The Lost Books of the Odyssey might have remained an enigma of history had it not been for Dr. James Stryszinksi. A Columbia-educated computer scientist specializing in cryptography, Dr. Stryszinksi had worked for the NSA since 1993 on projects related to the intersection of information theory, data compression and linguistics. His brief was to develop mathematical methods for detecting linguistic communications, possibly encoded, and, in many cases, with no lexicons or grammars available for the languages in question. Essentially, he had to find computationally detectable signs of linguistic structure in the countless ghostly voices and data transmissions acquired by the NSA's eavesdropping nodes.
Dr. Styrszinksi's project was a success--the mathematical and computational tools that he developed (which are far beyond the scope of this volume and in any case are still largely classified) numerically estimate the density of semantic information in arbitrary signals.1.1 For a while he happily optimized his programs to separate the static-filled walkie-talkie conversations of Islamic radicals in the Hindu Kush from the background radio signal generated by the Aurora Borealis, but as the technology matured he looked for other applications for his work. He found them, to his surprise, in the humanities, where there is a ready supply of untranslatable, encoded and determinedly cryptic historical documents.
At the time I was a Professor of Modern Languages at Trinity and although my training is philological rather than mathematical I had won some small fame in cryptographic circles. While doing a master's thesis on Talleyrand, I had immersed myself in the Louvre's collection of his papers and among them found a piece of stationery covered with clusters of what looked like random alphabetic scribblings, rather as though someone had been trying out a new pen. The remarkable thing about it was that Talleyrand had filed it with his most important papers. Previous scholars had given it a look now and then but it was generally assumed to be there by mistake. The groups of letters put me in mind of a code. I spent a few days trying various simple cipher keys and on the third day was immensely gratified to unlock a message in Latin.
The contents turned out to be nothing less than a list of questions for Talleyrand, dictated by Napoleon himself, on the likely diplomatic aftermath of a successful invasion of England. The letter, once considered worthless, recently sold at Christy's for US$262,144. No doubt others could have cracked the code as fast or faster, but no-one else had tried.
Inspired by success, I took up the study of archaeo-cryptographical documents. There was no shortage of material--untranslated and possibly untranslatable historical documents abound, from an anomalous fourteenth-dynasty Egyptian hieratic stele to the enigmatic Linear A tablets to the indubitably fake Codices of Mu. Some documents are in dead languages, others in code, others vacuous. I realized a few more successes, mostly with World War II era material; most notably, I decoded letters showing that in 1933 the Germans had sent an agent to England with orders to seduce and subvert computer science pioneer Alan Turing. (The agent was, in the event, unsuccessful.)
When James contacted me in hopes of finding a challenging application for his algorithms the first thing I thought of was The Lost Books of the Odyssey. This book had a been a tantalizing mystery for millennia, and although in the modern consensus it was regarded as fraudulent, I thought it would be worthwhile to prove it once and for all. I duly faxed him my mimeograph of the British Museum copy of the Lost Books (which is itself an early Renaissance copy of a (most likely) Roman copy).
The Lost Books, dating from classical Greece or earlier, is traditionally attributed to the Homerids, epic poets who claimed to be descended from Homer. Since antiquity the book has been known only in encrypted form, the code having resisted all attempts to crack it. Tbe book is mentioned here and there in the texts of antiquity; Herodotus wrote that the Lost Books contains a multiplicity of alternate histories of the Trojan War, most of them centered on the adventures of Odysseus, and the Greek historian Arcesius wrote that it consists of short episodes from Odysseus's life, the point of which is less to advance a plot than to take one image or theme and, paring away all that is inessential, present it with the greatest possible concision and clarity. Besides the assurances of the classical authorities, though, there was no reason to believe that the book was anything more than hundreds of pages of random Greek letters.
Scholars throughout history have had an ambivalent relationship with the Lost Books, drawn to it but suspecting an ancient hoax perpetuated by credulous bibliophiles. Enesech of Syracuse, a first century A.D. head of the great library of Alexandria, says that it is a venomous asp coiled in his collection, but one he spares, however grudgingly, for the sake of the gleam of its glorious scales.
I was skeptical when, five days after receiving my parcel, James called to say that his algorithm had showed that, with very high probability, the Lost Books encrypted meaningful linguistic information. Not only that, but the code had certain signatures characteristic of fractal compression techniques and some kinds of baroque music.1.2 Although this proved (or nearly) that there was meaning locked within the Lost Books, it told us nothing about the key. James said he had an educated guess about the class of encryption algorithm that had been used, but that if his guess were right decryption would be effectively impossible without the code key, no matter how many super-computers were thrown at the problem.
