I’m reading Toby Wilkinson’s excellent The Rise and Fall of Ancient Egypt
Among the great inventions of human history, writing has a special place. Its transformative power—in the transmission of knowledge, the exercise of power, and the recording of history itself—cannot be overstated. Today, it is virtually impossible to imagine a world without written communication. For ancient Egypt, it must have been a revelation. We are unlikely ever to know exactly how, when, and where hieroglyphics were first developed, but the evidence increasingly points toward a deliberate act of invention. The earliest Egyptian writing discovered to date is on bone labels from a predynastic tomb at Abdju, the burial of a ruler who lived around 150 years before Narmer. These short inscriptions already used fully formed signs, and the writing system itself showed the complexity that would characterize hieroglyphics for the next three and a half thousand years. Archaeologists dispute whether Egypt or Mesopotamia should take the credit for inventing the very idea of writing, but Mesopotamia, especially the southern city of Uruk (modern Warka), seems to have the better claim. It is likely that the idea of writing came to Egypt along with a raft of other Mesopotamian influences in the centuries before unification—the concept, but not the writing system itself. Hieroglyphics are so perfectly suited to the ancient Egyptian language, and the individual signs so obviously reflected the Egyptians’ particular environment, that they must represent an indigenous development. We may imagine an inspired genius at the court of one of Egypt’s predynastic rulers pondering the strange signs on imported objects from Mesopotamia—pondering them and their evident use as encoders of information, and devising a corresponding system for the Egyptian language. This may seem far-fetched, but the invention of the Korean script (by King Sejong and his advisers in A.D. 1443) provides a more recent parallel, and there are few other entirely convincing explanations for the sudden appearance of fully fledged hieroglyphic writing.
I like the image of the inspired genius “pondering the strange signs on imported objects”; does anybody know how widely accepted the sequence of events described here is?
Chinese writing may have originated the same way. It appeared out of blue following China’s first contact with Indo-European steppe nomads (the Chinese ended up borrowing a lot, including the entire Bronze Age chariot warfare complex).
Granted, the nomads were illiterate themselves, but they had extensive contact with literate societies of Mesopotamia, so it is quite possible that they transmitted to Shang China the idea and principles of writing, but not any concrete signs.
One difference is that the Koreans had been using Chinese script for about 1300 years by the time King Sejong proposed his syllabary– not even the first Korean syllabic writing system. See Peter Kornicki’s _Languages, Scripts, and Chinese Texts in East Asia_ (a wonderful book from all sides).
As far as I know, the Maya script is the only writing system certain not to be ultimately connected with the Fertile Crescent, at least through diffusion of the concept of writing. Rongorongo is an open question.
In Mesopotamia, we probably have a fuller view of the history of writing simply because they used clay and the evidence is better preserved. Denise Schmandt-Besserat has written about that, suggesting that clay tokens preceded the use of writing for accounting.
https://sites.utexas.edu/dsb/tokens/the-evolution-of-writing/
For areas like Egypt or China, though, I’m skeptical that we’ve captured enough of the evidence to really have a picture of what happened. If they developed writing on papyrus, palm leaves or bamboo, little if any of that is likely to have survived.
There are occasional finds from the Neolithic like the Dispilio tablet, the Vinča symbols or the Jiahu symbols that *could* be proto-writing. The problem is that those are pretty isolated finds, and may not be related to writing at all (and may not be related to later writing even if they are).
For areas like Egypt or China, though, I’m skeptical that we’ve captured enough of the evidence to really have a picture of what happened. If they developed writing on papyrus, palm leaves or bamboo, little if any of that is likely to have survived.
Good point.
I feel certain that the visual arts didn’t begin one Thursday afternoon when one person, Person A had the big idea of drawing cattle on a cave wall. It seems ridiculous that writing might have come from a single source.
Maybe I’m being thick, but doesn’t lack of a single source make it more difficult, as several people have to independently make the improbable leap forward…
Just how fully formed were early hieroglyphics? I had a vague idea that signs representing objects came first and signs representing syllables similar to the names of those objects later, and the leap to the first is definitely less far than the leap to the second.
@Y: I’d modify that slightly and say Mesoamerican scripts are independent of anything in the Old World–the Mayans appear to have had some predecessors. There’s evidence the Olmecs had writing (e.g., the Cascajal Block) and there’s a Zapotec script as well, both of which probably predate the appearance of the Mayan script. There’s also an Isthmian script which is more or less contemporary with the earliest preserved Mayan writing.
What I said before applies to Mayan writing as well. We know they used bark codexes, most of which have been destroyed. At least one from the Classic era has been found (I think in a tomb), but it’s poorly preserved. Older bark would not likely have been preserved.
AJP: It seems ridiculous that writing might have come from a single source.
Well, yes, and well, no.
One point is statistical. Whether there were two, three or four independent inventions of writing, they were very few and far between. This goes to show that invention of writing was something very rare that took very special circumstances. I don’t expect that to have been achieved by the work of one great thinker, but rather a series of small steps and lucky repurposings. An almost impossible combination of rare events.
Another point is statistical. Can we believe that a string of almost impossibly rare events happened contemporanously in the neighbouring civilizations og Mesopotamia and Egypt, rather than influence one way or the other? China is far away, but SFReader makes the point that the first writing seems to be recorded just as China was connected to a new level of transcontinental trade. Could this be chance? But even if travellers brought inspiration rather than the whole system, I don’t particularly like the description above of the “inspired genius pondering the strange signs”. As soon as the concept of writing is available, the development of a writing system is less of a leap. One might even argue that repurposing a foreign system for your own language is more of a leap than inventing a new one. I think the travellers — nomads or seafarers — would have had a good grasp of the system and explained it to their trading partners. “Look here: “Amphora – five lines – amphora again, with grapes this time – curl – a man and a spear — a tower and two sheep”. In Hanepti that is “Okam dedem mom fafishu anpulan kanimet”. In your language “Five units of wine from Anpulu of Kanima” — Kanima, that’s a town up north. The name sounds like “four lambs”. And Anpulu literally means spearman. Yeah, you can imagine the jokes, but he makes very good wine. And with these signs I don’t even have to open the bottle to know it’s his.”
Michael: That’s a good point about Central American scripts.
I was trying to properly equivocate the Chinese script. It maybe was, maybe wasn’t independent from of the Mesopotaamian invention of the script, but the Central American scripts undoubtedly were so.
Sorry–should have pointed out that I was making a very minor nitpick. This is an area of great interest to me.
It’s quite conceivable that the alphabet was the invention of a single genius, very likely helped along in his/her inspired generalisation of the rebus principle by being a native speaker of a Semitic language where it’s comparatively natural to dream up a writing system which ignores the vowels (and encouraged by example of Egyptian itself, of course.) Unlike writing in general, the (idea of the) alphabet seems to have emerged exactly once.
We probably would all still be struggling with cuneiform if it hadn’t been for the happy accident that the Afroasiatic languages have such bizarre morphophonemics. It’s all those Canaanite builders and their graffiti we have to thank. (And the Egyptians, of course.)
The real parallel is Sequoyah. Sejong and all his advisers read & wrote Classical Chinese fluently.
Also Shong Lue Yang, the inventor of Pahawh Hmong. Pahawh Hmong is, I believe, unique among syllabaries, in that the main shape of each character indicates the rime of the syllable (nuclear vowel and tone, plus the occasional coda), and diacritics encode the onset.
doesn’t lack of a single source make it more difficult, as several people have to independently make the improbable leap forward…
I had a vague idea that signs representing objects came first …
I see nothing improbable with the invention of symbolism-in-general. Aside from cave paintings (which are stylised, not representational), there could be scratchings in sand/loose dirt: we’re here (X marks the spot); the buffalo are there; there’s a stand of trees near them and it’s downwind; we’ll drive them towards the cliff over there. Nobody’s going to draw a representationally accurate map for that.
Rocks/trees marking territory need to be scratched/scored with some symbolism that couldn’t just be natural weathering.
Merely: we won’t find archaeological evidence until there’s trade and property rights and administration that needs to keep indisputable permanent record.
Iron metallurgy was a single source invention.
Wheeled vehicles may have been another.
Gunpowder too.
Much credit is assigned to the inventor of the wheel, though the idea may have come from playing with the earlier bead.
FWIW, writing systems researcher Peter T. Daniels, An Exploration of Writing (2018) p. 136:
“In Sumer, China, and the Yucatan,* pictograms turned into logograms.” “* Or in some other part of Mesoamerica (see #6.3 Details).”
