earlier time, from which point the sheets could have traveled many routes to the forgers. (Brückle, Smith, and Mayer, 2014: 35)16

Even today the Voss edition is obtainable in the used book market: one copy of the 1646 edition had been offered for sale in December 2003 for €280 (about U.S. $320). In addition, Smith [the alleged forger] noted no ownership or other marks of provenance in the book, and the most probable places for such information, the front cover and first pages, had been destroyed … Furthermore, security at Mar Saba would have been more concerned about books being removed from its library than about books being smuggled into it. The book is not difficult to conceal due to its small dimensions (approximately 15 cm × 20 cm or 6’ × 7 ¾’). (Carlson, 2005: 36–37)17

Two of the three acrostics have no explicit connection to Heraclides; the other could just as easily have been placed in the text to satisfy Dionysius’ rather scandalous sense of humor. If so, Heraclides just happened to step into a trap particularly suited for his corpulent frame … And one can think of other options. For example, it is possible that Dionysius knew that Heraclides was beginning a work on Sophocles … or, possible—in a world of limited book distribution—he arranged for this work simply to fall into Heraclides’ hands … and so on. (Ehrman, 2013: 13, 13 n. 8)18

Naturally enough, none of the above examples were used as sole arguments for persuading nonplussed scholars to accept the cases for inauthenticity, but rather to counter (sometimes preemptively) objections to the contrary—by means of establishing that a plausible story of the alleged deliberativeness of the alleged forger does exist. Yet, the instances of unwittingly spelling out the totality of WWFD, as happens in the above examples, are not the main problem in contemporary debates on forgeries. After all, the straightforward ones are the easiest to spot. Far more troublesome is the inadvertent concealment of WWFD that often follows from the more complex presentations of inauthenticity, as explicated below.

Concealed use of WWFD

Recently, Kipp Davis (2017) has considered the characteristics of a number of recent Dead-Sea-Scrolls–like manuscripts that have found their way into private collections. Noting the problems with their provenance, Davis enumerates numerous details of correspondence between these manuscripts: surplus of biblical citations already known from other sources, diminutiveness of size, poor material quality, color range leaning towards darker end of the spectrum, poor scribal quality, palaeographic anomalies, and surprising textual correspondences with modern scholarly emendations (2017: 238–61). The latter prompts Davis to ask, ‘How critically ought we examine readings in manuscripts that appear to be too good to be true,’ and, ‘How much more so does this hold for manuscripts of unknown or dubious provenance?’ (2017: 260).

I have chosen Davis as a prime example of a contemporary technique of employing framing to construct a bricolage that conceals the use of WWFD behind it for two reasons. First, despite the following criticism, I strongly suspect that Davis is on the right track, and that these recent Dead-Sea-Scrolls–like fragments are, in fact, modern forgeries. Second, Davis, as a matter of fact, has worded his conclusions so carefully as to effectively dodge the following criticism, as he ultimately claims only that we ‘should at least disqualify [the manuscripts in question] from discussion as genuine textual artefacts from antiquity and prompt further, urgent investigation into their provenance’ (2017: 263).

Davis’s argument constructs a bricolage by placing numerous individual observations of correspondence one after the other according to an explicitly worded principle: that even the weakest of them ‘is likewise … suspicious’, as it is perceived ‘in the light of a combination of other features’ (2017: 257 n. 78). Although this style of arguing from preponderance of evidence can be acceptable in the abstract, historians who employ volume as a (second-level) criterion are usually careful to note that a ‘cumulative approach to evidence runs the risk of circularity’ (Scott, 2014: 4). For this reason, scholars prefer to carry out their (first-level) analysis ‘under the watchful eye of a strict methodology, rigorously and consistently’ (Beetham, 2008: 12).

How, then, does Davis control the quality of his individual observations of correspondence? Were he to conduct a study of, say, intertextuality regarding the same manuscripts, he would presumably proceed akin to his fellow historians, and evaluate his voluminous examples according to some sort of constrained criteria, of which there are many examples within the crossroads of literary and historical studies to choose from.19 Of course, as previously discussed, no such thing exists for the topic of manuscript forgeries, and consequently Davis can be hardly faulted for not assessing his correspondences in light of non-existent tenets. Nevertheless, the observations of correspondence are left without qualifications. The end result is consequently no more and no less than a list of various ‘patterns of consistency’ between the questioned manuscripts following a single, loose rule-of-thumb: that commonality in physical and textual features points to commonality in provenance—‘perhaps modern common source’, i.e., forgery (Davis, 2017: 233).

