Hieratic — 3200 BC, Egypt

Hieratic "A"

Hieratic was a version of Egyptian hieroglyphics that developed in parallel with hieroglyphics.

While the question of “is it a writing system or isn’t it?” has come up several times already in these postings, hieratic is interesting because it brings up a slightly different question: “is it a distinct writing system or isn’t it?”

Hieroglyphics were used for relatively formal use, and frequently carved in stone.  Hieratic script was written with a brush and ink on papyrus or cast-off pottery or limestone shards for less formal texts. There is almost a one-to-one correspondence between hieroglyphic glyphs and hieratic glyphs, but they don’t look very much alike.  They look like different writing systems.

However,  cursive and printed Latin script don’t look very similar, either:

Cursive and printed Latin "r"s

It turns out that there are a number of writing systems that have similar-but-not-quite-the-same characters.  For example, the Chinese glyph for fire looks like a person with their arms up.  The version of that glyph most commonly used in Japan is similar, but the angle of one of the “arms” is flipped for the version most commonly used in Japan:

Chinese and Japanese versions of "fire"

Just as an English-reader would probably say that the cursive “r” and the printed “r” are the same character, Japanese and Chinese readers would probably agree that the “Chinese” and “Japanese” versions of “fire” are the same character.

Now look at these.  You might think they are from the same writing system:

Greek, Latin, and Cyrillic "A"s

In fact, they are not.  The left one I got from Greek, the middle from Latin, and the rightmost from Cyrillic.  You might say, “ah, but they are the same basic character because they represent the same basic sound, a vowel that is pronounced something like ‘ah’.”

That doesn’t work for this glyph:

Cherokee syllable "go"

This character is not a vowel pronounced something like “ah”, it is a syllable pronounced “go” from Cherokee.  This is clearly from a different writing system than Latin (or Greek or Cyrillic).

Thus, design similarity does not determine if something is a distinct writing system or not.  Hieratic looks very different from hieroglyphics, but I think it is nonetheless not a separate writing system.  It’s more like a crazy-different font.

Links: Wikipedia, Ancient Scripts, Omniglot

Posted in Abjad, Evolved slowly from parent, Rating: 4 "Huh, interesting!" | 4 Comments

Elamite cuneiform — 2200 BC, Iran

Elamite cuneiform determinative for people

Linear Elamite petered out after a few hundred years, ignored in favour of more prestigious languages/writing systems like Babylonian (which was an organic descendant of Akkadian in both language and script).  When the Elamite language reappeared after a few hundred more years, it was written in a modified version of the Akkadian cuneiform.  The number of signs varied over time, but there were never more than about 130 signs in use at any one time.

More of Elamite cuneiform has been deciphered than of Linear Elamite, in part because of similarities to Akkadian cuneiform, but also because of some very well preserved, long texts in multiple languages.  The most famous example is some big honkin’ inscriptions at Behistun, Iran that are 100m up a cliff (i.e. relatively safe from vandals) telling of Darius the Great’s exploits.  These inscriptions were not only important for deciphering Elamite cuneiform, but also for deciphering Old Persian cuneiform.

The upright wedge glyph at the beginning of this post is a “determinative” in Elamite cuneiform.  Most languages that were heavily logographic also had phonetic aspects that make it frequently ambiguous.  A glyph might represent a word/idea/thing, or it might represent a sound.  (For example, you can imagine a writing system for the English language where a drawing of a bee might mean either the sound “B” or the little insect that buzzes around.)  To give the reader some clues about the adjacent word, many heavily logographic languages had additional, unpronounced symbols called determinatives which classified the adjacent word.

The determinative glyph at the beginning of this post is for people: proper names, names for classes of people (like “king”), relatives, and some pronouns.  Sumerian cuneiform also had determinatives for such things as “trees and things made of wood”, “rivers or canals”, and “buildings or temples”.  Egyptian hieroglphics had determinatives for almost everything, like “vine, fruit, or garden”, “fire, heat, or cook”, or even “bodily discharges”.