David Eddyshaw: There’s a pretty good argument that alphabetic writing is, if not the work of a single genius, the work of a fairly small group of people. It may have happened at a place now known as Serabit el-Khadem, an Egyptian mining site in Sinai. However the workers seem to have been Canaanites.
Some time around 1750 BC someone got the idea to make a list of all the sounds in the Canaanite language and represent each by the corresponding hieroglyphic symbol. So the sound B was represented by the word ‘beth’ for “house” and indicated by the hieroglyphic for house. Of course in Egyptian the word for house would be spoken completely differently.
The Egyptians were into bureaucracy, so undoubtedly there was a scribe there to keep the records, and they must have had at least a few foremen who were bilingual.
I believe that hieroglyphics had some phonetic component, but still it was quite an intellectual leap to go from there to a totally phonetic system. Whether it was just one person or a group of people hanging out after work, you couldn’t know nearly 4000 years later.
There is a temple of the Egyptian goddess Hathor on the site, but the inscriptions show that the workers called her “Baalat”.
It’s possible that this breakthrough didn’t occur at this site, as there are other inscriptions around the same time period, and they are all hard to date.
According to my source (David Sacks, Letter Perfect: The A-to-Z History of our Alphabet, Knopf 2003), all alphabetic writing except for Korean derived from this one invention. I’m not sure I believe that for Devanagari. At any rate I don’t know enough about the subject to venture an opinion.
cave paintings (which are stylised, not representational)
Stylised isn’t an antonym of representational, see ‘stylised’ here (two-legged animals). The opposite of representational is ‘abstract’ (which they aren’t).
The sequence for inventing writing is something like, 1. Make marks in the snow, on the tree, scratched on the skin. 2. Notice that marks, blobs and cloud formations sometimes resemble animals and other objects. 3. Use marks (aka drawings) to communicate ideas like ‘grazing buffalo’ or ‘shooting deer’. 4. But these ideas are also communicated by noises. 5. Notice that phrases like ‘grazing buffalo’ can be drawn, there’s an equivalence. 6. Repetition: communicate by always repeating the same action when similar circumstances arise. Always make the same drawing to communicate the same word or verbal idea.
That would account for different forms of writing evolving in different societies. The concept of one person being responsible for devising the entire sequence is unnecessary and seems absurd. The same goes for ‘inventing the wheel’. It just evolved over time: 1. People found a large roundish stone and used it perhaps as a fulcrum. 2. Their children saw the rock roll down a hill. 3. When they grew up they used the rock and logs to move loads that were being transported (perhaps tree trunks to a river).
Devanagari is ultimately from an Aramaic script, though the Indians deserve the credit for inventing vowel symbols (on purpose, instead of by lucky misinterpretation, like the Greeks.)
There have, of course, been alphabets created de novo, but always by people already familiar with the basic idea (like King Sejong.) Korean actually isn’t an exception to this.
Phonology is older than writing in India, because when the Persians appeared on the scene and brought their Aramaic-writing bureaucrats in, Sanskrit was already a classical language not natively spoken by anyone, yet its exact pronunciation was considered extremely important.
Sejong, too, appears to have been a passionate phonologist – and his writing system, too, doesn’t quite come out of nowhere: officially it is adapted “from the Ancient Seal Script”, and nobody is more ancient than the undercover ancients…
The concept of one person being responsible for devising the entire sequence is unnecessary and seems absurd.
Nobody said that. The suggestion was that one Egyptian (or a small group, as with the Canaanites) might have been responsible for adapting the Mesopotamian system, which probably developed along the lines you suggest, to his local area and language. That’s no more unlikely than the Chinese (to take a current situation) stealing technology developed over time and with much effort elsewhere.
I’ve wondered, sometimes, whether the “default vowel” abugida nature of the Brahmi script branch (of which Devanagari is a member) is somehow derived from the similar structure of Persian cuneiform.
In principle, the timing fits, and the location isn’t too far off, but Persian cuneiform was only used briefly, and was apparently never very popular in the first place (I vaguely recall that only one, very short, text is known that isn’t a monumental inscription).
Agree with David Eddyshaw that the “alphabet” (in actual fact – abjad) would have been facilitated by the structure of Semitic (or rather Afroasiatic) languages. And as David points out, this was encouraged by the example of Egyptian where a large number of phonetic signs was used. The Egyptian phonetic signs represented 1, 2 or 3 consonants. Egyptians did have an abjad of sorts, because in theory, Egyptians could have used only their 1-consonant signs to write everything in their language. Indeed, the Meroitic writing system does this. But the Egyptians found that that the efficiency offered by the 2 and 3 consonant signs benefitted their writing system (a bit like using X and not KS / GZ in English or ψ in Greek) and outweighed any benefits of using just the 1-consonant signs. And of course, there is the sheer inertia of the literate classes who want to preserve their status through their complicated writing system.
The Semitic / Phoenician writing system is an abjad rather than an alphabet, because it only represent consonants. The first true alphabet is the Greek one, although there could be an argument that the Greek alphabet is not a pure alphabet either in the sense of 1 letter = 1 sound, because:
– there were sounds it didn’t represent: eg /h/ which was represented only in the archaic alphabet but not in the classical Ionic alphabet (well not represented by a letter anyway)
– relevant sound distinctions were only partially represented eg. short ο and long ω, but α could be long or short.
– letters stood for two sounds: ξ, ψ.
On Devanagari and other Indic scripts: Colin Masica (in his fine 1991 book THE INDO-ARYAN LANGUAGES, pages 133-137) argues that Brahmi -which Devanagari descends from -does not derive directly from Aramaic or indeed any script, but must instead be a case of “idea diffusion”, i.e. must have been created by (an) individual(s) familiar with some other script(s) as well as with phonetics. A parallel he does not cite but which might make a topic for a decent book or thesis is the similarity with the invention of Armenian and Georgian script: both were invented by individuals who must have known Greek, Syriac, possibly both, possibly other scripts as well, but who created scripts which were admirably suited to the phonology of the languages they were meant to write.
Yes, a study like that would be very worthwhile. We lose a lot by only focusing on one bit of the earth at a time.
Language: The suggestion was that one Egyptian (or a small group, as with the Canaanites) might have been responsible for adapting the Mesopotamian system
Fine. But it also says
the evidence increasingly points toward a deliberate act of invention.
The only evidence presented is of a deliberate act of copying Mesopotamian writing: Mesopotamia, especially the southern city of Uruk, seems to have the better claim. It is likely that the idea of writing came to Egypt along with a raft of other Mesopotamian influences
He says
Archaeologists dispute whether Egypt or Mesopotamia should take the credit for inventing the very idea of writing…
But it seems unlikely to me that anyone would come up with the very idea, it would have developed over time. That’s what I was writing about.
Well, I read “invention” in that context as equivalent to “developed over time.” I don’t think the author, or anybody but a crackpot, would literally suppose that writing sprang full-blown from the brow of one person frowning and scribbling one fine afternoon.
infinite number of monkeys scribbling for infinite amount of time will eventually invent the idea of writing
Is there an agency that contracts out teams of an infinity of monkeys, with or without typewriters? I have a few projects in mind (not Hamlet. That’s been done.)
The concept of one person being responsible for devising the entire sequence is unnecessary and seems absurd. The same goes for ‘inventing the wheel’. It just evolved over time
Yet one often finds single persons reinventing the wheel, even if only a square wheel. I see nothing absurd in the ur-idea of an ur-wheel springing from the brow of a single person. Not titanium counterbalanced wheels, of course. Too early for that.
As with “fittest” DNA mutations, we see today only ideas that have persisted. We don’t see the billions of square marshmallow wheels that didn’t convince the punters.
Last year I read an article in which the author undertook to explain why wheels don’t occur in nature. The reason was that wheels require axles – or something like that. It left me rather bemused.
Not long after I read about the discovery of a teeny-tiny wheel in the rear legs of grasshoppers. I seem to recall that it worked like a ratchet, or rather a clutch plate.
The reason was that wheels require axles – or something like that.
TL/DR is that, for wheels to work, there has to be a rotating part (wheel, or wheel pair plus axle) and a non-rotating part (actual vehicle), which would by necessity have to be non-connected (for the rotation to be stable without an ever-extending connection), and natural organisms just don’t come in disjoint parts like that.
It’s possible, in principle, for the entire body to be a wheel (or, theoretically, wheel pair plus axle), without a non-rotating part; a simplified version of this is how tumbleweeds work.
Heavens! Planthopper reinvents the wheel. I love to see asymmetry in nature (and in Nature).
So the sound B was represented by the word ‘beth’ for “house” and indicated by the hieroglyphic for house.
Was it? I thought the story was that the inventors did not understand hieroglyphs, otherwise they would have used the hieroglyph for /b/ instead. Any hieroglyph that looked vaguely like a house would have done. This particular case does seem to be correct, but does the hieroglyph for ‘ox’ look like a Proto-Sinaitic aleph?
Devanagari is ultimately from an Aramaic script, though the Indians deserve the credit for inventing vowel symbols
Brahmi script looks suspiciously like Ethiopic script, and the principle is exactly the same, so perhaps what the Indians added was not so much vowel marks as the inherent, unmarked vowel, which doesn’t exist in Ethiopic.
Is there an agency that contracts out teams of an infinity of monkeys, with or without typewriters?
Sure is:
https://en.wikipedia.org/wiki/Amazon_Mechanical_Turk
Those plant hoppers are really amazing. However, what they demonstrate is cogging gears, not a wheel and axle. The latter would be very, very hard to evolve.
The existence of alien Wheelers (described as an evolutionary dead end) in the galactic natural history of Terry Pratchett’s Strata might be intended as a clue that things are not all as they seem. Well actually, they are definitely evidence that things are not at all what they seem; but maybe in a more profound way than what the characters at the beginning of the book have figured out.
Wheels in nature: there’s the exception that proves the rule – the bacterial flagellum, which consists of a rotor made from a few protein molecules and a stator made from a few protein molecules set in the cell wall. Also, as a smaller version, ATP synthase and a few other such things.
Flagellum isn’t really a mechanical device, it’s a molecular contraption with a rotational symmetry which can disconnect and reconnect molecular bonds. There are plenty of those in nature. DNA helicases are the prime example. Since DNA is wound into a helix, one needs a lot of rotating to read or copy it. But at the molecular scale of things, there’s isnt an axle and a wheel, but rather bonds broken and replaced.
…if by “bonds” you mean hydrogen bonds, maybe, but those also exist between the flagellum and the surrounding water. There are no covalent bonds between the parts of a flagellum.
A feature of the Mulefa is their use of large, disc-shaped seed pods from their world’s enormous “seed-pod trees” in locomotion; the pods fit neatly onto a spur on their front and rear legs when each zalif has grown enough to use it. They propel themselves using their other two legs, like a cyclist without pedals.
https://en.wikipedia.org/wiki/Races_and_creatures_in_His_Dark_Materials#Mulefa
Which is it then, a mechanical device or a molecular contraption ?
Wheels-and-axle is an example of what Heinz von Foerster called a trivial machine. Computers are another example. Animals are much more interesting machines.
Suns too. This is explained by Edgar Morin in La Nature de la Nature, the first of the six-book series called La Méthode.
Stu, As usual I don’t know anything about this Heinz Foerster or his interests. Perhaps to include humans & other animals as machines is intended – like le Corbusier’s description of a house as a machine for living (in), une machine-à-habiter – to give distance to the usual cozy prejudices. That’s fine, but in any other situation we lose the advantages of having the category ‘machine’ in the first place.
hydrogen bonds, maybe, but those also exist between the flagellum and the surrounding water
even more importantly, ion-pair intermolecular bonds, and yes, it is exactly my point that the “motor” of the flagella is “greased and rolled” by the same types of forces which link together the other molecules around them, at the same angstrom-scale distances, So there is nothing qualitatively separating the moving parts; rather, they are a part of the same macromolecular complex built together on the same principles as the rest of the macromolecules.
In contrast, mechanical wheels have moving parts separated by microns or millimeters, and the specific interactions of the molecules along their interface play no useful role at all in all in enabling rotation.
But we are rolling too far from the language themes here 🙂
Well, I assume that one set of molecules are connected strongly and rotate rigidly (to a first approximation) relative to another set of strongly connected molecules, the two sets making their closest contact along a circle.
If that is not the definition of a plain bearing, I don’t know what is. That the same kind of molecular forces exist internally and between the parts is just a detail.
Crown: perhaps…[it] is intended to give distance to the usual cozy prejudices. That’s fine, but in any other situation we lose the advantages of having the category ‘machine’ in the first place
Morin and Foerster (look ’em up in the English WiPe) are not on about prejudices. You might think of it this way:
The notion of “machine” arose originally to describe non-living contraptions that did cool things predictably. Somewhat later, the notion was extended to living things because they too do cool things, although not quite as predictably. But good enough for medical science and pollsters.
Many people fret about the “reductiveness” of regarding humans as machines. But that’s primarily because the old sense of a non-living contraption still infects the notion of “machine”.
So let’s turn the tables, taking our primary notion of machine from living creatures (and suns). When we then attempt to transfer it to non-living contraptions, we find these fall short of the mark by comparison. Thus “trivial machines”. These can do cool things, but not on their own initiative. They can’t repair themselves except on the silver screen. Ee tee see.
Me, above: [A lot of text I wish I had waited for time and occasion to write and edit properly, including:]
I think the travellers — nomads or seafarers — would have had a good grasp of the system and explained it to their trading partners. “Look here: “Amphora – five lines – amphora again, with grapes this time – curl – a man and a spear — a tower and two sheep”. In Hanepti that is “Okam dedem mom fafishu anpulan kanimet”. In your language “Five units of wine from Anpulu of Kanima” — Kanima, that’s a town up north. The name sounds like “four lambs”. And Anpulu literally means spearman. Yeah, you can imagine the jokes, but he makes very good wine. And with these signs I don’t even have to open the bottle to know it’s his.”
What I mean is that we talk of courts and tax collectors, but we shouldn’t forget the value of information in form of tags and lists for travelling traders. They would have seen the added value in information on their goods, and once they had that information, they would have seen the added value from making (some of) that information available for their customers. This is a just-so-story, but on this level every origin hypothesis is a just-so-story. And it can be tested against evidence from shipwrecks in the Red Sea and nomad graves in the Tarim basin.
When it comes to system design, I think we are too hung up on alphabets and imperial administrations. For the Egyptian first encountering Sumerian writing (or verse visa, mutaetoes mutahtoes), there was no reason to think that the signs could be repurposed for a different language, any more than both sound and meaning of the words could. And she was right. When logogrammatic systems have been imported to languages they weren’t designed for, the results are awkward and suboptimal and the Japanese suffer for two millennia. It’s what happens when those already literate use the tools they have at hand. The default idea when meeting writing seems instead to have been to use the concept and make ones own system.
When the Semiotes abjadicated hieroglyphs, they just followed the usual pattern, though on a different system. The Egyptians had developed a way to represent foreign names and words in Egyptian, but there was no reason to think that the same sign(name)-sound-combinations would be useful when representing Canaaneic directly.
the results are awkward and suboptimal and the Japanese suffer for two millennia
…and, come to think of it, so did the Akkadians, in a conceptually very similar situation.
(And then there came the Hittites, who had to deal with Sumerian and Akkadian…)
I don’t know which is the worse writing system, Japanese or Hittite.
Haven’t I had the idea of travelling merchants before? Apparently.
The point that does not seem to have been made is that it in fact is possible to overstate the importance of writing and there is a lot of new evidence that seems to point exactly in that direction. There were many quite sophisticated civilisations that may not have had anything much resembling writing for example the Incas or the Harappans. Is also much easier for oral transmission of information to be quite reliable if that is what you have available.
What writing does do is make it a lot easier to record what happened for people like us who can then talk about it. This gives us the illusion of civilization being linked to written culture whereas that may not necessarily be the case. In the same way that we associate civilized architecture more with stone or brick rather than wood. But that is really just the availability heuristic in action.
Japanese is definitely worse, what with kanji, katakana, hiragana, and Latin letters. Hittite had only one syllabary, a bit more complex because it had to handle non-CV syllables, and only about 375 logograms. (Note that Unicode has 2800+ emoji, rather more than the 2100+ standard kanji, though highly educated Japanese may know 5000+.)
The Incas didn’t have writing, but they did have sophisticated methods of record-keeping.
Isn’t the primary moral from the Japanese example that a highly suboptimal writing system is really not nearly so burdensome and costly as orthographic reformers like to suggest, given how well the Japanese seem to have managed (in material prosperity, life expectancy, art, military success until they overreached, etc etc) despite that handicap? Even after centuries of having by contrast a well-designed writing system unusually beloved by orthographic buffs, the Koreans right next door are still trailing them (esp if you take a weighted average of North and South …).