It was at this point that I played my most substantial role, perhaps even a pivotal one, in the history of the Lost Books. Every scholar dreams that his hard-won erudition will, in a flash of insight, furnish the key to some ancient riddle. In this regard, I was fortunate to have read the works of the thirteenth-century Catalan Raymond Lully, a kaballist and mathematician. Lully is historically noteworthy for inventing a kind of precursor to combinatorial mathematics, although the application he had in mind was more divinatory than scientific--by mechanically combining elements of a list of God's attributes, he claimed to be able to reconstruct some aspects of the divine thoughts.
In a coda to Ars Magna, his great work on combinatorics, Lully mentions the Lost Books, a copy of which is known to have been in his library, and which he claims to have read using his new science. That Lully purports to have decoded the Lost Books is, in archaeo-cryptographic circles, widely known and little thought of--the history of the Lost Books is littered with such claims, all unverified, most absurd. I recalled Lully's boast because in his gloss of the decryption he mentions the eleven keys that unlocked the book. These sub-sidereal attributes, analogous to but lesser than the divine ones, are: Time, Memory, Desire, Revenge, The Gods, The Dead, Departures, Returns, Words, Deception and Doubles.
I emailed Lully's list of keys to James and he set up a program that tried them in a variety of plausible cryptographic frameworks. I will not dwell on the emotions of that time, the personalities involved, or the gratifying attention our work soon received. Suffice it to say that we successfully decrypted the Lost Books. Not only that, but that where James had been looking for just a single functional combination of keys, he found an ordered set of forty six.1.3 Each combination, consisting of between one and four of Lully's keys, unlocked a new chapter thematically related to the keys that had produced it.
The revealed contents of the Lost Books were full of surprises. The chapters vary radically in style and tone, which suggests a multiplicity of authors, for all that certain tropes and phrases recur across otherwise quite dissimilar stories. Based on internal textual and linguistic evidence, most chapters can plausibly be dated to between 1200 and 800 B.C., as would be expected,1.4 but there are exceptions, some of them extreme. The chapter Bright Land is clearly much older and is, in fact, in a language so ancient it is hardly Greek. Another chapter is about the death of Alexander the Great and as such can be dated no earlier than 323 B.C. Interestingly, this is not long after the usual estimate of the date of composition of Aristotle's Poetics, which includes a reference to the Homerids and their secret book.
The most surprising chapter is Record of a Game. Though written in very plausible Homeric Greek and thematically consistent with the rest of the book, this chapter makes reference to England, India and a relatively modern version of the rules of chess--on these bases, it can be dated no earlier than the Renaissance. If the Lost Books were an ordinary book it would be natural to assume that this chapter had been tacked on by a clerk or bookseller with more inspiration than reverence, but Record of a Game is encoded in the same way as the rest of the book. We can only conclude that the encryption techniques the Homerids used to conceal their book were passed down to, or rediscovered by, someone who lived eighteen hundred years after the Homerid tradition is generally believed to have been extinguished, and that this brilliant hypothetical scholar left no other outward sign of his breathtaking discovery. Raymond Lully would be a compelling candidate but ``Record of a Game'' is not consistent with the many surviving examples of his writing, and chess-historically speaking the story must have been written at least two hundred years after Lully's death.
The dates of composition were not the end of our perplexities. Even decoded, the chapters of the Lost Books are highly ambiguous and call for substantial interpretation. They are written not in ordinary language but in a compressed, grammarless shorthand consisting of sequences of noun/modifier pairs. What's worse, each word is specified using just three letters, from which the full word must be inferred based on context. There are 24 letters in the Greek alphabet, for a total of 13,824 possible three letter combinations, though only 2,179 occur in the text. Each three letter sequence encodes for, on average, 7.2 distinct words. The translator's perplexities are compounded by the absence of spacing and punctuation.
The reader unacquainted with Homeric Greek would find a raw list of three-letter sequences unilluminating. I therefore present the first sentence of one of the chapters in an expanded but uninterpreted form:
Ithaca arrived, rage trembling, spear poising, heart piercing, road obtruder.
I translate this as:
Odysseus set foot on Ithaca trembling with wrath, his spear poised to fly through the heart of the first man unwise enough to cross him.
This is one of many possible translations. Below is an alternative that, were it not for the context, would be equally plausible:
The day he came back to Ithaca, Odysseus found a man standing in his way, atremble with anger, his spear at the ready, so Odysseus ran him right through the heart.
Even with context, the absence of grammatical structure makes interpretation difficult. Fortunately, some reliable heuristics have emerged--for instance, in the absence of contradictory information, the subject of each sentence is generally Odysseus,1.5 and certain epithets occur only with a certain characters. For all that, my translation relies heavily on linguistic and cultural considerations, a careful reading of the core epic and a general anticipation of the thematic content of each story.