There is no doubt that Davis’s presentation makes these manuscripts look extremely suspicious: clearly they are not to be taken at face value. Yet the ‘suspicious effect’ of the bricolage stems largely from the framing of the discussion, by which I mean that none of the observed correspondences are contextualized in a qualitative sense, apart from the notion that, as a collective, these manuscripts are distinct from earlier Dead Sea Scrolls as a collective (Davis, 2017: 262). In other words, we might (and probably should) agree with Davis that the suspected manuscripts are diminutive in their size and rather dark in their overall coloring, but, despite the absolute numbers Davis provides, we are never told just how much weight these details are supposed to carry, and by which criteria they should do so. These are the core questions that are left open.

Unfortunately, framing manuscripts as suspicious in this manner is a surprisingly trivial task. Previously cited examples, such as the notion that Codex Tchacos is a modern forgery, are maintained on equal grounds—that specifically the Gospel of Judas is too little of a genre match with other ancient exemplars, too much of a topical match with contemporary points of discussion, too little of a substantial match with other ancient variants of the Gnostic myth, too much of a literary match with the Secret Book of John, and containing some (but apparently too many) grammatical anomalies (Arthur, 2008). Even more excessively, as documented by Paananen (2012), the manuscript of Clement’s Letter to Theodore was likewise deemed a modern forgery in 2005 with a breathtaking amount of individual details, ranging from its physical characteristics to palæographic anomalies and beyond, in a manner that constructed a ‘suspicious effect’ that pushed a number of prominent scholars to preemptively call the case closed. The assessment of the depth of the ‘suspicious effect’, whether it is warranted or not, required in the latter case a laborious process to consider each of the details framed as suspicious according to some sort of sensible, constrained criteria, including considerations of statistical nature regarding the various absolute numbers presented as suspicious (Solow and Smith, 2009).

An analogous case in the extreme would be one of Ignatius Donnelly’s (1888: 890), who notes: ‘The proofs are cumulative. I have shown a thousand of them.’ Yet it is hardly surprising that Donnelly’s case has not become accepted within its relevant academic discipline, as his ‘thousand proofs’ seek to establish Francis Bacon (1561–1626) as the true author of Shakespearean poems and plays. In themselves—or should we say, en masse—and in lieu of established, constrained criteria for their assessment, such ‘proofs’ are prone to become intuitively assessed as intentional, following the normal cognitive process of assessing causes to originate from deliberate acts of an agent; a truism within contemporary studies on human cognition, as is our tendency to its overuse (McCauley and Lawson, 2002).

When we thus approach Davis’s (2017) observed correspondences with intentional explanation instead of constrained criteria, we see a thoroughly concealed case of WWFD becoming manifest. Though nothing in the article text hints to the effect, every individual detail within it implies a deliberate act of a modern forger which, in turn, has led to the individual observation. Thus, what the forger would have done is the following: they would have gained access to small fragments of ancient writing material of poor quality (both regarding texture and coloring of the fragments), since to obtain qualitatively lower material from the antiquities market would have been easier than qualitatively higher ones. Next, the forger would have had to rely on their own, obviously rudimentary understanding of ancient languages and writing, producing specimens directly on the ancient fragments to the best of their ability. To mitigate the danger of obvious grammatical mistakes, the forger would have employed ancient examples of biblical texts, with the added positive that such manuscripts would extract higher prices from the market compared to more mundane text passages. The small size of the fragments would also naturally mitigate the problems with their handwriting, as the number of glyphs needed to be simulated would have been smaller. Nevertheless, a competent palæographer would still describe the writing of the forger as ‘crude,’ based on details such as ‘bleeding letters’ and ‘misaligned text’ among other script irregularities, while textual scholars would draw attention to the correspondences between textual variants in the fragments and contemporary scholarly emendations they seemingly corroborate, such as the forger’s clear blunder to have included a diacritical notation (superscripted Greek alpha) from Rudolf Kittel’s 1937 edition of Biblia Hebraica, a modern text edition the forger would have used as their model (Davis, 2017: 238–61).