It’s true, a highly suboptimal writing system does not actually, all by itself, doom a nation to penury and oblivion. Now that we’ve disposed of that straw man…
The straw is supplied by repeated invocation of the word “suboptimal” without its being clear, in each invocation, what the terminus ad quem is. “Suboptimal” to what specific purpose ? “Optimal” to what end ? Cui bono ?
Dominik Lukes gave an example above: What writing does do is make it a lot easier to record what happened for people like us who can then talk about it. So when your goal is to have things to jaw over that you might otherwise have forgotten about, an easy-to-deal-with writing system is not suboptimal for that purpose.
Although talking as well can lead to penury and oblivion, when it is pursued so intensively that nobody remembers to milk the cows.
Gives a counterexample.
Having learnt two ‘suboptimal’ writing systems, I have become increasingly sceptical of people who spout the mantra of how terrible they are. Yes, they have their bad points, but they are not as bad as critics (who I suspect are often frustrated foreign learners) make out.
Japanese: “Oh dear, all those characters and their different readings you have to learn! Definitely one of the hardest writing systems in the world to master.” Chinese characters do impose a load on the memory; there is no denying that. But the number of characters has been limited, alleviating this problem. And the multiple readings (which are only particularly severe for a few characters like 生 or 上) mirror certain aspects of the Japanese lexis, and in fact provide a lot of help in learning that lexis. So once you know the character 頼, which is read tanomu (頼む) ‘ask, request’ in plain speech and rai in slightly fancier words like 依頼する irai suru ‘ask, request’, it helps you remember the connection between words like tanomu and irai suru and choose the word suitable to the occasion. This is one of many hundreds of such connections that run through the Japanese vocabulary. Sure, writing these words in romaji is ‘easier’ at first blush, but it fails to show these rather important linkages. The retort, I suppose, is that Japanese shouldn’t structure their vocabulary like this and should just use romaji, which doesn’t require consciousness of them. So what else do you want to tell the Japanese to do with their language?
Japanese has been using this ‘suboptimal’ writing system for hundreds of years, and despite its difficulties it has sustained a thriving literature that a good proportion of the population, even in pre-modern times, were sufficiently literate to read.
Having learnt the characters and their readings when I was younger, I’ve come to realise that they mostly become second nature to literate Japanese, a part of the ‘texture’ of the language. For the foreign learner they form a steep learning curve, which is, I suppose, why many think the Japanese system is ‘suboptimal’ and needs to be changed. I would suggest that Chinese is much harder because of the sheer number of characters available to write a relatively smaller number of distinct syllables.
Mongolian traditional script: One commenter here way back said that they’d given up on the Mongolian traditional script, it was so damned hard. I struggled with it when I was self-taught, but once I was taught properly it lost a lot of its power to inspire terror and frustration. It has its problems (single letters used to indicate multiple sounds, idiosyncratic spellings, silent vowels), and dictionaries are badly designed, but rest assured, children who learn it in Inner Mongolia grow up perfectly adept in using it. By contrast, Cyrillic, supposedly a more rational, more ‘phonetic’ script, has a LOT of annoying, often senseless, spelling rules, rules that users of the traditional script have trouble mastering. In fact, spelling rules for the traditional script are a lot easier to remember than those for Cyrillic.
One advantage of the old script is that it can accommodate pronunciation differences among dialects. So ᠮᠥᠷᠡᠨ mören or müren ‘river’ can accommodate dialects that use these variant pronunciations. Cyrillic has standardised this to мөрөн (mörön), leaving müren out in the cold. Only if you are a staunch believer in lockstep standardisation can you claim that the change is from ‘suboptimal’ to ‘optimal’. Curiously, the linguistic authorities in Mongolia have used this kind of phenomenon to artificially distinguish between words like tövshin ‘smooth, stable, quiet, peaceful’ and tüvshin ‘level’, which were originally just variant pronunciations. They also flipflop on the ‘correct’ form of words, like tseej ‘chest’, which they recently decided should be standardised as cheej, to be enforced by teaching it as the correct spelling in schools.
There is no doubt that various scripts leave something to be desired — including the spelling of English — but ‘suboptimal’ strikes me as an ethnocentric, even a legacy colonialist way of looking at things.
So ᠮᠥᠷᠡᠨ mören
Horizontalization of a vertical script is an abomination.
> for wheels to work, there has to be a rotating part (wheel, or wheel pair plus axle) and a non-rotating part (actual vehicle), which would by necessity have to be non-connected (for the rotation to be stable without an ever-extending connection), and natural organisms just don’t come in disjoint parts like that.
Couldn’t the rotating part and the non-rotating one be separate organisms in symbiosis?
> 2100+ standard kanji
I’m wondering if there is a more meaningful way to quantify this, such as the number of kanji components, e.g. semantic components and phonetic components. Either way, comparing 2100(+50+50(+26)) to 26, the way it’s sometimes set up, is not fair at all.
The relative complexity of the Japanese and English orthographies is very relevant to me now, since my son is beginning to learn both. One thing that I can say with conviction is that the Japanese one definitely has a smoother (is that the correct term here?) learning curve. Learn 50 hiragana (well, +50 katakana) and you can read anything, assuming it’s been written using hiragana/katakana. Learn the 26 letters and their most common pronunciation, and you can read…. maybe 25% of English words? The fact that the Japanese system has configurable difficulty as a built in feature shouldn’t be underestimated.
Here’s an article from Robert D. Holmstedt, posted this month, that (if I understand it) proposes alphabet creation from Canaanite (city of Byblos) interaction with Egypt:
https://bibleinterp.arizona.edu/sites/bibleinterp.arizona.edu/files/images/B%26IAlphabet.pdf
Much of the vaunted difficulties of scripts like Japanese and Mongolian traditional are due to entry barriers.
Let me take an example. The word diachronic is a highly specialised term in English. It’s not a word that you can spell after learning your 26 letters. To understand it you need to have an idea of the meaning of its components. This means an appreciation of its Greek roots: roughly ‘through time’. If you’re an ordinary person in an English-speaking society, the first connection you make will be to the word ‘chronic’, which is already a partially slang term meaning ‘chronically bad’. You also have to remember the spelling with ‘ch’, which is not the first way you would think of spelling words like this in English but will have become familiar with through the years as one typical way of doing so.
In Japanese, the same word is 通時的 tsūjiteki. This is instantly grokkable (although not manifestly understandable in its technical sense) if you know the characters 通 tsū or 通る tōru ‘pass through or along’, 時 toki or ji ‘time’, and 的 mato ‘target’ or teki ‘ical’ — the only possible reading here being teki. There is a considerable upfront commitment of time and effort required to acquire this script — that’s called ‘education’ or ‘literacy’ — but it underlies much of the Japanese vocabulary and helps people understand and master words like 通時的 with about as much ease as English speakers master ‘diachronic’.
In Mongolian, ‘diachronic’ is цуваа цагийн tsuvaa tsagiin, where цуваа tsuvaa means ‘file, column, procession, trail’ and цагийн tsagiin is the genitive of цаг tsag ‘time’. Again, grokkable but not immediately understandable in its technical sense. The Cyrillic looks like a perfect representation of the spelling, but it embodies a silly little rule. The genitive ending for a masculine vowel is normally -ын -yn, but because цаг tsag ends in a г (‘g’), the rule is that it should be written with the feminine genitive ending -ийн -iin instead. No big deal, of course, but petty. (A different orthography was developed for Buryat, which, if I’m not mistaken, would write the words цуваа цагийн as субаан сагай subaan sagai — a good spelling system, especially if you want to split the language up.)
The traditional script writes this word as ᠴᠤᠪᠤᠭᠠ ᠴᠠᠭ ᠦᠨ chobog-a chag-ün, where ᠤ is ambiguous between /o/ and /u/. Yes, this is confusing, but if you are a native speaker 1) you will already know this word, including the knowledge that it is pronounced with /u/, and 2) you will know that the ending is likely to be read as a long vowel rather than consonant + silent vowel (i.e., ‘g’) because there are tens, maybe hundreds of words just like this. If the first vowel is /u/, the final vowel will almost automatically be read as -aa. The spelling of ц ‘ts’ as ᠴ ‘ch’ might be confusing, but the problem here is that Khalkha has diverged, splitting the consonant into two phonemes, ‘ch’ and ‘ts’. In most Inner Mongolian dialects, all instances of this consonant are pronounced ‘ch’. Cyrillic has clearly been developed to write only one dialect. The genitive ending in the traditional script is perfectly regular for words ending in consonants and does not need to be specially memorised.