The Lost Books is legible for the first time in centuries but some mysteries remain. Even though the technique for decrypting the Lost Books is understood, no-one has yet discovered how to do the corresponding encryption. (This is counter-intuitive but I am told that the techniques for encryption and decryption are not always the mirror-images one might expect.) Also, James tells me that the book's information density is absurdly high, almost impossibly so. More precisely, the encoded text of the Lost Books is not big enough (in information theoretic terms) to hold the chapters we have extracted from it.
This anomaly has excited considerable debate and given rise to three interpretations. The first is that the Lost Books is a peculiar mathematical kind of literature in which the elements of one story are constrained to be used in the other stories in a formally specified way. Consider a crossword puzzle in which ``homer" is a horizontal entry--its five letters appear in as many as five vertical entries, whose letters then constrain yet more horizontal entries, and so on. The proposed mechanism underlying the Lost Books is analogous, but more complex, encompassing both the lexical and semantic levels. This theory predicts that most chapters will be recombinations of the elements of other chapters in the collection and that there will in general be a great economy of words, tropes and symbols. This is the theory that we favor, as it requires no extraordinary assumptions and the Lost Books amply bears out its predictions. The difficulty is that composing story collections under such constraints is very hard--so hard, in fact, that no (modern) mathematician has been able to devise an algorithm for doing so, though the area has recently been the focus of intense research activity.1.6
The second interpretation is that the Lost Books is an illusion. In this view the encoded text is the product of an inspired mathematical exercise by some pre-Pythagorean numerological mystic who used an algorithm to subject the traditional Odyssey to a complex but ultimately vapid distortion. The unlocked ``chapters'' are essentially meaningless--the meanings we purport to find are the product of over-interpretation and of chance. The proponents of this view contend that the Lost Books were composed not by bards thousands of years dead but, thanks to the universal human capacity to see form in purest entropy, by me.
I can not take a fully objective position on this matter. It is not even clear which position would be the more modest one to take. I will say that I am in essence a pedant, a sifter of old books who has not before this been accused of originality, and that in the Lost Books there is sometimes a grace that my own writing has never acheived. I do, however, thank my academic rivals for their high opinion of my creativity.
The third interpretation, which it would be ungenerous to call fanciful, is from outside the academy. It is espoused by the Theosophists, a sort of New Age religious society. The basis for their view is in the writing of the minor nineteenth century classicist Cyrus Maurer, who published what was first a famous and then a famously fraudulent translation of the Lost Books. He wrote that Homer was neither a bard nor a bardic society but, rather, a mathematician. Maurer contended that Homer's works are a priori literature, in the sense that they can be generated by the repeated application of a few simple mathematical operators. According to this theory, The Lost Books of the Odyssey can be derived from mathematical first principles in the same way as, say, Fermat's Last Theorem.
The Theosophists have been trying to discover these literature-producing operators for most of the twentieth century, so far without any success, despite the fact that some of the participants in this seemingly marginal enterprise have been truly eminent--most notably, Kurt Gödel, of the famous Gödel's Incompleteness Theorem, worked with them for some three and a half years. They argue (in a monograph entitled ``Maurer and the Inevitable Literature--the Hieratic Theorems of the Homerid Sages,'' available through the Theosophical Society's press) that the high information density of the Lost Books proves that it is a product of pure analysis rather than mere human artifice.
A consequence of the Theosophist position is that the Lost Books is not just some book, but an ideal book, whose text is woven into the very fabric of mathematical logic. Their view is not without precedent--Muslim theologians contend that the Q'ran is prior to the universe and that its text is a direct emanation of God's will, just as students of the kaballah believe that there is nothing accidental or contingent in the Torah.
There have been many publications on The Lost Books of the Odyssey in the learned annals of philology, cryptography, archaeology and the history of mathematics, and I refer the reader to these literatures for a thorough examination of the issues I have here only touched on. This volume is intended not for professional academics but for lay readers--the translation is liberal, prioritizing readability and perceived authorial intent over literality and pettifogging linguistic considerations, and uses no more footnotes than are necessary for an appreciation of the text. I hope the general reader will forgive the occasional anachronism and that my colleagues will forgive my license.
During my stint at the Sorbonne I was befriended by a historian of the French Revolution who had one of the best wine cellars in France. Among his many excellent specimens was a bottle of '03 Moutons et Croix bordeaux, a famous vintage of a famous wine, of which there were only ever a hundred and three bottles. Even at the time it was laid down it was considered a great wine and contention for the bottles was fierce (there is a book that describes the circumstances of the consumption of 97 of them). His was the last bottle of its kind, and we brought it up from his cellar on a drizzly May evening in 1987. My friend was near tears as he uncorked it--he explained that the mystery and promise of that vintage would, within the hour, disappear from the world forever. I feel a similar melancholy now that the Lost Books has been revealed, that the mystery, secure for ages, is gone for good, that I will never read another new word of Homer's, and after finishing this book, reader, neither will you.