Thus, to read languages like Elamite, you needed to know a very large number of logograms, syllables, and determinatives, some of which were ambiguous.  You had to — on the fly — figure out whether to pronounce the characters or not, and if a symbol represented a word or sound.  I will never complain about English spelling again.  (Addendum: I lied.  I have complained about English spelling again.)

Links: Wikipedia, Ancient Scripts, Omniglot

Posted in Logograms, Rating: 5 "Whoa!!", Syllabaries, technology influenced | 1 Comment

Linear Elamite — ~2200 BC?, Iran

Linear Elamite “shi”

Linear Elamite — also called “Old Elamite” — was the outgrowth of proto-Elamite in modern-day Iran.

We tend to think of languages as either “deciphered” or “undeciphered”, but Linear Elamite is currently partially deciphered and likely to stay that way until more artifacts are uncovered.

It has been difficult to decipher because there are only twenty-two texts as of this writing.  In addition, Elamite appears to be a language isolate.   However, some of the artifacts have long sets of text in both Linear Elamite and Akkadian cuneiform, which helps.

There are about 80 glyphs in the twenty-two existing texts, which is more than most syllabaries but way less than most logographic languages.  It is likely that it was mostly syllabic, but with some logograms here and there.

Links: Wikipedia, Ancient Scripts, Omniglot, Cuneiform Digital Library Initiative

Posted in Logograms, Rating: 1 "Dull, only here for completeness", Syllabaries | 1 Comment

Proto-Elamite — 3000BC?, Iran

Undeciphered Proto-Elamite character

Proto-writing happened over an extensive area over thousands of years.  While proto-writing developed into “proto-cuneiform” in modern-day Iraq and from there into Sumerian cuneiform, it developed into proto-Elamite in Iran (and from there into Elamite).  We don’t hear much about Elamite because the Elamites weren’t around for a hugely long time, and they have been quiet for a very very long time.

Like proto-cuneiform, what we know of proto-Elamite is from clay tablets that appear to have had things pressed into them.  Some glyphs seemed to be made using reeds (like the one adorning this posting), and others appeared to be made using small round things that were pressed into the clay.

The fact that there were two proto-writing systems that experts can tell apart suggests that writing might not be that unusual an invention.  Or, perhaps the jump from no symbols at all to protowriting was more important than the jump from proto-writing to writing.

Proto-Elamite is undeciphered, but appears to mostly have been accounting records (like Sumerian).

It is less linear than its direct successor, Linear Elamite, in both how the glyphs are shaped and in how they are laid out.  Not only do some of the marks look like they were made with round things (not linear things), but they aren’t really laid out in a line like English text.  Sometimes they are, but sometimes they are laid out in a line of boxes, where each box has a glyph made out of lines (like the above glyph) and some pressed-round-thing-shapes.  If you think of the nature of receipts and financial record-keeping, this makes sense: you might have a box for “3 goats” and another box for “6 skins of beer”.  The goats and beer don’t really have an order to them.  This is very different from a sentence of “the three goats drank six skins of beer”, where the sequence of the goats and beer is important, so you don’t have the beer drinking the goats.

Links: Wikipedia, Ancient Scripts, Omniglot, photos of Elamite artifacts

Posted in developed by illiterate(s), Logograms, Rating: 4 "Huh, interesting!", Undeciphered | 1 Comment

Quipu — 3000 BC?, Peru

Quipu

The Incan empire was very large, very powerful, lasted for a very long time, yet had no writing system — apparently.  This seemed odd.

However, even the conquering Spaniards recognized that the Incas did have a way of storing accounting information.  The Incas tied knots into cord in a particular pattern which encoded numbers.  These macrame objects were called quipu (or kipu or khipu or quipo).  People forgot how it worked for quite a few centures, but they have recently figured out how the numbers were stored.  Due to fortunate redundancy in the quipu, there is extremely high confidence that they  understand exactly how the numbers are encoded.