The traditional script is, admittedly, more complicated than Cyrillic, but native linguistic knowledge plus familiarity with patterns in the script mean that a person who has been educated in it will not find it so much harder than someone learning Cyrillic. There is a greater barrier to initial acquisition but things get much easier after that. And while it has vagaries of spelling (like English), it does not have petty spelling rules like Cyrillic, of which the use of a feminine ending after the letter г ‘g’ is just a minor example.
Thanks, Stu. So let’s turn the tables, taking our primary notion of machine from living creatures (and suns). When we then attempt to transfer it to non-living contraptions, we find these fall short of the mark by comparison. Thus “trivial machines”
Good point. As long as we don’t throw out the baby, and lose the original categories that distinguish animals from trivial machines.
Yep.
Don’t forget to include the suicides in the school system in the costs, because those are scarily numerous.
And indeed, I’ve seen more criticism of the status quo from inside China than from inside Japan, even though news from Japan should be more accessible to me. What’s your impression? You’ve actually been to both places and talked to people…
Also, in Japan, simplification can creep in slowly: just gradually decrease the amount of kanji and increase the amount of kana. In China, any change couldn’t help being drastic and very noticeable.
Standard Soviet procedure: we come in, decide which dialects belong to the same language and which don’t, pick one dialect of each language, declare it the standard, and develop an orthography for it that contains oddities which only make sense in Russian. (And sometimes also contains misunderstandings of the facts on the ground.) No concept of a pluricentric language; no concept of the fact that, if dialects are mutually intelligible, everyone can just use their own; no concept of the fact that a prestige dialect* or a koiné might develop on its own anyway.
* Let alone several for different purposes, as in Ancient Greek… though I don’t know any other examples of that happening.
Yep, English is really unusual that way among alphabetic and syllabic orthographies.
Is this some kind of attempt to distinguish [g] from [ʁ]?
Is this some kind of attempt to distinguish [g] from [ʁ]?
I’m pretty sure it is.
The genitive of бага is багын, that of баг is багийн.
Two things about Cyrillic.
Search on the Internet and there is always a significant proportion of misspellings. (Try амрасан and амарсан, for instance, which is a pretty mild case. There are much worse, like явъя – correct vs явья – wrong.)
There are quite a few ‘spelling dictionaries’ in the bookshops. They are like ordinary dictionaries but focus on correct spelling and common errors (highlighted in red). I have one that runs to over 600 pages.
I misspoke myself. There are spelling dictionaries in the bookshops, possibly more than one. They seem to be displayed in considerable numbers, which suggests either they are selling well or not selling well at all…
Mongolian short vowels are basically indistinguishable in spoken form. So spellings must be memorized, you won’t hear difference anyway.
‘Mongol’ is actually pronounced as ‘mongl’ or maybe even ‘mngl’ – you’ll have to be aware that there is an ‘o’ there, but you won’t hear it.
Kalmyk orthography omits short vowels altogether, but it’s not quite right – short vowels are actually pronounced, but you simply can’t tell what it is.
Bulgarian short vowel “ъ” is probably close.
Horizontalization of a vertical script is an abomination.
Especially when it’s based on a R-L script and ends up looking like one upside down.
The short vowels are epenthetic vowels. You don’t actually need to memorise the spelling; the epenthetic vowels are predictable in pronunciation and placing (subject to the caveat below).
The biggest problem with the Cyrillic Mongolian script is the treatment of epenthetic vowels in verb forms. Cyrillic has annoyingly complex rules that result in vowels being written where they are not pronounced and vice versa. Амрасан (amrasan) above is a result of this confusion. People who write it thus are influenced by other verb forms in which ‘amar-‘ is indeed written ‘amra-‘ according to the rules of Cyrillic spelling.
amrasan
this is pronounced something like ‘mrsn with ‘ being an indeterminable short vowel (you can hear that there was a vowel, but you can’t hear which vowel).
obviously Mongolians have difficulty figuring out how to spell it – who cares if it’s spelled amarsan or amrasan – you can’t hear a single “a” in it anyway.
Chinese characters do impose a load on the memory
And on the visual system generally. I can’t even find one that I’m looking for in a long list. (Then again, I mostly can’t find anything visually in a cluttered field, Latin-script and to a lesser extent Greek and Cyrillic words being an exception.) Writing one from memory, as opposed to copying it (which I can’t do either, but knowing the stroke order might help) is out of the question.
To understand [chronic] you need to have an idea of the meaning of its components. This means an appreciation of its Greek roots
I don’t think so. The OED says “Pertaining to or designating a method of linguistic study concerned with the historical development of a language; historical, as opposed to descriptive or synchronic.” Of course the etymology is given, but it’s not intertwingled with the definition. To me the word is just a synonym for historical, useful primarily because the contrast synchronic/diachronic sounds cooler than mere-shmere descriptive/historical.
If you’re an ordinary person in an English-speaking society, the first connection you make will be to the word ‘chronic’, which is already a partially slang term meaning ‘chronically bad’.
Well, some English-speaking societies. I have never heard Americans use chronic as merely ‘bad’, and neither have the American dictionaries, which don’t even list it as chiefly Brit., their cautious way of saying “We have no American evidence for this sense whatsoever”. Even UD, which is mostly American, gives only the senses ‘marijuana of a particularly strong variety’ and ‘marijuana laced with cocaine’. It’s true, though, that the associations of chronic are mostly negative, clearly by association with chronic disease, even though in technical language a disease can be both chronic and trivial, like teenage acne.
Sure, writing these words in romaji is ‘easier’ at first blush, but it fails to show these rather important linkages.
No doubt English obliterates important semantic connections by writing king, royal, regal instead of king, kingly, kingish or 王 alone or with two modifiers of some sort.
Horizontalization of a vertical script is an abomination.
And yet verticalization of horizontal scripts doesn’t seem to be a problem, at least on book spines, crossword puzzles, and tall signs. One point is that vertical Latin can run from top to bottom or bottom to top (English and German book spines respectively), whereas all vertical scripts strongly prefer left-to-right when written horizontally. (Ogham is a partial exception, as most vertical examples are bottom-to-top.)
Don’t forget to include the suicides in the school system in the costs
I doubt that would change much, as long as society remains on a one-failure-and-you’re-out basis. The virtue of “If at first you don’t succeed, try, try again” is very modern indeed, showing up in the 19C U.S. and then only in primary education. Bankruptcy, for example, remained a social disgrace until quite recently.
Standard Soviet procedure
Standard procedure for any imperial state. Polycentricity (as opposed to variation) is always treated as an unfortunate accident, definitely not something to emulate.
Mongolian script as upside-down Aramaic, and how it mixes with other scripts.
For foreign learner, this represents a problem only if you want to learn to write in Cyrillic Mongolian correctly.
Good luck, because adult, educated Mongolians spend decades reading and writing in it every day and still make frequent mistakes.
Boy, that sounds like yet another terrible writing system! I know people make do and even prosper with all sorts of terrible writing systems, but one can’t help but wish they were better.
Sorry, my comment turned out wrong (fuzzy thinking).
You don’t actually need to memorise the spelling; the epenthetic vowels are predictable in pronunciation and placing
That should have been: you don’t need to memorise the spelling. The correct vowel is always predictable in writing, even if it is quite indistinct when heard.
The placing of the vowels in writing is also predictable according to pesky rules, which doesn’t always match where they are actually pronounced.
this represents a problem only if you want to learn to write in Cyrillic Mongolian correctly
Well, yes, but Mongolians still have a pretty prescriptive mindset (this is ‘right’, this is ‘wrong’). And yes, despite this ‘educated Mongolians spend decades reading and writing in it every day and still make frequent mistakes’. For a so-called ‘phonetic script’ it is remarkably ‘sub-optimal’.
I once tried to put the spelling of verbs in some kind of order at Spelling of Verbs in Mongolian Cyrillic Script. I adopted the approach of ordering the verbs into ‘classes’. It was the wrong approach because there are still exceptions. You have to remember the spelling according to certain rules relating to the behaviour of certain letters (‘l’s or ‘g’s, etc). Putting the verbs into classes doesn’t do the trick. I can never learn spelling according to that kind of rule. The script, by the way, was actually created by Mongolians. They partly referred to the old script…. And despite its shortcomings, I actually like the old script better…
Glossika Mongolian course has 3000 sentences transcribed in IPA phonetic transcription.
