Saurashtra — <1880AD, India

Saurashtra "pha"

The Saurashtra people have a very unsettled past.  They lived in Gujarat, but then Gazni Mohammed invaded around 1000 AD and the Saurashtras took off for Devagiri, farther south.  They stayed there for two centuries, but then that empire collapsed and they were on the move again, even farther south to another kingdom.  They were traditionally weavers, and frequently were royal weavers, enjoying the patronage of the local king.

Perhaps their being a perpetual minority for the past thousand years (and living by the grace of the local king for hundreds of years of that) contributed to them not sticking fiercely to their own language and/or script.  Indeed, my sources don’t cite any material in the Saurashtra script before the late 1800s, when Medhavi Sri T.M. Rama Rai set up a press in Saurashtra script.  In 1920 AD, however, a resolution was passed declaring that Saurashtra should be written with the Devanagari script.  In practice, however, the language is now mostly written in a modified version of Tamil script.

The Saurashtra script is a pretty generic Brahmi-derived script, with one exception: it has a diacritic for consonants, turning e.g. “ma” and “ra” into “mha” and “rha”.

Links: Wikipedia, Omniglot, Unicode proposal

Posted in Abugida, Rating: 2 "Not all that interesting" | Leave a comment

Sinhala — 700 AD, Sri Lanka

Sinhala "ma" (Elu)

Sinhala has my vote for the prettiest script on the planet.

Sinhala has two sets of characters: the Elu set represents all the spoken Sinhala phonemes, while the Mixed set represents characters for loan words, mostly from Sanskrit and Pali (another liturgical language). Almost all of the Mixed set can be represented or approximated with the Elu set, but using the Mixed set marks them as loan words.  This is similar to how Hiragana is used for native words and Katakana is used for loan words in Japanese.

One complication is that there are some characters in the Mixed set which have the same sound as a character in the Elu set.  There are also two characters in the Elu set for sounds that are no longer pronounced distinctly from other sounds.  This means that it is quite easy to misspell words — something that is not easy in languages with a closer one-to-one relationship between sounds and characters.

Aside from having two sets of characters, Sinhala is a pretty vanilla Indic abugida, with a virama (called a hal kirima in Sinhala) and consonant conjuncts to kill the vowel and diacritics to change the vowel, plus a  diacritics to show nasalization.  Unlike in most Indic writing systems, the virama does change its shape depending upon the shape of the glyph it modifies.

Links: Wikipedia, Omniglot, Ancient Scripts

Posted in Abugida, Rating: 4 "Huh, interesting!" | 1 Comment

Malayalam — ~830 AD? ~1600 AD, India

Malayalam "nya"

The Malayalam script is used in Kerala, the southernmost province on India’s western shore.  Kerala has been a destination for trade and travellers for thousands of years; Kerala is the easternmost point on the only surviving map of the Roman postal service, for example.

The script has had a long evolution, which makes assigning one date difficult. However, the evolution of the Malayalam script is punctuated with three significant events.

Background: The Malayalam spoken language was originally written in Vatteluttu.  Like in the Tamil-speaking areas, most people moved to a hybrid of Grantha for for Sanskrit (in the Indo-European or Aryan language family), and Vatteluttu for the local language (in the Dravidian language family).  This hybrid was commonly called Arya-eluttu.

First significant event: Around 1600 AD, the poet Thunchath Ezhuthachan wrote his poetry in the Malayalam language, but based on Sanskrit themes, using Arya-eluttu.  When he needed a sound that was not in Arya-eluttu, he would use the Vatteluttu glyph.  His poetry was wildly popular, and served to popularize the Arya-eluttu hybrid.  It was not a perfect fit for the Malayalam language, as there were some distinctions that the script did not capture.

Second significant event: In around 1850 AD, a German missionary named Hermann Gundert published the first Malayalam-English dictionary, a Malayalam grammar, and a Malayalam translation of the Bible.  Gundert also added a few new vowel diacritics to disambiguate the vowels, and they stuck.  (This is the first time that I noticed that a foreigner made a change to a writing system that stuck!)

Third significant event: In 1969-1971 AD, the government reformed the script.  In particular, they separated ligatures into independent glyphs.  For example, diacritics which modified the vowel were subsequently written with two separate (non-conjoined) glyphs, the modifier to the right of the modified glyph.  (Presumably this was to make printing with movable type easier.)  They also eliminated some (but not all) consonant conjuncts.

Malayalam has a virama (called chandrakkala in Malayalam) to kill vowels, but also has six special consonants (called chillus) which never have a vowel following them.  The chandrakkala is written above the consonant glyph, not to its right like the vowel diacritics; I guess there were limits to how much they were willing to reform the language in 1970 AD.