The encoding system for numbers is so well understood in quipu that it is also absolutely clear that there are some knots which do not represent numbers.  There is speculation that these might encode non-numeric information.

Once researchers started thinking about the cords as possible writing, they started seeing many things that could be conveying linguistic information.  Quipu had physical artifacts strung onto the cord (e.g. shells) which didn’t make much sense in the context of numbers.  There were also multiple levels/branches of cord.  Some knots were tied right-over-left, some left-over-right. The twist in the cord could be clockwise or anti-clockwise.  The cords were attached to other cords in two different ways.  Finally, the cords were made of different materials and were sometimes dyed.  All these variations could conceivably encode information.

In 1996, a manuscript was uncovered which purported to explain non-numeric quipu symbols.  The authenticity of the manuscript has been challenged, as the owner will only let one research team look at it, but it suggests that something was woven into the top of the cord representing a (multisyllabic) familiar word, and subsequent numerical values referred to the number of the syllable in that word.

There is obviously more work to be done here, but there is quite a lot to work with: more than 700 quipu are currently known.  This is an exciting time for quipu research!

Using knots to record numbers and perhaps also words is not unique to Quipu.  Okinawans used a very similar system of knotting rice straw called warazan or barazan until the early 20th century; there are reports of knotted number aids being used by Hawai’ians and the Maori of New Zealand.

Even wampum belts in what is now the Northeastern US/Southeastern Canada apparently communicated information.  A wampum belt was not just a collection of beads, but a story.

Links: Ancient scripts, Wikipedia, Khipu database project, Wired, Facebook

Posted in developed by illiterate(s), first in its area, Rating: 5 "Whoa!!" | 1 Comment

Aztec — 1400 AD, Mexico

Aztec "snake"

There are no pre-Columbian Aztec books remaining: the colonists managed to get every last one.  However, there are a few books written by Aztecs left from the period after the Spaniards occupied the land, with translations/explanations written in Spanish.

The Aztec writing system was very similar to Mixtec.  Neither had a syllabary; both used text to support pictures instead of using pictures to support text; both used visual puns to represent words, especially city names.

lowercase "g" with descender a different colour than the body

This doesn't happen

Aztec used colour as an integral part of the communication, which is very unusual.  In English writing, the letters can be in different colours, but the colour is usually not meaningful.  (One exception is on the Web, where hyperlinks are conventionally in blue.  However, changing the colour of modern glyphs does not change the meaning of the word.)  Latin letters also (almost) never have different colours within that a letter.  The descender of a “g” is almost never a different colour than the “o” part of the “g”, for example, and even if it were, it would still be a “g”.  However, in Aztec, glyphs were multi-coloured, and changing the colours could sometimes change the meaning.

In modern writing systems, we are used to text being linear.  Latin script is read left to right (then top to bottom); Arabic script is read right-to-left (then top to bottom); Chinese and Japanese are sometimes read top-to-bottom (then right-to-left). However, in cultures that read left-to-right, it is assumed that if you have a bunch of pictures, you start at the left and move to the right.  The reading order affects us readers so much that we European-language-readers unconsciously tend to start looking at any scene the left and moving to the right.   Arabic readers look at the right first and then move left.  This tendency is so strong that in Hollywood movies, monsters almost always appear from the right to increase the surprise.

However, Aztec (and Mixtec) look more like annotated pictures, and like annotated pictures today, don’t have a clear order.  If you don’t have a cultural bias to read left-to-right or right-to-left, and if your writing does not have an inherent order, how do you tell your readers which direction the narrative proceeds?  In Aztec, they sometimes solved this problem by showing footprints leading from one point to another, in the same way that we might use arrows.

Links: Ancient Scripts, Wikipedia

Posted in Logograms, Rating: 5 "Whoa!!" | 4 Comments

Mixtec — 1200 AD, Mexico

Mixtec writing

The Mixtec writing system is from Southern Mexico, very close to the Mayan civilization in distance and contemporaneous.  The Spaniards only did half as good a job destroying the Mixtecs books — there are a whopping eight pre-Columbian Mixtec books still in existence instead of just four like the Mayan books.

