Saturday, October 21, 2006

BOOK: Andrew Robinson, "Lost Languages"

Andrew Robinson: Lost Languages: The Enigma of the World's Undeciphered Scripts. Mc-Graw Hill, 2002, 0071357432 (hc), 0965421244 (pb). 352 pp.

This is an interesting book about various undeciphered (or formerly undeciphered) writing systems. It includes the three famous examples of successful decipherment: Egyptian hieroglyphs, Mycenaean Linear B writing, and the Mayan hieroglyphs. Then there are several chapters about scripts that are still undeciphered, including both well-known and less well-known ones: Linear A, the Indus Valley script, the Etruscan alphabet, the inscription on the Phaistos disc, two pre-Maya scripts from Central America (Zapotec and Isthmian), the rongorongo writing of Easter Island, the Meroitic script (Egypt-inspired, used in Sudan in the late 1st millenium BC), and the Proto-Elamite script (the earliest writing system of Persia, just a bit younger than the early cuneiform of neighbouring Mesopotamia).

Physical aspects of this book

From the point of view of form and production, this is an unusual book — it has many elements that remind one more of a textbook than of a traditional trade publication. For example, there's the unusually wide format; the pages are almost as wide as they are tall, and much of the text is set in two columns. (I guess the width of the pages was chosen so as to accommodate the illustrations better, which is a good idea as there certainly are lots of illustrations.)

Another very unusual thing is that the whole text is set in a sans-serif typeface; I think this is the first time I've read a whole book in a sans-serif face, and I now quite agree with those who say that seriffed faces are easier to read and sans-serifs should not be used for extended amounts of text. And, most curiously of all, the book is printed in two colors: the text and illustrations are mostly in black, but occasionaly a bit of blue is used in headings and to frame or highlight some particularly interesting part of an illustration. Perhaps it was originally really intended to be some kind of textbook — after all, McGraw-Hill does also publish textbooks. But it doesn't anywhere explicitly say that it's intended to function as a textbook (nor it is obvious to me what sort of student would find this sort of book useful — it's a very fine book to satisfy the curiosity of the general reader, but a student of history or archaeology will probably want to specialize in just one of the scripts mentioned here, and will want much more detail about it than a book like this one can provide).

This particular copy that I bought has another curious aspect: it isn't the McGraw-Hill edition; it's “A Peter N. Nevraumont Book” published by “BCA” (whatever that means) “by arrangement with McGraw-Hill [. . .] Created and produced by Nevraumont Publishing Company” and it doesn't even have an ISBN, only a “CN” (I guess that would be some sort of catalog number internal to the publisher) of 106839. However, it is listed on amazon as having the ISBN 064169959X.

Successful decipherments

The chapters about the three deciphered scripts are particularly interesting as they show the various small steps and discoveries that eventually led to the successful decipherment. About the other, still undeciphered, scripts discussed in this book, so little is known that the author cannot do much more than describe the currently known inscriptions, how they have been discovered, and what attempts have been so far been made (often of a rather crankish nature) to explain and decipher them. There is not much hope that these other scripts will be deciphered in the foreseeable future, as we don't have enough material in them, no bilingual texts, and often no knowledge of the languages that were written using those scripts.

In fact, if we compare the three cases of successful decipherment (Egyptian, Linear B, and Mayan writing), we see that they have several characteristics in common:

Firstly, something was known of the language: Champollion knew Coptic, a descendant of the ancient Egyptian language, and put it to good use (pp. 64b, 69b, 72a). Michael Ventris, the decipherer of Linear B, made the crucial steps towards decipherment soon after he realized that the language used in the inscriptions is an early form of ancient Greek (and not some completely unrelated language, as has been believed until then and as is still considered to be the most likely for the Linear A inscriptions from Crete); pp. 98–101. And in the case of Mayan inscriptians, languages related to the one used by the pre-Columbian Maya in their inscriptions are still spoken by the present-day Maya that inhabit roughly the same area (p. 111).

