Category: Linguistics

(Originally published at The Growlery, 2008/08/21)

Prior to writing my previous post about Jerome Eisenberg’s conclusion that the Phaistos Disk is a recent forgery perpetrated by its excavator, Luigi Pernier, I unfortunately did not have access to the original article in Minerva magazine in which Dr. Eisenberg announced his findings (Eisenberg, Jerome M., ‘The Phaistos Disk: one hundred year old hoax?’, Minerva, July/August 2008, 9-24). Happily, once he found my post, Dr. Eisenberg commented on it and later sent me an electronic copy for my consideration. I can now report that while I previously thought I knew a lot about the disk, I now have a much better knowledge of the disk and the nature of the hoax claim. Unfortunately, I remain unconvinced, even though I admit that I do want to believe the hoax claim, but I think that the evidence from the sign-forms just isn’t strong enough, that it relies on unproven visual similarities to too great a degree. Let me explain what I mean.

In the previous post, I focused on Eisenberg’s evidence from a) the uniqueness of the artifact’s manufacture, which is unlike the Linear A tablets; b) it uses ‘movable type’ of which no other example has ever been found; c) the idea that Luigi Pernier’s rivalry with Arthur Evans would lead him to do this. In dealing with the first two, I pointed out that comparisons with the Linear A tablets aren’t necessarily that useful if in fact the PD was part of a highly specialized text genre – i.e. it would be like using a monumental inscription to proclaim a handwritten note to be a forgery. Of course we don’t know that it’s part of such a genre, or indeed what genre it could have represented at all – hence the mystery.

The one thing I didn’t focus on is the sign-forms or graphemes on the Disk. In fact, Eisenberg spends a good deal of his paper looking at resemblances between Phaistos signs and signs on other inscriptions from the ancient world in order to assert that the latter formed the models on which Pernier based his forgery. In particular, he aims to show that there are similarities between the Phaistos graphemes and authentic artifacts made much later, but that were known to 19th century archaeologists / epigraphers and thus could have been known to Pernier.

This is an unusual line of argument; it is in fact a sort of cousin to the standard techniques by which experts on scripts postulate cultural borrowings from one society to another. If we have a 10th century BC Phoenician inscription and a very early, 8th century BC Greek inscription that use many similar letter-forms, we make the reasonable inference (all right, it is more complex than this, but you get the idea) that the Phoenician script is ancestral to Greek. In particular this is the case because there is known cultural contact (e.g. trade) between the two societies, and more importantly, because there is not just a graphemic similarity but also a phonetic similarity – the signs don’t just look the same but they have the same / similar sound-values. What Eisenberg is doing, effectively, is turning these resemblances on their heads. If there are similarities between the PD signs and known inscriptions from elsewhere, then those inscriptions may have acted as a model for the forger. If the inscriptions are later in date than the PD, Eisenberg argues, it is far likelier that these artifacts served as a model for the disk’s forger than that the Disk script served as a model for the later artifacts. Similarly, if the PD shows influences from several different regional styles, this suggests that a forger just cobbled together signs from different inscriptions to make something really unique.

Now, the reason I’m unconvinced is that I just don’t think the similarities bear up, and that even where they do they don’t point unequivocally to a hoax. For instance, let’s have a look at Phaistos sign 03:

Now, this is seen by Eisenberg as being modelled after an 18th Dynasty Egyptian wall painting (16th century BC) in which the figure, a Cretan captive is facing the other direction, has extensive facial features, has hair (long, flowing hair), and has a torso with arms. The only major similarity is the two circles on the face. But this seems to go directly against the notion of the Disk as a hoax; the time is right, the captive is Cretan, so the most parsimonious explanation is that they are both genuine representations of some sort of facial decoration (indeed, as Eisenberg suggests, it may be a Cretan ‘double earring’). But, writing, “It was certainly derived from the wall painting”, Eisenberg proceeds to write as if it is now a given that Pernier did, in fact, use this as a model for sign 03 (Eisenberg 2008: 17).

When we get to one of the more unusual characteristics of the Disk – the presence of five hand-incised dots on each side of the disk, and ‘word-separating’ vertical lines – I’m in my element, because these, Eisenberg sees as being modelled after the Cretan five dots = the numeral 50 and vertical bar = the numeral 100. This is dangerous territory though – dots and lines are ubiquitous in scripts and numerical systems. And are we really to believe that Pernier needed a model to think of the idea of adding bars and dots to a forgery? These are stylistic elements found in virtually any script worldwide, and are not indicative of anything. One of the real problems with the study of writing systems is the assertion of cultural relationships based on passing visual similarities, and one of the things that we do not yet know how to do well is to know how similar two graphemes must be before a claim of diffusion can be sustained. This is the same sort of reasoning used to argue for a hoax in this case, and ultimately its inclusion greatly weakens Eisenberg’s argument, and made me look much more critically at the remainder of his claim.