The above story of the forger makes a plausible prima facie case for the recent Dead-Sea-Scrolls–like fragments as modern fakes by its framing of their observed correspondences as suspicious, fit for an imaginary model of the deliberateness of the imagined forger, concealed behind its self-built bricolage. Additionally, many of the more abstracted ‘features of forgeries’ are referenced in its text, such as the previously criticized notion of ‘too good to be true’ as a telltale sign of forgery (Davis, 2017: 260). Furthermore, Davis illuminates the perennial conundrum of interpreting anomalous details with either innocent explanations or hypersensitivity to the potential of forgery in his discussion of a single, amorphous glyph in a manuscript fragment from the collection of the Museum of the Bible (MOTB. Scr. 003175), which has been interpreted inter alia both as ‘a very clumsy [Hebrew letter] vav’ (Abegg Jr. et al., 2016: 215) and ‘eerily similar to a diacritical notation that was printed in Rudolf Kittel’s 1937 edition of the same passage’ (Davis, 2017: 261).

As previously mentioned, Davis (2017: 260, 263) concludes by effectively circumventing the above criticism by formulating its reservations as questions (‘How … ought we examine’; ‘How much more so’) and by proposing a very modest policy suggestion for future studies of the matter (‘should at least disqualify … as genuine … and prompt further, urgent investigation’). Notwithstanding the fact that this strategy is one I would personally have also chosen under similar circumstances, the lack of answers to these questions relegates them to mere rhetorics and might even discourage the development of constrained criteria for the task. By this, I mean that as long as we are content to accept voluminous swarms of suspicious, WWFD-certified details to set the stage of discussion, it is hard to imagine the means by which the debates on fakes and forgeries might make methodological progress of any kind.

Hyperactive use of WWFD

Though the use of WWFD in a concealed, unwitting manner is the most widely spread, some contemporary scholars have also exhibited its opposite, a hyperactive usage of narrating the deeds of the imagined forger. Recently, Javier Martínez (2014) has proposed an exploratory method based on an observation by Robert W. Mitchell (1996: 851), that ‘the only understanding we have of deception is from failed deceptions or those revealed by their perpetrators.’ From this insight Martínez infers that ‘studying successful historical forgers thus requires that there not be a scholarly consensus as to their status as forgers’ (2014: 163). Choosing a contested document, Hippias’s Olympionikon Anagraphe authored in the fifth century B.C.E., and its list of Olympic victors whose authenticity is debated, his conclusion is succinctly put:

Hippias of Elis stands out from other ancient forgers in various ways. Many scholars to this day do not recognize Hippias as a forger, but seek to defend the veracity of his work. His work demonstrates subtlety and insight into the psychology of the reader. His deception is skillfully hidden in virtually every facet of his work. He not only succeeded in deceiving his fellow Greeks for generations, but continues to this day to deceive modern scholars. (Martínez, 2014: 177)20

More remarkable than the certainty expressed in the above conclusion is the chain of inferences from which it is derived. Hippias, as Martínez narrates the story, performed nine deliberations, and either acted or withheld himself from acting in each of the nine instances, in a manner that made his forged list of victors ‘so subtle and so well-constructed that modern scholars to this day continue to debate the authenticity of his work’ (2014: 164). Hippias, for instance, resisted the urge to refer explicitly to any sources that might have become subsequently invalidated, but placed himself as the sole authority behind his work, to make it stand or fall together with his own reputation, should his readers deny him this trust. Nor did he refer overly much to contemporary, commonplace wisdom that could have likewise become invalidated later on, endangering the continued success of his forgery. Furthermore, instead of outright invention, Hippias utilized the names of some genuine victors—ones that other people could verify—to frame the whole composition with an illusionary trust (Martínez, 2014: 165–71).