Secondly, there was a substantial amount of material. This enabled e.g. Ventris and other decipherers of Linear B to look for frequently occurring patterns of characters, infer that they represented various grammatical inflections (for different cases etc.), so that he could start deducing where the Linear B characters fit on his ‘grid’ (since Linear B is a syllabic writing system, each character tends to represent a combination of a consonant and a vowel, and most of the syllabary can be compactly represented by a grid with one row for each consonant and one column for each vowel); see pp. 90, 97, 100.

In the case of Linear B, decipherment efforts moved practically nowhere for several decades until all the available material was edited and published and thus made accessible to people interested in deciphering it. Before that, for example, Evans sat on his Linear B material for decades, hoping that he would be the first one to decipher it (but he didn0t get far, unsurprisingly given that he was firmly convinced that the language of those inscriptions had nothing to do with Greek, and he also believed that the script was more pictographic than phonetic); pp. 84–7.

Thirdly, there was some bilingual material. The most well-known case is the Rosetta stone, containing the same text in Greek and in Egyptian. This enabled the decipherers to look for personal names, which could be reasonably assumed to be written approximately the same way in different languages.

Ventris similarly benefited from similarities between Linear B and the Cypriot script, which had been deciphered thanks to bilingual Cypriot/Greek inscriptions (pp. 81, 83, 98b); after the first steps, he could also make use of known place-names such as Knossos and Amnisos (p. 99b).

In the case of the Mayan writing, there was the “Landa alphabet”: Landa was a 16th-century bishop who ordered the destruction of many Mayan manuscripts that fell into his hands, but (somewhat surprisingly) he also wanted to record some information about their writing. He apparently thought the Mayan writing was a simple alphabet and got a native speaker to write down one character for each phoneme (p. 120). I remember reading about this in Sprague de Camp's books (Lost Continents p. 32, Citadels of Mystery p. 8), where I got the impression that the Landa alphabet was perfectly useless. But it seems to have been slightly useful after all: it did give the present-day decipherers some hints as to the pronunciation of some of the Mayan characters, just enough to get them started in the right direction and get the ball rolling, so to speak (pp. 119–21, 123). It was difficult enough even so, but without those hints provided by the Landa alphabet it might have been more difficult still, perhaps even impossible (p. 272a).

Unsuccessful decipherments

At least some of these three elements are usually absent in the case of the other (still undeciphered) scripts described in this book. Sometimes we don't know anything definite about the language (e.g. about Etruscan; or about the language used on Crete for Linear A, pp. 198–9); for the Indus Valley script, some have proposed that the language is from the Dravidian family, but the earliest related language that is known is from the 3rd century BC, nearly two thousand years after the Indus Valley civilization, so even if the Dravidian hypothesis is correct, the two languages cannot be very closely related (p. 278). Sometimes we have too little material, e.g. just a few rongorongo boards, or even just one single artefact in the case of the Phaistos disc. In most of these cases we also lack anything bilingual, even just a few recognized personal names or place-names that we might otherwise hope to find on the inscriptions in the unknown script and then work from there.

Anyway, how much material do we need to decipher an alphabet? John Chadwick, one of the pioneers of Linear B decipherment, spoke of a ‘critical mass’: “a quantity of text which will ensure that a few correct guesses will produce a chain reaction leading to more solutions. There is no formula known to me for determining the critical mass; it depends of course on the complexity of the script, and I should guess that it contains n squared where n is the number of different signs in the script.” (P. 36.)

Linear B

For the most part, this book was quite interesting to read. Occasionally the author goes into more detail about some script than I was really interested to read, but these potentially boring parts are never very long. The illustrations were nice — many of these old scripts are beautiful to look at; Linear A and B are the notable exceptions, being quite boring and dull. But this is hardly surprising if we remember that they were also used, apparently, almost exclusively for boring purposes such as business record-keeping. However, I am still somewhat amazed by the fact that the Minoan civilization, which by all accounts was brimming with joie de vivre more than almost an yother, has not taken the trouble to commit any of that exuberance and liveliness to writing, and has instead satisfied itself with writing down only the sort of boring commercial data that wouldn't tell us anything interesting about their history even if we were able to decipher their script.

I was interested to learn that, apart from the well-known Linear A and Linear B scripts, another and still earlier writing system was used on Crete, namely a kind of hieroglyphs (pp. 76, 183). Of course, even less is known about this script than there is about Linear A, and there is also much less material in it.