It’s a horror likely to frighten any reader foolishly hoping to learn Mongolian.
Worse than transcription of Johanna Nichols’ excellent Ingush grammar, hell, even Coastal Tsimshian looks pretty in comparison.
Example:
p⁼ʲi ʃɔkʰɔɮadant t⁼ʊrʰtaʲ. xarʲəŋ ʧʰamt? ʧʰʲi ʃɔkʰɔɮadant t⁼ʊrʰtaʲ ʲʊː?
Bi shokoladand durtai. harin chamd? chi shokoladand durtai yu?
I like chocolate. how about you? do you like chocolate?
And funnily enough, this intimidating IPA phonetic transcription still doesn’t reflect actual pronunciation of spoken Mongolian.
Maybe it reflects speech of TV anchors who make a point to pronounce every word as it is written, but average Mongolian doesn’t speak like that – you won’t hear any of these ɔ, a, ʊ, aʲ vowels in actual speech.
This is a slightly strange transcription, partly phonemic, partly showing minute phonetic detail.
ʃɔkʰɔɮadant should have the final vowel reduced to schwa (or something) — a problem you just pointed out above.
Another problem is the treatment of aʲ, which is actually pronounced /æ/.
But the second /t/ in t⁼ʊrʰtaʲ looks like a meticulous rendering of pre-aspiration (ʰt).
yet another terrible writing system!
Is kesklinna the genitive (‘of the downtown’, with a long nn) or the illative (‘to the downtown’, with an overlong nnn)? That’s Estonian for you. To be sure, if it’s a road sign, it’s the latter, but still…
I have never heard Americans use chronic as merely ‘bad’
It’s used that way in Australia. My bad.
all vertical scripts strongly prefer left-to-right when written horizontally
Only through convention. Right to left was common pre-war in certain contexts.
No doubt English obliterates important semantic connections by writing king, royal, regal instead of king, kingly, kingish
You are talking different historical backgrounds. English never developed that way. Japanese did (exploiting the fact that Chinese characters had a system of multiple readings), and it is an element in its vocabulary development. It would be difficult to create an ex post facto system for English vocabulary that could match the Japanese one.
indeterminable — now, that is admitting defeat. We call it indeterminate because we have determined that that is what it is.
>Here’s an article from Robert D. Holmstedt, posted this month, that (if I understand it) proposes alphabet creation from Canaanite (city of Byblos) interaction with Egypt:
HIs criticisms are much more compelling than his theory. I’d first underline a point that contradicts a lot of what is written here:
>The fact that the Byblians had already created a new, more much abstract writing system with the syllabary by the beginning of the second millennium suggests that the Byblian context had the sufficient conditions for writing innovation. Indeed, as revolutionary as we may consider it, the alphabet is but a simplification of syllabic writing:
Of course, the Byblian obligate syllabary is the more important development, not the later abjad. It’s hard to understand why anyone would think otherwise except a sort of ethnocentrism of alphabet-using moderns. Abstracting language entirely away from logographs is a big step. From there, dropping vowels, and later picking them back up, is trivial. Although frankly, the development from rebus to obligate syllabary is also stepwise. This is why the idea of an “inventor” of the alphabet or even the abjad is a little silly. They’re all standing on the shoulders of giants. The only real invention I can see is the rebus itself. That is the moment of abstraction from picture of thing to representation of sound via an early graphic pun. The rest is an extension and refinement of the initial principle.
As to who actually developed the alphabet, I think Holmstedt is interesting but ultimately not convincing. After cogent criticism of the Petrovich “Hebrews” theory for its reliance on names that are not Hebrew, but broadly West Semitic, he then falls back on the idea that the word Baalat = lady must be the Biblian Baalat. Perhaps. But it seems likely that any West Semitic speaker who encountered Hathor might well have called her Lady.
His strongest point in favor of Byblos is that the Wadi el-Hol and Serabit el Khadem graffiti are surely just remnants of a much larger and earlier corpus on perishable materials. It’s doesn’t even make sense to treat them as a terminus ante quem. Their content is not the content of the developers of a communication system; their medium is the most likely to survive rather than the easiest to experiment with; and their separation from each other by at least a century and probably more is statistical evidence that they are lucky finds in a longer arc of development.
If the abjad development predates the earliest rock inscriptions by decades or centuries, we’re in a time when Byblos as the most powerful West Semitic trading partner of Egypt becomes a likely candidate. If perishable material is the most likely medium for experimentation, then Byblos is well placed as an importer of papyrus, and they do have a goddess called, like many other goddesses, Lady.
It doesn’t seem definitive, but certainly interesting.
But you’re left with why question of why the northern coast drops the curvilinear abjad and instead develops another, cuneiform-based alphabet that it holds onto for centuries (the Ugarit alphabet), while others run with their abjad. I guess you could find reasons in varying relations with Egypt and Mesopotamia and in the availability of papyrus vs. appropriate clay.
I tend to think eventually, someone will delve deeply enough into the right tell to find early alphabetic texts on clay, and then we’ll know.
Is there any West Semitic goddess (or god) who we have reason to believe was not routinely referred to as “baalat” (or “baal”)?
YHWH
From there [syllabaries], dropping vowels, and later picking them back up, is trivial.
This strikes me as profoundly wrong. It only seems trivial in hindsight. This “trivial” insight has occurred to precisely zero other users of syllabaries, and there is a good reason for that: it involves abstracting away from speech to units which cannot be uttered in isolation at all. The vowel/consonant dichotomy itself is a far from trivial linguistic discovery; to miss that is to betray a deep lack of historical perspective. It’s on a level with supposing that it’s obvious that the world is round.
units which cannot be uttered in isolation
Mmmhm.
(Related, probably, somehow.)
Pfft.
Some units which cannot be uttered in isolation …
Still, it represents a major increase in abstraction over a syllabary; and I don’t think that abstraction is obvious except in hindsight. (If it were, that step would have taken place commonly among syllabary-users.)
The Ugaritic script is a “semi-syllabary” – just. It distinguishes vowels after aleph.
I’ve just read the Holmstedt paper. I agree that his criticisms is better than his arguments.
The syllabary argument is no argument for Byblos specifically as the cradle of the alphabet without a clear derivation of one system from the other.
His graffiti argument is refuted by his own criticism, in which he recognizes that the importance of inscriptions is an artefact of conservation. It’s also unnecessary since, as he says, most writing in a trading port would have been on perishable materials.
The Ba’alat argument, on the other hand, refutes itself. The qualifier gbl “of Byblos”, used for the female godess in Byblos, tells me that the title “Lady” was not exclusively associated with Byblos. (But it led me to a fun thought. What if Pallas Athene were a calque of the epithets of Phoenician city godesses?)
That said, the Phoenician trading ports aren’t any less interesting even if Holmstedt makes a weak case for Byblos in particular.
I’ve seen buses in China where the writing goes front-to-back: left to right on the left side, right to left on the right side!
One reason it looks scary is the unique use of p⁼ t⁼ for what I’d write b̥ d̥. Another is the very abstract use of aʲ for, as mentioned, æ. Once these are out of the way, it all makes sense to me… except ʲʊ.
I would say that Japanese RTL is really TTB with single-character columns, since the page progression is RTL. Most Ogham stones are written BTT on the edge, but some go BTT, then LTR across the stone, and if necessary TTB on the opposite edge! Given that all manuscript Ogham is LTR, I think it really makes more sense to see it as a natively LTR script that is able to go vertical in either direction, like other horizontal scripts.
The Unicode vertical text model characterizes each character as being always upright whether vertical or horizontal, rotated 90 degrees when vertical, or replaced by a different glyph when vertical with a fallback to either of the other two positions if no vertical glyph is available. A character with a diacritic is treated the same as its base character, unless the diacritic is an enclosing square or circle, in which case the combination is treated as always upright. Sometimes it’s contextual: an English word in Chinese context is rotated, but a Latin-script acronym also used in Chinese typically uses upright letters.
(Related, probably, somehow.)
Nice, what a beautifully stubborn approach to consonants. “Only stops really deserve to be recognized as consonants”, right?
Speaking of stubborn things, it also made me realize that there was nothing unique about the stubborn survival of the Basque through the Iron Age and into the early centuries of the Roman era.