Mixtec doesn’t really look like writing in the way that Westerners think of writing.  It’s not that the shapes of the glyphs is odd — you should be used to that by now if you’ve been following this blog.  Rather, the text seems to be in service of enhancing pictures (mostly with dates and names), while we are used to our writing being mostly text, occasionally with pictures in the service of enhancing the text.

Text serving pictures is not unknown in the Western world.  Go into practically any old church in Europe and look at the stained glass.  You will see pictures that tell stories, with a little bit of text on them to enhance or clarify the pictures (mostly with names).

Now, it might be that the Mixtecs had lots of books where text dominated, but that those were preferentially destroyed, and the only ones the Spaniards left were ones that they felt were just innocuous picture books.

I think that is unlikely.

  • First, the Mixtecs kept writing after the Spaniards showed up.  There are not a huge number of post-colonial books, but it’s more than eight.  These post-colonial books frequently had Spanish annotations/translations written next to the text/figures they represent, so it’s not that we don’t understand the writing system.
  • Second, the Mixtec writing system was strongly logographic: they did not seem to have a syllabary amongst their symbols.
  • Third, many of their symbols are more like what we would think of as picture wordplay than writing.  Many place names, for example, were drawn as little pictures that evoked the name of the place, sort of as if I drew laundry hanging on an electrical cord over a 1000-pound weight everywhere where I wanted to talk about “Washington, DC”  (washing+ton+DC [current]).

Links: Ancient Scripts (excellent), Wikipedia, nice writing sample/example

Posted in Logograms, Rating: 5 "Whoa!!" | 2 Comments

Maya — 250 BC, Guatemala

Maya "Kopan" logogram

Maya script is a very rich and complicated script, and the colonizing Spaniards did an outstanding job of eradicating it.  For example, they managed to destroy all but four books written in Maya script.

Fortunately, many carved examples of Mayan writing were not so easily destroyed.  In addition, notes taken by the very man who ordered the destruction of the books were a hugely important clue in relearning how to read Maya script.

Maya script was logosyllabic, with many logographic symbols in addition to a full syllabary.  To complicate matters, some glyphs could represent either a logogram or a syllable.  To make it even harder to decipher, each syllable could have several different glyphs for it (sort of like how the first sound in “fan” is spelled “f” in some words and “ph” in other words).

In addition, the syllabary was made up of consonant-vowel (CV) syllables like “ka” and “lo”, yet the language had syllables that were actually consonant-vowel-consonant (CVC), like the English “cat”.  They had relatively complicated rules for how to combine two CVs to make a CVC (i.e. how to show that a vowel should not be pronounced).

That’s not all: for multi-syllabic words, they smashed glyphs for syllables together into one blocky character.

To make it even harder to decipher, the reading order wasn’t left-to-right, or right-to-left, or top-to-bottom, or even bottom-to-top, but rather in a vertical sawtooth pattern!  You start at the top left, then go right one, down and left, right, down and left, right, etc. until you get to the bottom.  You then go to the top of the third column and start over.

But wait, it gets worse: there were several different languages that were all written in Maya script.

It’s understandable why nobody made any headway in deciphering the script until 1952.  It’s amazing it got deciphered at all.

Links: Wikipedia, Omniglot, and an excellent article at Ancient Scripts.  Glyph drawn by Zykasaa.

Posted in Logograms, Rating: 5 "Whoa!!", Syllabaries | Leave a comment

Epi-Olmec — 900BC? ~450 BC?, Mexico

Epi-Olmec

How many independently-developed writing systems have there been?  One hundred years ago, educated people would have told you without hesitation, “three: Sumerian or Egyptian in the Middle East, Chinese in China, and Mayan in Central America”.  Very well-educated people would say four, including Indus script.