P. 99 mentions some of the first Mycenaean Greek words that Ventris recovered from Linear B inscriptions, such as “ke-ra-me-u (potter), ka-na-pe-u (fuller), i-e-re-u (priest) and i-je-re-ja (priestess)”. I am always delighted, even touched, to discover how many words that have a fancy ring to them nowadays actually have perfectly decent, ordinary, down-to-earth origins. Ceramics has always seemed somewhat fancy to me. But here I learned that the word has a charmingly humble origin: keramos means ‘clay’, kerameus means ‘potter’, and so on. And as for ka-na-pe-u, if this is the man who stuffs furniture, is the word related to our modern canapé, which is a kind of sofa (probably usually stuffed too)?

“Those who approach decipherment expecting sensational revelations—of great battles and the fall of civilizations” etc. “are likely to find their expectations confirmed [. . .] even if they have to invent an underlying language” (p. 44). Thus, the more mundane the contents of a text are according to a proposed decipherment, the more likely the decipherment is correct and not the result of some crank's feverish imagination. When Chadwick and Ventris “found the names of four classical Greek gods on a single tablet [. . .] Ventris had been horrified, because this was exactly the sort of too-good-bo-be-true result that previous, eccentric attempts at decipherment had been offering as ‘proof’. ” (P. 45.) Fortunately it turned out to be the description of a ritual, listing the gods to which sacrifices had been made: a “fairly mundane explanation” after all.

A substantial number of new Linear B tablets had been discovered at Pylos just before the WW2; soon after the war was over, Emmett Bennett edited and published a book about them and sent a copy to Ventris in England. “[W]hen he [= Ventris] went to pick up the packet, a suspicious postal official asked him: ‘I see the contents are listed as PYLOS TABLETS. Now, just what ailments are pylos tablets supposed to alleviate?’ ” (Pp. 87–9.)

Ventris was very talented for languages. While on holiday in Rome with a friend, “he was able to get them into a part of the Vatican closed to the public by chatting to the Swiss Guards, in what they mistook to be the Swiss-German dialect of a native speaker.” (P. 92a.)

The Mayan script

The chapter about Mayan glyphs also has some interesting pages on the Mayan number system and calendar; these were also the first parts of their writing to be successfully deciphered (pp. 112–6).

“The very name Yucatan is derived from ‘uic aithan’—the phrase spoken tot he Spanish conquistadors by the Maya when asked what their land was called: it means ‘what do you say, we do not understand you.’ ” (P. 120a.)

During Landa's efforts to get a native Maya speaker to explain the Maya writing, at some point “communication with his informant clearly broke down, as we can see from the following noted phrase: ma-i-n-ka-ti which means ‘I don't want to’—presumably the informant's response to Landa on being requested to write further phonetic values of the mysterious glyphs.” (P. 121.)

There's a photo of Yuri Knorozov, the chief decipherer of the Maya writing, on p. 122. The photo must be quite old; the contrast is quite high, some very bright white, lots and lots of very dark black, and little of any shades of gray in between; the expression on Knozorov's gaunt, grim face is best described by >-(, and in his arms he is holding a cat that is staring at the camera as if it was determined to win some kind of award from If I didn't see from the caption that it was Knorozov, I would assume that it must be the supreme villain from some cold-war-era James Bond movie. Priceless. (It's similar to the one on this web page, but even better.)

The “Maya scribes loved to play with their system and use it to spell words in several different and unpredictable mixtures of phoneticism and logography, not just two or three, as in the Egyptian hieroglyphs.” (P. 132.)

The Etruscans

An interesting quote from Seneca about the Etruscans (p. 162): “ ‘The difference between us and [them] . . . is the following: while we believe that lightning is released as the result of the collision of clouds, they believe that clouds collide so as to cause lightning. For since they attribute everything to the gods' will, they believe, not that things have a meaning insofar as they occur, but rather that they happen because they must have a meaning.’ ”

Apparently, the vowel o “did not exist in Etruscan” (p. 170a) — I'm amazed; how is it possible for a language to fail to make use of a sound as obvious and easily pronounced as o?