That all along Spain’s Mediterranean coast, non-IE Iberian languages were just as thriving, despite (or maybe because of) this area being overrun by the descendants of the Steppe pastoralists in the Bronze Age. I realized that the Iberian languages with their semi-syllabary writing system shared quite a bit with the Basque, including a whole set of numerals, a set of suffixes, and elements of toponymy … but of course too little is know about the Iberian to make a conclusive case about its genetic relationship with the Basque. But in terms of plain DNA genetics, both peoples were essentially identical through the Iron Age (half derived from the descendants of the Steppe migrations overall, and nearly fully descended from the Steppe on the male lines). Does the DNA equivalency make the case of the common roots of Iberan and Basque a bit more attractive? Does it hint that both languages were invasive to the Iberian peninsula (only the former didn’t survive the massive colonization of the Mediterranean coast in the Carthaginian and Roman times) ?
Would code cracking be more difficult in a syllabary? Seems like it would. I’m not offering that as a reason for keeping a syllabary. It was just an intriguing thought.
It seems to have begun as a misinterpretation of the Phoenician alphabet. The letters for plosives were misunderstood as CV syllables because they couldn’t be pronounced otherwise.
Would code cracking be more difficult in a syllabary?
Michael Ventris managed it.
Japanese RTL is really TTB with single-character columns
Do you have a particular basis for this statement?
And why do you think that LTR is a more natural choice? Would it perhaps be that that is the model that they adopted from the West? Or is something more profound at work. Please enlighten us.
This “trivial” insight has occurred to precisely zero other users of syllabaries, and there is a good reason for that: it involves abstracting away from speech to units which cannot be uttered in isolation at all.
Perhaps it was all a huge mistake. Perhaps syllabaries are a more natural way of analysing speech sounds. Perhaps the phoneme isn’t as natural a unit as phonologists think it is. (Someone at LanguageLog suggested something like this.) Chinese, Japanese, Korean, traditional Mongolian… all traditionally worked with syllables.
Syllabaries are certainly more natural in the sense that when people progress to writing sounds as opposed to meanings that is what they nearly always produce if they haven’t encountered the Idea of the Alphabet. But Japanese is a kind of counterexample: it lends itself to syllabic writing more than most languages, but this creates contortions in the traditional analysis of verbal morphology because the script has no good way of writing consonant-final verb stems.
Korean writing is an alphabet, surely, unless you mean Hanja? And isn’t the traditional Mongolian script fundamentally alphabetic too?
And why do you think that LTR is a more natural choice? Would it perhaps be that that is the model that they adopted from the West?
Of course the Han-derived scripts are written LTR because of Western influence. If they came heavily under Arabic influence instead, they might have been written RTL. But when is RTL in fact used? One context is newspaper headlines, which can be thought of as vertical (as the text below them normally is) but limited to a single row rather than being multi-row columns.
As I mentioned, LTR is now outdated, although still occasionally found on sides of buses, etc. But your stated “strong preference” for left-to-right is just a result of historical factors, not inherent preference.
As for “Japanese RTL is really TTB with single-character columns”, that is just nonsense. It came about not because people were writing in “single-character columns”, but because there is, in fact, no inherent preference for either direction. Indeed, RTL was conceivably adopted in those more nationalistic days because it was regarded as more “natural”, or (what is the same thing) less Westernised.
Korean is alphabetic, true, but it combines letters into square shapes that correspond to syllables. Perhaps it combines the best of both worlds, but the nod to syllabicity is significant.
Traditional Mongolian script can be decomposed into letters but is taught to Mongolian children as syllables, set out in syllabic tables. I can assure you, it makes the script much easier to learn than learning it letter by letter, which is extremely confusing — you might understand this better if you look at the script and think how you’re going to decipher it.
Unicode adopted the ‘letter-by-letter’ approach to represent the script, which has led to a lot of problems with implementation. Inputting the traditional script would be far easier if they had adopted a syllabic approach. Input methods in Inner Mongolian adopted the syllabic approach, setting aside private-use code points, and they work a lot better than Unicode.
The traditional Chinese phonetic analysis splits syllables into initials and finals. This is enshrined in bo po mo fo and is certainly not harder than pinyin. In fact, it avoids the problem that occurs when ‘iou’ is abbreviated to ‘iu’ in pinyin, which leads many foreign students to pronounce it incorrectly.
For RTL, see https://japanese.stackexchange.com/questions/13076/under-what-circumstances-does-japanese-read-from-right-to-left
However, I am extremely dubious of Sjiveru’s assertion that “horizontal text was treated as a single row of vertical text. This meant that since you start on the right when reading vertically, you started on the right here as well.” This interpretation suggests that LTR is some kind of “universal norm” whereas RTL is an aberration brought about by people who streamed their ordinary orientation in the wrong direction due to some abstract schema.
Snailboat’s explanation is better: “Historically, there are examples of R-to-L text that are clearly not a special case of vertical text, e.g. this image, which contains both L-to-R and R-to-L. The explanation I remember reading is that both L-to-R and R-to-L appeared around the same time, but R-to-L never became popular”.
John Cowan: “No doubt English obliterates important semantic connections by writing king, royal, regal instead of king, kingly, kingish”
That is indeed a criticism of English by Esperantists such as Claude Piron: moon/lunar, city/urban, etc. impose an extra learning cost. All respect to Piron, but I don’t buy that argument in the context of basic communication and finding hiking buddies around the world. You don’t need lunar for basic communication, you can use moon’s or a phrase. It *does* apply in the context of *scientific* communication: you do need both moon and lunar to write a scientific paper, and that’s a cost with no benefit. The benefit is to literature and poetry, where the many lexical strata in English offer choices of register, associations and sound. There’s an important contrast between “Whispering lunar incantations / Dissolve the floors of memory” and “The moon has lost her memory”.
Would code cracking be more difficult in a syllabary?
A lot depends on the type of language (e.g. noun and verb morphology), the type of code (presumably we’re assuming a straight substitution cipher, but it’s not very clear), and especially the completeness of syllabary – does it have loads of signs for complicated syllables, like Yi, just assemble everything from CV syllables, like Cypriot (not sure if there are any extant examples), or only have signs for CV because that’s what the language has, like Japanese?
In some situations, a Ventris-type method would arguably make decoding easier – though I’m not sure whether it would actually be easier than cracking an alphabetical code (even without frequency analysis), unless we have a bunch of short regular texts (which would admittedly arguably be the kind of thing usually transmitted in code).
the problem that occurs when ‘iou’ is abbreviated to ‘iu’ in pinyin
A more infamous problem is abbreviating ‘wei” to ‘ui’. Because it sounds like the word khuy ‘penis’, it makes surnames like Hui and words containing that element practically unusable in Russian. Anhui becomes Ankhoi, and so on.
All that depends on the language. There’s a lot of literature on where the syllable boundaries in English even are, including whether there’s even a single answer to that, and what that means for the phoneme inventory (the minimal pair nitrate vs. night rate comes to mind). As I did in the LL thread you’re probably thinking of, I recommend starting here.
Me: the results are awkward and suboptimal and the Japanese suffer for two millennia
@Bathrobe: I admit that I added that point for punch, not for truth. The discussion that followed has been very enlightening. What hadn’t occured to me before is that the Japanese hybrid system may actually be conceived as a way to unite the best of two worlds, a logographic system for roots and concepts and a syllabary for clitics, particles and suchlike. I don’t think Japanese is actually there, but conceptually, maybe not that far away.
Back to the Levant, even if I think Holmstedt fails to make his case, the paper’s worth reading for the ideas, and we need to keep trying to understand the remarkable innovation of writing in the early -2nd millennium Levant. Something in their economy, culture and language led them to ungrok and rehack the concept of writing, not once but at least three times. Suggestions:
1. The existing logographic models of Egyptian hieroglypics and (by then) Babylonian cuneiform were difficult to adapt to Semitic languages.
2. The investment in training of logographic scribes wasn’t worth it in the multicentric and plutocratic societies of the coastal Levant.
Point 1 is sort of trivial, but it’s true about all but the most isolating languages, and nevertheless logographic scripts were adapted to anything, and all the neighbouring languages using logographic scripts were related and typologically similar to West Semitic.
Point 2 is more promising, but it’s also a just-so-story with no specific evidence to go for it beyond making sense. The idea would be that merchant houses, ships and caravans were all small operations, and the wealth and political position of their owners came and went with the luck of their latest expeditions. They needed clerks, but being unable to plan years ahead they would naturally develop a system that any bright kid could hack their way along in without much formal training. City authorities would also need officials on the docks and city gates, but as soon as there was a class of trained clerks around, they would recruit from that.