But the heart of the issue is that Eisenberg is working at cross-purposes here. On the one hand, he wants us to believe that the Disk is so unique, so different from other inscriptions that it cannot possibly be genuine. On the other, he wants us to use evidence of similarities with known scripts as proof of ‘forger’s models’. While a hoax can, of course, be both unique and based on models, we’re left with the impression that virtually any similarity or difference can be evidence of forgery, and that just isn’t sound argumentation. So I’m not convinced. I do still think the idea is worthy of consideration, and I do think that it is worth trying a thermoluminescence test, not only because it can settle the hoax issue but also because it can resolve the question of the artifact’s age even if it turns out to be genuine. In this respect, I believe that Eisenberg and I are in full agreement.

In conclusion I want to thank Dr. Eisenberg for sending me this paper, and also for inviting me to the upcoming International Conference on the Phaistos Disk, which unfortunately I am unable to attend due to my new work commitments. It does highlight however the real value of blogging as a means of social interaction and information exchange.

But of course the real question remains unanswered: should it be disk or disc?

(Originally posted at The Growlery, 2008/08/04)

The Phaistos Disk is one of the more enigmatic and bizarre artifacts in the field of ancient writing systems. Found in Crete in 1908 by the archaeologist Luigi Pernier and associated archaeologically with the Minoan civilization (dating to roughly 1850 – 1600 BCE), it remains completely undeciphered and has no obvious connection either to the Minoan (Linear A) script or to any other known script, deciphered or otherwise. Now, a very notable claim has been made by the American art historian / art dealer Jerome Eisenberg, an expert on forgeries, that the Disk was in fact an elaborate hoax constructed by Pernier himself, which Eisenberg has published in his own magazine, Minerva (Eisenberg 2008).

I’m not an expert on Minoan writing by any means, but my scholarly focus lies heavily in the study of ancient scripts and the anthropology and archaeology of literacy. I use Yves Duhoux’ hilariously entitled ‘How not to decipher the Phaistos Disc’ in my course on the anthropology of literacy (Duhoux 2000). Moreover, the century of scholarship on the Phaistos Disk is legendarily riddled with cranks, frauds, and loons, and as I have more than a passing interest in pseudoarchaeology, Phaistos-related material is of ongoing interest to me. Honestly, it would make a lot of things a whole lot simpler if we could just deny the disk’s authenticity – but this is no ordinary hoaxbusting exercise, and the importance of the artifact demands that we give the claim close scrutiny.

Phaistos Disk, Side A. Source: Wikimedia Commons.

Before we get to the Eisenberg claims, we need some context. So, firstly, what do we know about the PD?
– It is a fired clay disk, roughly 16 cm in diameter and 1-2 cm thick.
– It was found at a Minoan palace site at Phaistos in southern Crete.
– It appears to have been fired intentionally (with care) to produce a permanent record, whereas other Minoan documents were fired accidentally (e.g., when buildings were burned).
– Glyphs are stamped on both sides using distinct punches or stamps, not carved/incised into the clay.
– It has 241 signs in total, consisting of 45 distinct characters / glyphs. However, the total ‘signary’ (all the signs in the system) was probably greater, since some rare signs almost certainly do not appear in this particular text.
– The ‘text’ is divided into 61 sections of up to 5 characters apiece.
– The prior two facts suggest that it may have been a syllabic writing system, recording syllables rather than single phonemes; it has too many unique signs to be an alphabet but too few to be a logographic (word-signs) or some other sort of system. However, this does not rule out the possibility that it was not phonetic writing at all (e.g., if it was a calendar or a game).
– Because the signs in the centre are slightly compressed, it seems to have been written from the outside spiralling inward.
– Judging by the overlap in some signs, it was stamped/written from right to left, suggesting that that is how it was meant to be read.
– There is no useful resemblance of the glyphs to those of any other writing system in the Mediterranean or elsewhere, although it was found in close association with a Linear A (Minoan) tablet.
– Its date is established solely through its archaeological context, and while the early second millennium is the most likely period, it may date as late as 1400 BCE.

Now, on to Eisenberg’s paper. The first thing worth noting is this is not a peer-reviewed academic venue, and the author is the founder, editor, and publisher of the magazine. A better analogy would be to think of it as editorial opinion. It also is not the result of any particular new research undertaken by Eisenberg or anyone else. In fact, as seen in the comments here, Dr. Eisenberg has been making this claim for nearly a decade, and there is no new evidence that demonstrates the likelihood that it was a forgery. Pernier, the artifact’s excavator excavator, is labelled as a forger, not on the basis of any particular evidence, but has simply been ascribed motives (rightly or wrongly) that might lead him to falsify the document. So we don’t have anything like the revelations in the early 1950s that debunked the Piltdown hoax on the basis of physical or chemical analysis; neither do we have the spectactular video evidence that revealed Fujimura Shinichi planting fake discoveries at his sites in the Japanese Paleolithic hoax in 2000 (Hudson 2005). It is a highly circumstantial case. It is nonetheless one that ought to be vetted seriously, both because it is plausible on its face and because Eisenberg has been responsible for several other (much more solid) hoax-busting episodes over the past few decades.

The starting point for Eisenberg’s claim of a Phaistos ‘hoax’ is the uniqueness of the artifact, both the object itself and the writing on it. Given that no other examples of this form of writing have been found, it is striking (pun intended) that its creator would have made 45 distinct seals to stamp into the clay rather than simply incising the signs as necessary. No actual stamps/seals resembling the signs have been found, either, suggesting that this early instance of ‘movable type’ was used to create only one artifact, and then the process was abandoned entirely. In his popular Guns, Germs and Steel, the evolutionary biogeographer Jared Diamond (1997: 239-259) asserts that the PD was indeed a very early and remarkable example of movable type, but one that could not be exploited by the Minoans because in other respects their society lacked the technology and organizational expertise to develop it further. Eisenberg’s perspective is different – he argues that the uniqueness of the artifact’s medium suggests that it is a hoax, designed by Pernier to intrigue and mystify other scholars and to boost his own prominence, and that of Phaistos, in relation to his rivals (particularly Arthur Evans).

The PD is a singular artifact and a very short text, making it literally impossible to decipher unless more examples of the writing system are found. Yet John Chadwick, whose career was built upon his work with Michael Ventris in deciphering the Mycenaean Linear B script (Chadwick 1990), was plagued by purported Phaistos decipherers and purportedly received one new solution per month; there is a fairly thorough list of purported decipherments in this Wikipedia article. Basically, every remotely plausible script tradition has been claimed as an influence, and the disk itself has been asserted to be in languages ranging from Greek to Egyptian to Basque to Atlantean (!!!). Alternately, it has been suggested to be a game board, a calendrical document, or some sort of mystical text. Unless more documents in the same script are found, no one is going to be able to resolve the matter definitively. If it were confirmed to be a hoax, however, everyone could just stop looking. Eisenberg is suggesting, in effect, that the futility of the search rests in part on Pernier’s ingenuity in creating such a mystery.

The crux of Eisenberg’s argument, however, lies in the physical properties of the artifact: the fact that it was very carefully, intentionally fired, and that it has a very cleanly cut edge in comparison to other Minoan clay tablets, and here, he finds fault with Pernier. Because it is so different from other Minoan clay artifacts in this regard, this sends up a red flag for Eisenberg suggesting that its uniqueness may be due to Pernier’s ignorance of these facts. The counterargument to this, however, would be that while Minoan clay tablets with Linear A writing are all economic documents not intended for long-term archiving, the PD, if ancient, is almost certainly of a very different textual genre and script tradition than these texts. This doesn’t disprove the notion that it may be a hoax, but neither does it act as substantial confirmation. For instance, if the disk is a gaming board, a calendar, or a devotional inscription, its makers would have a good reason to fire the clay at the time of manufacture, and a potentially good reason to cut its edges so cleanly. It simply was not the same sort of text as the copious clay economic documents. We need to answer the question, “Could the Minoans have chosen to preserve some forms of information permanently and not others?”

One potential resolution to the mystery lies in its dating. The artifact has never undergone any sort of radiometric dating, and indeed for most of the past century could not have been dated except through archaeological context, as discussed above. However, thermoluminescence dating allows archaeologists to non-destructively determine the date when clay was fired, and if TL dating were used on the disk, one could find out if it was truly of ancient manufacture. Yet this test has not been permitted by the museum that holds it (in Heraklion, Crete), because, Eisenberg claims, “no Greek scholar or politician would dare to help ‘destroy’ such a national treasure”. This is unfortunately true; museums are rarely open to this sort of inquiry, even from major scholars. Archaeology is frequently tied up in nationalistic fervor and institutional pride, and the failure to undertake a standard, well-accepted test will haunt the study of the Disk from now on, now that the claim has been made so publicly. Thus, I regard Eisenberg’s public claim as a valuable stimulus, hopefully forcing the issue of the thermoluminescence dating. It would also be highly informative even if the PD proves to be ancient, because the TL could establish whether it was an early second millennium artifact (1800-1600 BCE) or more in the range of 1400 BCE.

Ultimately, this is suggestive, and I would not exactly be astonished if Eisenberg’s claim were to be verified, and if the PD turned out to be a fake, but I cannot agree that the matter is now settled. Because literacy is not simply an ‘on/off’ phenomenon – we must deal with the possibility of different text genres, different media, and different purposes for writing – we can’t use the Linear A clay economic documents to prove the disk’s anomalous nature. A date from an independent lab would go a long way toward resolving my doubts. This would still leave the question of how it was done and by whom – remember that there is no direct evidence against Pernier. However, I for one look forward to this claim receiving greater attention over the next couple of years.

Chadwick, John. 1990. The decipherment of Linear B, 2nd ed., Cambridge University Press.
Diamond, Jared. 1997. Guns, germs and steel: the fates of human societies, W.W. Norton.
Duhoux, Yves. 2000. ‘How not to decipher the Phaistos Disc: a review article’,American Journal of Archaeology, 104, 3, 697-700.
Eisenberg, Jerome M. 2008. ‘The Phaistos Disk: one hundred year old hoax?’, Minerva, July/August, 9-24.
Hudson, M.J. 2005. ‘For the people, by the people: postwar Japanese archaeology and the Early Paleolithic hoax’, Anthropological Science, 113, 2, 131-139.