Unfortunately, archaeologists keep messing up the picture.  In addition to finding Vinca and eight Chinese maybe-they-are-writing-maybe-they-aren’t symbols afterwards, in 1986 people fished a big slab of rock out of a stream that with symbols on it that is clearly writing.  This is recent enough that the academics haven’t settled on a name, but “Epi-Olmec”, “Isthmian”, and “La Mojarra” script are all names used for this script.

They have also found fragments of script which appears to be Zapotec, which might predate Epi-Olmec by a bit, but that dating is disputed.

Then in the early 1990s, some road builders in an area that had been inhabited by Olmecs stumbled over a smaller stone slab, now called the Cascajal block, with symbols on it.  The block has been dated to around 900 BC, and sure looks like writing to me: it has 28 unique symbols in a string that is 62 characters long.  The block is rather large and the characters are on a side of the block that has been deliberately flattened/smoothed.

Unfortunately, there aren’t enough characters or an obvious enough relationship to other scripts to be able to decipher the Cascajal glyphs.  Thus, Cascajal will have to stay in the maybe-writing-maybe-not category.

Epi-Olmec, on the other hand, is clearly writing, and is clearly related to the other mesoamerican writing systems.  In addition to artistic similarities, all of the mesoamerican writing systems use the same notations for numbers and for dates.  So regardless of whether or not the Cascajal block was writing or not, Mayan was not the first mesoamerican writing system.  Sorry, Mayan.

However, the good news for hemispheral pride is that writing was clearly independently invented in the New World.  While one can imagine scenarios where at least the idea of writing travelled from the Middle East to East and South Asia, camel caravans are highly unlikely to have reached Mexico.

Links: Wikipedia, Ancient Scripts.  I also highly recommend the Ancient Scripts article on mesoamerican writing systems.

Posted in Logograms, Rating: 4 "Huh, interesting!", Syllabaries, Undeciphered | Leave a comment

Indus script — 2600 BC, Pakistan

Indus script

Like the Sumerians and the Egyptians, the Harappans inhabited a fertile river valley (this one in Pakistan) in around 3300 BC.  Like in the Middle East, the Harappans developed a sophisticated civilization.  Like in the Middle East, the Harappans put symbols on many things, though they took 500 years longer than the Middle Easterners.  Unlike in the Middle East, it is not clear if those symbols were writing or not.

Like Vinca, the collections of symbols in “Indus script” or “Harappan script” are short.  The longest “phrase” is only 17 characters long, and the average string is only 5 characters long.  If it is a writing system, it is undoubtedly logographic due to the number of distinct glyphs — 417 found so far.

Most of the glyphs have been found on what appear to be seals.  It might be that the symbols are arbitrary symbols that people used to claim ownership or creation of a work, in much the same way that Facebook profile photos claim ownership and creation of status messages.

Indus appears to not have been influenced by other scripts, but it is quite possible that they got the idea of writing (or making symbols, if it is not writing) from elsewhere.  They had been writing in the Middle East for about five hundred years by the time they started making their symbols, and five hundred years is a long time.  Five hundred years is only slightly less than the time between Gutenberg’s printing press and moon shots, and there is evidence of indirect trade routes between Pakistan and Egypt in the Harappan time period.  (Essentially all lapis lazuli comes from Afghanistan, and there are lapis lazuli artifacts in both Egyptian and Indus archeological sites of that time period.)

I suspect that no messages were sent from Cairo to Karachi.  Papyrus was expensive (and could be re-used); clay was heavy.  Even more important, what’s the use of sending a letter to someone who couldn’t read in a language they didn’t speak?

It is possible that a trader brought along writing supplies for his own personal record-keeping.  While he probably didn’t give receipts to his customers, they might have seen what he was doing and inquired.  Or perhaps a trader commented about how the half-brother of his wife’s sister’s cousin’s son had a job converting sounds into marks. From that, perhaps some Harappan guy decided to try it himself.

The world will probably never know.

Links: Wikipedia, Ancientscripts, Unicode proposal

Posted in first in its area, Logograms, probably developed by illiterate(s), Rating: 2 "Not all that interesting", Undeciphered | 3 Comments