Also, or maybe as a counterpoint, the West Semitic languages ended up using an abjad, which on the face of it is particularly ill suited to represent the important morphological processes in those very languages. It’s like they went out of their way to avoid blurring the image of the root. Could the abjad be what happened to he syllabary when it was taken up by officials trained in recognizing the image of the word and supplying all phonology by context and rote learning?
@David Eddyshaw: The most prominent surviving West Semitic deity is (among others titles) “baal g’vurot,” lord of wonders, in the liturgy.
>2. The investment in training of logographic scribes wasn’t worth it in the multicentric and plutocratic societies of the coastal Levant.
Imagine the poor scribe in a trading group that dealt with both Egypt and Mesopotamia. It’s one thing to learn a logographic system, and/or to be tri-lingual. But imagine needing to be tri-lingual and learn two logographic systems!
On the other hand, think of the job security …
(Scribal desire to maintain a monopoly of literacy is the only possible explanation for the Pahlavi writing system, easily the worst ever to use an alphabet.)
Me three. The throwaway suggestion that the alphabet was intended as an improvement on the Byblian syllabary is really, really hard to imagine (and indeed there is no precedent).
Even the criticisms of Goldwasser’s hypothesis aren’t all that solid. Based on what little evidence we have, the proportion of the population that was literate in ancient Canaan fluctuated wildly; any amount of political and economic upheaval will do that. So why should illiterate mine workers have been too stupid to put themselves in something similar to a Sequoyah situation? The very fact that the letters are hieroglyphs used not for their Egyptian, but for their Semitic pictorial sound values argues for such a situation – even though not exactly the same situation, because the principle that signs stand for consonants was transmitted.
And that principle, to go back to the OP, comes straight from the logographic principle. Early in the development of a logographic script, it isn’t really language that is represented, but mainly just nouns and verbs. Morphology is not written. In China that meant the pre- and suffixes were ignored and can only be reconstructed nowadays; in Egypt it meant the vowels were ignored.
=====================================
The Ugaritic alphabet has been mentioned. I thought that’s just the alphabet, in a font designed for clay tablets?
The Ugaritic alphabet has been mentioned. I thought that’s just the alphabet, in a font designed for clay tablets?
That’s what it usually taken to be, though I can’t recall having ever seen a convincing derivation of Ugaritic from Proto-Canaanite (or, come to think of it, any derivation of Ugaritic from anything at all – it seems to be usually ignored in discussions of history of the alphabet).
The oldest known abecedaries (inscriptions listing the alphabet) are Ugaritic, but, again, it’s perfectly possible that Proto-Canaanite (or even older) ones existed and just didn’t survive (or even did survive but have yet to be found/excavated/identified).
As far as the Byblos syllabary is concerned… well, as far as I can tell, it’s a complete mystery. It does kind of look like something intermediate between Egyptian and Phoenician, but it has yet to be deciphered (and may never be, because the corpus is so small).
It’s possible, I guess, that there were several independent Sequoyah-esque events under similar pressures, and the Byblos syllabary is one while Proto-Canaanite as we know it is another. Such an explanation might fail Occam’s razor, however.
The Ugaritic system is nearly but not completely an abjad; it deviates only in having three alephs, depending on whether the following vowel was a, i or u.
There are actually two different sets of abecedaries in Ugaritic: most commonly the familiar aleph-beth-gimel, but also the Old South Arabian/Ethiopic-order one h l ḥ m …
…which means, from Greek onwards, e l h m. If n is next, you just solved the riddle of where elementum comes from.
Oh, nice! … but no prize. Next in line is a q.
Can’t find the thread with mash ‘mung beans’, so am posting this here:
Steblin-Kamenskiy fils links the word mash with māṣa, considered a Wanderwort.
Elements. Michael David Coogan has written two (or more?) excellent articles
showing that the Latin elementum then English element probably came from L-M-N
in several alphabets. “Alphabets and Elements,” Bulletin of the American Schools
of Oriental Research 216 (1974) 61-3; “)LP, To Be an Abecedarian,” J. of the
American Oriental Society 110.2 (1990) 322. His work has been cited
approvingly, e.g. by Victor Avigdor Hurowitz in articles on Hebrew acrostic
Psalms and Wm. Hallo in his book on Ancient Near East Origins. And before
Coogan, F.A. Wolf, J. B. Greenough (Harvard St. Class. Phil. 1 [1890] 97-99), and
others supported this origin. But the Oxford English Dictionary etymology offers:
[a. OF. element, ad. L. elementum, a word of which the etymology and primary
meaning are uncertain, but which was employed as transl. of Gr.
στοιχεῖον in the various senses: < a component unit of a series; a constituent part of a complex whole
(hence the 'four elements'); a member of the planetary system; a letter of
the alphabet; a fundamental principle of a science.]
Etymology books by Klein and Shipley propose an origin from Greek for elephant,
ivory letters. But they provide no good examples.
Can’t find the thread with mash ‘mung beans’, so am posting this here
That’s the GAN: WHODUNNIT? thread; I’ve copied your comment and moved the subsequent one there.
Thanx!
Has anyone looked at the Brian Colless theory that the earliest alef-bet WAS a sort of syllabary, where the consonantal sign was rotated to indicate the vowel.:
http://cryptcracker.blogspot.com/2014/04/early-hebrew-syllabary.html
I don’t have the knowledge to assess that, and I’ve sometimes suspected crackpot, but he seems to be a professor with some peer-reviwed articles.
I also don’t have the knowledge to assess Brian’s proposal on this, but from reading and correspondence on other subjects, I’ll say he’s certainly a researcher of substance.
That very post says it’s not the first. Colless calls it the neo-syllabary, distinct from the Byblian one.
Lots of interesting ideas in there; I wonder if there are enough inscriptions to do statistics with (my impression, from this one post, is no). The variation in letter shapes and orientations may simply mean that these features weren’t fixed yet (as the direction of writing was not); as long as the letters remained recognizable, anything goes. Quite similar things are seen much later with Germanic runes.
Canadian syllabics, on the other hand, is an abugida (better term than alphabet in this connection) where the vowel or lack of it is encoded by rotation (for the vowel) and small size (for the lack of one). Fortunately, most of the languages that use it have only four vowels.
>from reading and correspondence on other subjects, I’ll say he’s certainly a researcher of substance.
Thanks. I had read through the Crypt Cracker blog a while ago (hence my failure to remember that he wasn’t actually speaking of the original alef-bet when he wrote about an alef-bet syllabary.) He takes things like Phoenician relics in Jamaica or Texas more seriously than most, and that made me wonder. But Phoenicians exploring the Atlantic could have gone off course, so who knows. Some suspension of disbelief seems reasonable.
If there are splitters and lumpers in biology, then the study of ancients scripts has literalists and doodlists — those who see letters on every ancient ostracon, and those who see them nowhere. He is definitely a literalist by this definition. If three lines meet on a potsherd, it’s a text!
He also has a lot of interesting ideas. There may be few people in the world really able to evaluate some of his ideas, and I’m not anywhere close to being one of them. Glad to know he’s worth considering.
Wow. Really great comments thread.
Here are a few comments of my own now, that few will ever read.
It’s mostly a feedback loop of technological advances that allowed writing to come into being.
It is completely reasonable that writing only spontaneous came about a very small number of times. However, it makes sense that probably some early versions just died out without anyone else seeing it (because it didn’t provide a large enough benefit to the community, and didn’t spread anywhere else).
However, when people started moving and trading goods over large distances (metals, wheels, horses, etc.) the economic benefits of even rudimentary writing were magnified. Civilizations that had basic writing and record keeping were more successful at trading, and thus traded more with foreigners, who learned about writing, and traveled more, and brought home more technologies. And the seeds spread quickly because people interacted more quickly.
As the best biological example. 100% of known life in the universe came from a single event, that was just successful enough to not go extinct, and mobile enough to spread and multiply.
Good point. Mute Miltons …
No use having a brilliant idea if nobody around you can see any point in it. (Perhaps that’s why the inventor of the alphabet must have been, after all, a man, for all my pointedly leaving the matter open above.)
I just decided to buy the Wilkinson book. Thanks for the recommendation.
One question. I bought it as a Kindle on Android book. Pagination is more than a little weird in Kindle on Android. I’ve read the intro, which puts me on p. “133 of 10300”. Can you tell me how many normal book pages this will be?
A .pdf file has 526 pages.
Amazon sez: “Hardcover: 656 pages.”