Hippias, in short, ‘was perhaps among the greatest forgers of all time’ (Martínez, 2014: 164). We can be certain of this fact, because all the literary details in Olympionikon Anagraphe are narrated as part of his ‘expertise … in the subtlety with which he carried out his craft’ (Martínez, 2014: 176). Thus, Martínez writes:

Hippias knew from years of studying human behavior that ‘to generate fictional truths, the observer [reader] is to infer other propositions which are congruent with the representation, much as when a listener attends to the propositions in a story to make inferences’. (Martínez, 2014: 168)

Hippias is easily able to pass the scrutiny of questions about local promotion … He neither erases Elis completely from the early record nor does he insert Elis so frequently as to raise suspicion. (Martínez, 2014: 171)

This is part of the brilliance of Hippias of Elis. Whereas other forgers would go out of their way to legitimate their own work, Hippias knew that a better strategy would be to have others legitimate his work for him. … it was only Hippias’ deep understanding of human psychology that allowed him to pull off his deception. (Martínez, 2014: 174)

Nowhere is the ingenuity of Hippias exhibited better than in the subtleness of his introduction to the seventy-sixth Olympiad, which compels his readers to make a certain associative connection:

Hippias relied … on his trust in the intellectual prowess of his readers and the connections that they would inevitably draw without his help. First, Hippias was confident that the name ‘Scamander of Mytilene’ itself would be sufficient to call to mind the momentous Olympic festival [in 476 B.C.E.] that his readers had heard about on their fathers’ and grandfathers’ knees [being the first Olympiad after the Battle of Thermopylae in 480 B.C.E.]. Additionally, his readers would share Hippias’ penchant for numerology. Once they connected the name of Scamander with the famous Olympic festival, they would then be able to make the calculation of four years multiplied by seventy-five previous Olympiads to produce the significant number 300. With that number in mind, they could then deduce that the Olympics had been held for 300 years until the celebration of the 300 Spartans who faced the Persians at Thermopylae. (Martínez, 2014: 173)

Compared to the unconcealed and occasional, unwitting mistake of spelling out the practice of WWFD, a hyperactive style makes no apologies for its fiction. No detail is let to escape from the process, for instances that fit more easily to the story of the forgery are narrated as such (‘there is no indication that [Hippias] was paid by anyone for this work’), while all the rest are narrated as part of the expertise of the master forger, whose ‘genius’ and ‘innate understanding of human psychology’ is so eminent that the pitfalls experienced by later forgers are all subtly avoided by him—the latter characterization having been inferred from the notion that Hippias’s Olympionikon Anagraphe is so masterful as a forgery that it continues to divide the scholarly opinion as to its status as a (masterful) forgery (Martínez, 2014: 171–72).

To be fair to Martínez, I would nevertheless say that he partially circumvents the above criticism by connecting his study on ancient forgers, and not forgeries (2014: 163). That is to say, Martínez’s idea is to ‘examine Hippias of Elis as a forger and highlight the ways in which he managed to manipulate his audience into accepting his work as genuine … focus[ing] on the psychological nuances upon which Hippias relied to deceive his audience’ (2014: 164–65). In other words, there is nothing to prevent a close reading of Hippias as a forger, regardless of whether he, in fact, was a forger or not, and consequently forged his list of victors or not. In fact, such a reading would fit nicely with Martínez’s notion that ‘studying successful historical forgers … requires that there not be a scholarly consensus as to their status as forgers’ (2014: 163). The study itself, however, lacks the explicit statement that Martínez’s reading of Hippias as an ancient forger follows from his reading of his list of Olympic victors as forged by Hippias. As such, his strongly worded conclusion stands out as the prime example of the hyperactive use of WWFD among contemporary studies on forgeries.

Conclusion

It is a widely accepted notion that academic scholarship cannot be distinguished from other knowledge-generating endeavors except by its dedication to a particular, rigorous methodology, one that demands constant debate about its applicability. The limits of scholarly means are often at the heart of these debates. Samuel Sandmel (1962: 1), for example, warned biblical scholars half a century ago of the dangers of parallelomania, or the tendency to ascribe literary relationships between ancient texts as an ‘extravagance among scholars which first overdoes the supposed similarity in passages and then proceeds to describe source and derivation as if implying literary connection flowing in an inevitable or predetermined direction.’ Sandmel’s warning, as Terence L. Donaldson (1983: 194) has documented, had been precedented by a longstanding fear of ‘an illegitimate and uncritical application’ of parallels within religious studies. In related fields, such as literary studies, the situation has been much the same: parallels that are not qualitatively assessed are not to be used as they can be used to prove anything (Byrne, 1932: 22–23).

