Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
The Evolution of Writing (utexas.edu)
42 points by scott_s on July 29, 2018 | hide | past | favorite | 11 comments


while this tries to present a nice tidy picture and dump all progress and technological development to somewhere near the holy land of certain peoples even a wikipedia lookup would reveal serious flaws

indus script dates back to nearly the same as sumerian and egyptian around 3200 bc with some potsherd markings dated further back to 3500 bc

if indus script was related to sumerian it wouldnt be undeciphered till now

the first evidence of brahmi is from lanka which is probably as far from levant as china

it is funny that the intrusive nature of the alphabet amidst the syllabaries is not remarked upon but just accepted especially since the other intrusive party of the time 1500 bc are indo europeans and most likely indo aryans in the mitanni and kassite eras

the near complete further non development of the alphabet to better fit the languages indicating the borrowed nature of it

as opposed to say brahmi or chinese

no mention of the tale that the phoenicians gave greeks the alphabet or dicussion of their origin which is still unknown herodotus or someone said they were of the african coast i believe ie not native but everyone pretends this is just fake

the strained efforts to relate everything back to egyptian heiroglyphs because europeans want that heritage too


The Indus Valley Script is a really tough thing to place because there's so little evidence to go on, and India has a particularly nasty problem with nationalistic cultural appropriation.

Deciphering unknown scripts is insanely difficult. In fact, all of the big scripts started with use of a bilingual inscription: Egyptian via the Rosetta Stone, cuneiform via a bilingual Akkadian/Persian inscription, and Mayan via a (botched) bilingual alphabet. There's nothing of the sort for Indus Valley Script, and the texts themselves are too short for even cryptology to be of much use. It's not even clear that Indus Valley Script is true writing as opposed to proto-writing, i.e., if it was capable of encoding full utterances within the language spoken of the time.

The link between the Phoenician abjad and the Greek alphabet is quite well-attested, not least of which is the order, shape, and phonetic correspondence between the letters is damn near perfect. Compare the correspondence of Phoenician and Greek versus Linear B and Greek, say. It takes pretty big blinders to say the Greek alphabet couldn't have come from Phoenicians.


I'd bet that the most useful of these practices were also highly-mobile and cross-fertilized by travellers. (It only takes one.) We now know that humanity's been moving around for a -very- long time. 4500 years is relatively short.

"Merv is the oldest and best-preserved of the oasis-cities along the Silk Route in Central Asia. The remains in this vast oasis span 4,000 years of human history." - http://whc.unesco.org/en/list/886


The article mentions how writing evolved through several stages: (a) clay tokens, (b) replaced with 2-D signs, (c) evolving into phonetic signs, (d) settling into two dozen letters each representing a single sound.

Over the past 50 years, another stage has occurred via the keyboard. They allow the 95 Ascii symbols to be entered easily, and control keys enable changes to the fonts, sizes, colors, or even tokens displayed. The pinnacle of this change is the IME's for Chinese and Japanese, enabling people to type in spoken sounds (pinyin or kana) with a selector, but allow Hanzi or Kanji to be displayed. Quite useful for tiny smartphone screens.


The point I expected this page to say that it doesn't really mention is that there's a fairly clear progression in the development of writing. Early writing systems all seem to start with some sort of pictography. This runs into the problem that not every word can be represented with a clear pictogram, and the solution ends up being some sort of phonetic punning mechanism (i.e., the rebus principle). Even Chinese hanzi evidences these sort of mechanism, as the vast majority of its characters are based on combining a phonetic component with a semantic component.

After you start phonetic punning, most writing systems drop the pictographic portion and become solely based on phonetic constituents (Chinese being the major exception, although Japanese did evolve the next step into kana). These constituents are usually roughly based on a consonant-vowel pair and writing systems in this phase are called syllabries, but they're often not really full syllables, and there are tricks for representing more complex syllables (for example, Mayan uses the principle that the syllable pattern CVC is represented with CV-CV, where the vowel is the same in both characters.).

Now, in Phoenician or thereabouts, an interesting turn happened. Because of the nature of Semitic languages, the syllabry form dropped all of the vowels and just represented the consonants (what's now termed an abjad). The Indo-European speakers (principally Etruscan and Greek) who borrowed from this alphabet repurposed some of the letters which corresponded to missing consonants (such as the glottal stop in 'aleph) to represent vowels instead, as vowels are rather more important in those languages. Other languages that borrowed from abjads indicated vowels with systematic vowel marks, such as Devanagari or Canadian Aboriginal Syllabics; these are known as abugidas. Virtually every alphabet, abjad, and abugida known to exist is believed to derive originally from this first "wrong" turn when developing syllabaries.

An interesting point when discussing the history of writing systems is the development of the Cherokee script. The Cherokee script was invented by Sequoyah, who developed it entirely on the basis that he knew that the white settlers had a writing system that could be reliably read and a copy of the Bible but without any other knowledge of writing. He originally attempted to develop a mapping of word => character, but realized that the mapping is too complex, especially since the text he had got by with much fewer characters. Instead, he chose a syllabary, borrowing some of the letter forms from Latin script but not their phonetic mappings (which is why ᏣᎳᎩ is actually pronounced "tsalagi").


With regard to vowel-centric versus consonant-centric languages, English tends to be consonant-centric, and that's partly why our spelling is difficult. For example, "flutter" could be spelled "fluttr": we don't need the 2nd vowel and thus wouldn't have to guess whether it's spelled "fluttur", "fluttor", etc. English could use a refactoring. Examples: http://wiki.c2.com/?RefactorEnglish


What you're referring to is essentially schwa-reduction. Unstressed vowels in English get reduced to a neutral vowel sound (represented by schwa, an upside-down e), and different letters are used to represent that vowel depending more or less on linguistic history.

But vowels do matter in English. Consider the following words that differ only in the vowel they use: bat, bait, bit, bot, bite, boat, boot, butt, bet, beat, bout. Furthermore, all of those words refer to very different things. Semitic languages are based on a triconsonantal scheme, where three consonants define a root, and the choice of vowels implies inflection. Dropping the vowels in Semitic languages does not impair meaning all that much.


I didn't intend to imply all vowels can be dropped or wild-carded in English. As far as schwa, if we do refactor English, we should probably stick to existing "normal" letters rather than introduce symbols such as an explicit schwa. I agree this risks either ambiguity, or some potentially non-obvious rules to remember.

As it is, I feel sorry for non-native English learners; English's spelling is a mess. The only saving grace is that being semi-phonetic means one can often make a decent guess.


Re: With state formation, new regulations required that the names of the individuals who generated or received registered merchandise were entered on the tablets. The personal names were transcribed by the mean of logograms...pictures of words with a sound close to that desired (for example in English the name Neil could be written with a sign showing bent knees 'kneel').

Seems like a chicken-or-egg conundrum here. How can you create a regulation for something that nobody knows how to do? (No jokes about a dysfunctional Congress, please.) Somebody must have realized the logogram technique BEFORE suggesting or making an edict.


> clay tokens representing units of goods were used for accounting (8000–3500 BC). ... The token system had little in common with spoken language except that, like a word, a token stood for one concept. Unlike speech, tokens were restricted to one type of information only, namely, real goods. Unlike spoken language, the token system made no use of syntax.

That is, for 4.5ky Sumerians had a form of abstracted accounting record very unlike a writing system proper, and it was used on a large territory where people likely spoke different languages / dialects.

(In a sense, it was an early era of digital communication.)


These symbols are non-unique identifiers for physical objects. More like a digital trading floor: price is syntax.




Consider applying for YC's Winter 2026 batch! Applications are open till Nov 10

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: