Phallus and fallacy

The Phaistos Disk (PD) is a thick, baked-clay object, about 16 cms in diameter; it was found in 1908 by the Italian archaeologist Luigi Pernier in the Bronze Age ruins of a Cretan site. Dateable to about 1600bc, it had been impressed on both faces with a series of discrete signs, most of them pictographic, arranged in inwardly curving spirals - the world's first use of "moveable type". Until recently, attempts to decipher this unique script, so far confined to the disk, have met with failure.

On the other side of the globe, on the incredibly remote Easter Island (or Rapanui), the Polynesian inhabitants had created another enigmatic script, Rongorongo, which has long defied analysis by scholars from many countries. First described by outsiders in the mid-19th century, and subject to the destructive effects of cultural subjugation, missionary zeal, and the iconoclasm of the Easter Islanders themselves, this highly stylised writing system - which in appearance recalls the cipher of Sir Arthur Conan Doyle's "The Dancing Men" - is preserved on various wooden objects in collections scattered around the globe.

Steven Fischer, director of the Institute of Polynesian Languages and Literatures in Auckland, claims to have cracked not only the Phaistos script but also Rongorongo. In his biographical account Glyph-breaker, he says: "No one had ever deciphered two wholly different historical scripts before. Until now. This is the story of those two achievements."

When we decide that a script, say ancient Egyptian or Maya, has been "deciphered", what exactly do we mean? What does Fischer mean by "decipher"? Has he in truth produced a global decipherment of these scripts as, say, Champollion did with Egyptian hieroglyphs, or Ventris with Linear B? A generally accepted use of the term "decipherment" is given by Maurice Pope in his authoritative book The Story of Decipherment (1975): "the explanation (by transliteration or otherwise) of the individual signs of a script", and he specifies that the term does not mean understanding the sense of particular texts written in it; it does mean the matching of each sign to a definite sound or word in the ancient language. Thus, Etruscan has been "deciphered", yet we know little of the grammar or vocabulary of the language and hence of its written texts. And under such a definition, while we have a good grasp of the meaning of most Maya texts, only about half of the script has actually been "deciphered", as work proceeds on the logograms and phonetic-syllabic signs that compose it.

Etruscan, of course, is a special case, since the alphabet in which it was written was very close to that of early classical Greece. But all of the demonstrably successful whole or partial decipherments of ancient scripts seem to have met most or all of the following criteria:

* the database must be substantial, with a large number of texts which encode complete sentences

* the script should encode a known language, or at least one reconstructable from extant relatives or daughter languages

* there should be one or more bilinguals, one member of which is in an already-known and readable script (the Rosetta stone, and Bishop Landa's exposition of the Maya script, are outstanding examples)

* if the script is logographic or logo-syllabic, there should be extensive pictorial references to apply to the texts (as in Egypt and the Maya area).

Now let us see if Fischer's claim to have singlehandedly "broken" both the Phaistos and the Rongorongo scripts has any foundation.

Discounting two additional objects which may or may not have a few Phaistos-like signs on them, the PD is the only extant member of its class; with 122 signs on one side, and 119 on the other, this is not a very large database, to put it mildly. In contrast, the Linear B corpus, comprised of many hundreds of clay tablets, totals no fewer than 57,000 signs, yet it still gave Alice Kober, Ventris, and others considerable trouble to "break". I have said that the PD texts are laid out in spiral fashion, and are apparently to be read from right to left, from the outside to the centre, with vertical lines enclosing from one to seven signs within "fields" (Fischer differs from other scholars in not thinking that these "fields" define words). About half of the signary looks like pictures of real things: human heads and bodies, animals and parts thereof, a ship, a hafted axe, etc.

Fischer's first move was to convert these texts into the numbers assigned to the PD signs by Arthur Evans in his Scripta Minoa I (1909). Then he began to search for repeating structural patterns which might reflect actual grammatical patterns, a not-unreasonable procedure, for it was just such a methodology which led Kober to positing - always with commendable caution - that Linear B was recording an inflected language, a breakthrough which eventually resulted in Ventris's epochal decipherment. But let us keep in mind that Fischer has been working with a database that is less than 1 per cent of what Kober and Ventris had at their disposal. There are, it is true, a few such patterns discernible on the PD: for instance, a head with a "Mohican" hairdo (see red box on illustration), followed by a dotted disk, is often in an initial position on Side A, and once on Side B. Other repetitive sequences are less obvious.

Fischer is on surer ground in his assertion that the Phaistos script must be a syllabary. The size of the signaries of purely syllabic scripts vary between 40 (for Persian), and 87 (for Linear B); the upper limit for alphabetic systems is 36 (for Russian); while the lower limit for logographic and logo-syllabic scripts is around 500 (Maya, for example, has over 800 hieroglyphs). In the complete Phaistos system, there must have been more than the 45 discrete signs that appear on the PD, but probably not more than is normal for a syllabary of the consonant-vowel (CV) sort, which this in all likelihood was.

Now, if it is a syllabary, reasons Fischer, then it must be cognate to other early Aegean systems, such as Minoan Linear A, Mycenaean Linear B, and the Cypriote script (a written form of Greek, which he reasonably calls "Linear C"). Here he moves into high gear with several untested (or untestable) assumptions: * the known values of Linear B, which encodes an early Greek dialect, can be applied to "Minoan" Linear A, just as Linear C values were once successfully applied to Linear B (in fact, Ventris showed that this was far from the case)

* Linear A thus also encodes an early form of Greek

* 18 signs on the PD can be shown to be cognate with Linear A signs (to Fischer's satisfaction, at any rate, if not to this reviewer's), and thus these Linear A values can be applied to the PD cognates.

Ergo, the texts on the PD must also be in an early Greek dialect.

Further values for the PD signs can be derived on the acrophonic principle, that is, the sound values for some signs represent the initial sound of words for the objects they pictorially represent (assuming that these objects from the real world have been correctly identified).

Fischer's task was now to establish a syllabic grid based on these values, and see where it took him when he substituted his sound values for Evans's numbers.

And lo and behold! Fischer now has complete, readable texts in an ancient form of Greek. While I am certainly not a Hellenist, the translations resulting from his alleged decipherment of the PD strain credulity. There are passages that bring to mind the 17th-century polymath Athanasius Kircher's vain and usually ludicrous attempts to decipher the texts on Rome's Egyptian obelisks. Here is one such passage by Fischer:

"Safeguard me, Idaians: I am sore afraid. Loose me now. My night, my great: Ye loose me now. These afflictions so terrible and so great, verily so molestful: Ye loose me now. Down to the sea, everyone! Yea, deliver me of my great afflictions!" By piling one assumption on another, Fischer has, I am afraid, been led into the same morass as Kircher had stumbled into three centuries ago. His attempted decipherment of the Phaistos Disk has, by his own admission, failed to convince John Chadwick, the leading student of Aegean scripts, and it has left me with grave doubts whether he or anyone else will ever "break" this script. It simply meets none of the criteria for a successful decipherment. In his relentlessly self-celebratory Glyph-breaker, Fischer, like Kircher before him, has talked himself into believing in an illusory success.

While the second part of Glyph-breaker covers his involvement with the Easter Island script, this system receives far more extensive and serious treatment in his massive Rongorongo, which supersedes all previous books on this fascinating subject. Here, Fischer's most valuable contribution is a detailed history of research on Rongorongo (including the wild imaginings of the lunatic fringe), since its discovery in the mid-19th century until today, and a full catalogue of all known Rongorongo inscriptions. These alone will make this book a must, not only for Polynesian specialists, but for all those interested in non-European scripts.

Unhappily, however, Fischer's is not an entirely unbiased account, since it is coloured throughout by two significant preconceptions: 1) that he has dated the origins of Rongorongo and 2) that he has actually deciphered it, and thus understands its character. I will now deal with the problem of origins. Archaeological research has shown that Easter Island has been inhabited since about ad400, but, asserts Fischer, there is not the slightest evidence that the islanders knew anything of writing until November 20 1770, on which date some of their leaders, having already watched the Spanish officers sign an official Spanish document of possession, were coerced into putting their own marks to the document. Inspired by this example, once the Europeans had departed, they proceeded to elaborate their own script.

Yet in a letter of 1864, only 94 years later, the missionary Eug ne Eyraud states there were "hundreds" of inscribed wooden tablets on Easter Island. Those handful (21, or 25 according to Fischer) that survived the great period of destruction of the next half-dozen years are inscribed in a sophisticated system that shows little inner variation and an extraordinary degree of uniformity; Fischer ascribes this to the establishment of scribal schools led by Rongorongo experts. But when Eyraud was on the island, and when Tepano Jaussen, bishop of Tahiti, first learned of the script in 1868, knowledge of it had virtually died out. Fischer would thus have us believe that a writing system of this complexity had sprung up by stimulus diffusion, been codified, flourished, was taught in special academies, and subsequently died out within a span of less than 100 years. Not a very likely scenario! Perhaps the islanders (not always noted for veracity) were nearer the truth when they asserted that Rongorongo had been brought to Rapanui in the distant past by their legendary culture hero, Hotu Matua.

Fischer's alleged decipherment of Rongorongo was announced in 1996, and enthusiastically endorsed by archaeologist Paul Bahn in that same year in the journals Nature and New Scientist. Here, in a nutshell, is what it is. In 1886, William Thomson, paymaster of the USS Mohican, spent ten days on the island, gathering a great deal of very useful ethnological information, including a transcription of a chant (Atua Mata Riri) sung by a Christian convert named Daniel Ure Va'e Iko. Almost all of its 48 verses describe the copulation between two entities, to produce a third entity; the entities might be natural species like sharks, birds, plants, and the like. The verses always take this triplet form: "X copulates with Y, and there issues forth Z."

Now, on the Santiago Staff - one of the longest and best preserved of the Rongorongo inscriptions - there are frequent sign triplets. These always begin with an anthropomorphic figure (let us label this "X"), with a large extension curving up from the figure's left shin. Fischer identifies this extension as an erect phallus (although the resemblance escapes me). Then follow another sign, Y, and a third, Z; Y and Z are different in each case. Having identified these written triplets, and the objects supposedly portrayed by the Y and Z signs, Fischer moves ahead and applies the Atua Mata Riri procreation chant to the inscription, to produce this supposed translation of one passage in the text: "All the birds copulated with the fish, there issued forth the sun".

According to Fischer, much of the Santiago Staff text is in this vein, as are those on many other tablets (even though the alleged phallus is missing in these).

If Fischer's methodology is valid, is this even a "decipherment" in the Maurice Pope definition of the word? I think not, as there is no proof that any sign in Rongorongo can now be matched with a specific sound or word in the Rapanui language. Fischer claims that most Rongorongo signs are logographs (with the "phallus" being a silent semiasogram representing an act), but since he has failed to find phonetic signs to complement, or substitute for, the logograms, we will never know how they were pronounced or even what they really stood for in utterances - chants or otherwise. In fact, if Fischer is correct in his assessment of the script, the tablets would have been, as the late Alfred Metraux always held, nothing more than aides-memoires for the chanters, and thus irrevocably undecipherable given the total disappearance of knowledgeable informants.

Is Fischer even right about his identification of the underlying meaning of most Rongorongo texts? His "decipherment" using the Atua Mata Riri chant has been subject to blistering criticism by the linguist and computer scientist Jacques B. M. Guy, on the Internet site for Easter Island. Guy's most telling point is a lapse in Fischer's reasoning, which he calls "the fallacy of the excluded middle: dogs have four legs, tables have four legs, therefore tables, like dogs, wag their tails and pee against trees". Because the Santiago Staff has signs in groups of three, and Daniel Ure's recitation has protagonists in groups of three, Fischer leaps over the middle to insist that the Santiago Staff, like Atua Mata Riri, is a story of how animals, plants, and natural phenomena came into existence.

Once again, Fischer has unhappily talked himself (and a few others) into believing in a decipherment that never was. I myself am pessimistic: I believe that the Phaistos Disk and Rongorongo will be forever mute.

Michael D. Coe is emeritus professor of anthropology, Yale University, and author of Breaking the Maya Code.