The lack of such debate on method—or, even the lack of a sound, authoritative voice like that of Sandmel—brings us to the core issue at hand: that the contemporary debate culture on fakes and forgeries has been filled to the brim with suspiciousness and hypercriticality. This has culminated in the all-powerful WWFD, an explanatory tool with a scope too broad to be used sensibly. As the numerous examples cited above exhibit, WWFD as such is fit to construct a forgery out of anything.

There is without doubt a specific enticement in the act of placing oneself into the imagined shoes of a forger and coming up with a small justification narrative for a particular feature of the alleged forgery. This rests on the asynchronous relationship between the conceivable and the justifiable. That is to say, the use of WWFD ends up constructing an implicitly compelling narrative of a forger, one that rests on the notion of conceivability as its sole criterion, which holds as long as the narrative merely avoids logical impossibilities (whereas an argument that seeks to justify such conceivability requires that the criteria for its justification are met). Additionally, as previously noted, the act of WWFD likely follows from our hypersensitivity to intentional explanations when other, established criteria are lacking, strengthened by a powerful ‘feeling of epistemic satisfaction,’ because ‘we see how the reasons we constructed could also make sense of ourselves when performing that action under a suitable intentional description’ (Buekens, 2013: 442). And the ordinary cure for the excess of intentionality—vigorous debate on the limits of scholarly means—brings us back to the very beginning: ever since Mabillon and his contemporaries ventured on the quest for constrained criteria to distinguish the real from the faux, only loose rules of thumb have become somewhat established in academia.

Where to go from here? One stepping stone on the path onwards is to reformulate the central question that scholars are prone to ask of a suspected forgery. Instead of investigating whether or not the artefact is genuine, scholars might rather consider whether the artefact can be distinguished from a genuine artefact. The adjustment is slight, but its effect forceful. Material reality limits the historian’s ability to answer questions according to what can be justified by referring to the material reality (for example, documents or artefacts), which necessarily forms the basis of the historian’s conclusions. In other words, it is trivially true that a forger could conceivably have done many things that are inaccessible for the purposes of constructing them post hoc from the material artefact(s) themselves. For example, in the case of Robinson and the ink present in papyrus fibres cited above, if there is no difference between dipping damaged papyrus fibres in ink when manufacturing a forgery and authentic papyrus having been damaged so that its fibres contain some amount of (ancient) ink, then this detail cannot distinguish between the real and the faked. As such, the fact that the papyrus fibres contain ink cannot be used to argue for difference (forgery), but only for similarity (authenticity), as the case may be. As I have argued elsewhere with my co-author (Paananen and Viklund, 2015), this slight adjustment of the research question would bring the standards in contemporary studies on forgeries closer to those used in forensic sciences.

A second, more assertive step follows from the above. Should we accept the fundamental truism in forensic sciences that ‘a perfect forgery cannot be detected by anyone’ (Osborn, 1929: 367), our objective shifts to demarcation: to decide which features are admissible as evidence either for or against ascribing the status of forgery to an historical artefact. The crucial notion here is that of constraint—i.e., the control of method in the specific sense used throughout this article—namely, that a constrained criterion is one that cannot be employed to back up both an argument and its counter-argument. For an example, I have recently (Paananen, 2019) attempted to formulate practical constrained criteria to assess alleged forgeries in one of their under-explored aspects of mystification, a literary technique I have labeled concealed indicator of authority. In short, I argue that such deliberate concealment of authorial details by the forger within a forged artefact (or its immediate paratextual material) is recoverable as admissible evidence for forgery, according to specific criteria that is both determinated (for such categories as literary primer, method of decipherment, and plain-text solution) and constrained (for such phenomena as inattentional blindness, cryptanalytic hyperactivity, question-begging, principle of single solution, and room for failure to communicate).

Every other loose rule of thumb used in academia in the past three hundred years could be put through a similar analysis to discover equally determinated categories and constraints that would serve as criteria in future debates on forgeries. I propose that such a research program would be a worthwhile contribution to scholarship. Criteria of any kind are necessarily arbitrary choices to fulfil a practical goal: to facilitate scholarly discussion. Meticulous attention to details in efforts to uncover forgeries is of paramount importance, for we are working at times of suspiciousness and hypercriticality, amidst the repeated epistemic highs offered by the application of WWFD to the task of constructing (yet another) non-provenanced artefact as a forgery.