From an illustration on p. 178, it seems that the Etruscans did not make paragraph breaks the way we do today, but instead used ‘insert paragraph’ marks very similar to those used by modern proofreaders.

Linear A

The development of writing in the Minoan-Mycenaean culture was not quite so straightforward as Evans had originally imagined (the hieroglyphic script leading to linear A which led to Linear B, and all of this happening on Crete): “all three scripts have been found outside Crete, and the spans of their dates are now seen to overlap; [. . .] Linear A and Linear B may be ocusin scripts, rather than the first being the parent of the second.” (P. 183b.)

Two French researchers published a “five-volume collection of Linear A inscriptions, Recueil des Inscriptions en Linéaire A (known familiarly as GORILA)” (p. 193). (From their last names, Godart and Olivier, and the first letters of the title words.)

The origins of the language used to write Linear A are unknown; it is “only possibly Cretan in origin” (p. 183b). Various other hypotheses have been proposed: that it is a Semitic language; that it is another early form of Greek; or that it is an Indo-European language originally from Anatolia, e.g. Lycian (pp. 198–9).

The Phaistos disc

Page 298 discusses whether the Phaistos disc may be a modern fake, “ ‘a joke perpetrated by a clever archaeologist from the Italian mission to Crete upon his fellow excavators. Taking a thermoluminescence test, which should date the firing of the clay at about 100 years ago, can solve the mystery of the disc.’ ” (Jerome M. Eisenberg, cited on p. 298.) But, perhaps unsurprisingly, the Greek authorities have so far not appeared to be willing to try anything of that sort: “No one is going to thank the person who proves the disc to be a fake. ‘For who would want to look at an ex-enigma, or buy a Phaistos disc postcard, or any souvenir, if the disc had already been deciphered?’ asks Bennett.” Actually, I don't think that decipherment would impact the souvenir sales — the British Museum, for example, is still whoring out the Rosetta stone for all it's worth. But proving that the disc is a fake might be harmful to the trinket business. Nevertheless, fake or not, it's a beautiful artefact. And “[i]n fairness, it must be said that the hoax theory is very mucha minority opinion” (p. 303a), and there are lots of arguments against it.

Interestingly, the characters on the Phaistos disc “are undoubtedly impressed, not incised, into the clay (unlike the characters of Linear A and B)” thus making it “ ‘the world's first typewritten document’ (Chadwick)” (p. 304a).

“There is an empirical formula for working out the probable number of signs in an alphabet or syllabary from a small sample of the alphabetic or syllabic writing. It has been shown to work well” with many scripts both ancient and modern (pp. 308–10). The formula is (L × L)/(L − M) − L where L is the total number of characters in your corpus, and M is the number of different characters (p. 310a). A bit of mathematics certainly shows that this converges to M as L grows towards infinity. The formula has been used on the Phaistos disc and suggest that the full Phaistos syllabary would have around 56–57 signs.

The Phaistos disc has been very popular with all sorts of cranks claiming to have deciphered it. They kept pestering archaeologists such as John Chadwick, who complained that it “has been a millstone round my neck for decades” (p. 312b).


“[A]s we now know, the Egyptian languages written in hieroglyphic and demotic are not identical, but they are closely related, like Latin and Renaissance Italian.” (P. 60.)

There's an interesting paragraph on the origins of the inhabitants of Easter Island on p. 222. It's well known that Thor Heyerdahl proposed that at least a part of them came from South America rather than from the other Polynesian islands lying to the west of Easter Island. I have also read in various places that Heyerdahl's arguments didn't manage to persuade the other experts, and hardly anyone accepts his hypothesis nowadays. Here in this book at least some of the reasons why are mentioned: “the scientific evidence—archaeological, ethnological, linguistic and genetic—overwhelimgly supports the” theory that Easter Island was settled only from Polynesia; among other things, “a 1990s analysis of the DNA in early skeletons from Easter Island (predating the colonial contacts of the 18th century) revealed no trace of genetic contact with South America.”

“If we could decipher the Indus script, we would perhaps learn if the root sof Indian civilization really did possess some special genius, different from the other ancient civilizations, as suggested by the material artefacts at Mohenjo-daro and Harappa and the spiritual emphasis of subsequent Indian culture—Buddha instead of Alexander, so to speak.” (P. 321.) This is an intriguing idea, but I'm not too optimistic. So far, whenever we have had an idealized opinion of some ancient culture, it has subsequently been found to be not quite so ideal; the Maya were eventually found to be bloodthirsty, and even the Minoans were not above the occasional human sacrifice (Nat. Geo., Feb. 1981, 205–24). On the other hand, those whom we are apt to demonize eventually turn out to be ordinary people after all (e.g. the Carthaginians). I'm afraid that human nature is equally bad all over the world, in all times and places.


  • Maurice Pope: The Story of Decipherment (2nd ed., 1999). Covers “every significant deciphered script” (p. 46 here).

  • Michael Coe: Breaking the Maya Code.

  • Hans J. Nissen et al.: Archaic Bookkeeping: Early Writing and Techniques of Economic Administration in the Ancient Near East. “[T]heir fascinating unraveling of the proto-cuneiform and proto-Elamite texts [. . .] perhaps the most important book on the origins of writing yet written” (p. 210a).

  • It might be interesting to read more about the interactions between ancient Egypt and the areas south of it, e.g. the Kushite culture of Nubia and Meroe. I suppose that the elements of Egyptian civilization slowly trickled across its borders and influenced the adjacent areas, but it must have been a slow process. I hope I'll eventually find some interesting book about this subject. Some years ago I started reading a History of Black Africa by Joseph Ki-Zerbo, but found it terribly boring and gave up before I read more than a fifth of it. I must admit that it contained a lot of information about early states in sub-Saharan Africa. Perhaps I should give it another try some day.

  • I also recommend this bit of Linear B humor.


Anonymous Anonymous said...

Keep on posting such themes. I love to read stories like that. By the way add more pics :)

Wednesday, February 03, 2010 12:39:00 PM  
Anonymous Anonymous said...

Great post as for me. I'd like to read something more concerning this matter. Thanks for posting this material.
Joan Stepsen
Indian escorts in London

Friday, February 19, 2010 2:54:00 PM  
Blogger Seamus Sweeney said...

I was happily Reading this book earlier on a train journey. And then I lost it in the back of a taxi! The irony. I was greatly enjoying it, but only got as far as the beginning of the Meroite chapter. I'm clinging to the hope the taxi driver may realise what happened and call with it... but that might be like wishing for another few rongorongo boards appearing from nowhere...

Sunday, December 12, 2010 10:45:00 PM  
Anonymous tuition agency said...

Thanks for sharing. Appreciate it.

Saturday, February 26, 2011 10:22:00 AM  
Anonymous Anonymous said...

“There is an empirical formula for working out the probable number of signs in an alphabet or syllabary from a small sample of the alphabetic or syllabic writing. It has been shown to work well” with many scripts both ancient and modern (pp. 308–10). The formula is (L × L)/(L − M) − L where L is the total number of characters in your corpus, and M is the number of different characters (p. 310a). A bit of mathematics certainly shows that this converges to M as L grows towards infinity.

Are you sure you've transcribed the formula correctly? According to my Excel version, it converges to L, not M, and is therefore inaccurate.

Tuesday, June 12, 2012 1:27:00 AM  
Anonymous Ari said...

Your review of "Lost Languages" has been a really, really interesting read for me. I'm a big fan of linguistics and the sheer amount of bad information one can find online on the topic makes me really cautious.
Thing is, living at the boundary of Eastern Europe isn't really a good chance of finding this book. Amazon doesn't deliver here and other service I found only has 2nd hand books and it's really expensive: with the transportation cost it costs just as much as I earn in 1.5 months.

I wondered if you have an electronic version of this book or a scan, or anything really. I really wanted to read more quality things on deciphering, as Mr. Ventris and Mr. Champollion are big heroes for me (I'm in high-school but I think these are some good heroes; most of the people my age around me would want to be famous soccer players or oligarchs).

Anyway, sorry for taking from your time and any help or suggestion is greatly appreciated.

Saturday, March 05, 2016 10:27:00 PM  

Post a Comment

<< Home