Malayalam, like Tamil, was written on palm leaves, which enforced a very curvaceous style.  Like Tamil, it is a very pretty script.

Links: Technology Development for Indian Languages, Wikipedia, Omniglot, Ancient Scripts

Posted in Abugida, Evolved slowly from parent, Rating: 4 "Huh, interesting!" | 1 Comment

Tamil — 700 AD, India

Tamil "i"

Tamil is descended in part from Grantha and in part from Vatteluttu.  Grantha and Vatteluttu are relatively closely related, and the two were used together for a time, so it was relatively easy to smear the two together.

Tamil has changed significantly as technology has changed.  When Tamil was written on palm leaves, there was a tendency not to write the virama (which in the Tamil script was a dot above the character) because of the tendency of palm leaves to tear.  Once Tamil started being printed, people started using the virama more regularly.

The Tamil spoken language has far fewer consonant clusters than most Indic languages, which means that there are fewer vowels that need to be killed, which means that there is less need for the virama (called pulli in Tamil).  It also means that there are fewer consonant conjuncts in the Tamil script than in most Indic scripts.  Almost always, the virama is used to kill the vowel.  There are only a few consonant conjuncts.

Tamil was once my choice for “prettiest writing system”.  I have a new favorite, but I still find Tamil to be a very pretty script.

Links: Technology Development for Indian Languages, Wikipedia, Omniglot, Ancient Scripts

Posted in Abugida, Rating: 3 "I did not know that", technology influenced | 1 Comment

Grantha — 500 AD, India

Grantha "a"

Around 500 AD, the Tamil people of Southern Inda started using Grantha, a slightly different form of the Brahmi alphabet, to write Sanskrit (the language of sacred Hindu texts), while still continuing to use Vatteluttu to write representations of the Tamil spoken language. They needed to use a different script for Sanskrit because Vatteluttu, like Tamil-Brahmi, did not have characters for sounds that were not in Tamil.

As Tamil borrowed words from Sanskrit, they would preserve the origin of the spoken language in the script that they used when they wrote it.  For example, if a word’s stem came from Sanskrit but the ending was Tamil, they would use Grantha to write the stem and Vatteluttu to write the ending.  This was very similar to how the Japanese would use Chinese characters for the stems and Hiragana for the endings, but in the Grantha/Vatteluttu case, both writing scripts were phonetic.

Eventually, the two scripts merged, or perhaps Grantha “won”.  It isn’t surprising that Grantha “won” as Vatteluttu was not ideal for Tamil even discounting loan words, as Vatteluttu was missing a way to represent terminal o, e, or u.

While Sanskrit is now usually written in Devanagari, Tamil speakers will sometimes still use Grantha, especially for ceremonial purposes like wedding invitations.

Links: Wikipedia, Omniglot, Ancient Scripts, C. Radhakrishnan, Unicode proposal

Posted in Abugida, Evolved slowly from parent, now ceremonial, previous script didn't quite work, Rating: 4 "Huh, interesting!" | 4 Comments

Vatteluttu — 500 AD, India

Vatteluttu "na"

In southern India, the Tamil people first used the Tamil-Brahmi script, and then over time, developed the Vatteluttu (also called Vattezhuttu) script.  From what I can tell, this was an evolutionary change and not a sudden invention.

Vatteluttu was used from about 500 AD to about 1500 AD, mostly used to write Tamil and sometimes Malayalam.  Tamil-Brahmi was also widely used to write Tamil and sometimes Malayalam.  This multiplicity of scripts was able to happen because there was not a centralized government to impose a practice — either by edict or by custom — of using one or the other script.

Eventually, two different scripts displaced both Vatteluttu and Tamil-Brahmi, one for Tamil and one for Malayalam.

There were two regional variants of Vatteluttu: Kolezhuthu and Malayanma.  There was almost no difference between the scripts, except that Kolezhuthu script didn’t allow endings in a, e, or u.

Links: Wikipedia, Ancient Scripts, C. Radhakrishnan

Posted in Abugida, Rating: 1 "Dull, only here for completeness" | 3 Comments

Tamil-Brahmi — 400 BC?, India

Early Tamil-Brahmi "ma"

The Brahmi script had two major branches: a northern branch (frequently called Ashokan Brahmi, after the king who put up the Edicts of Ashoka) and a southern (frequenly called Tamil-Brahmi).  So far, all the Brahmi-derived scripts that I have described were from the Northern branch, with the possible exception of Oriya.

Tamil-Brahmi is a bit different from Ashokan Brahmi.  It has characters for sounds not found in Ashokan Brahmi; it went through three different phases trying to deal with terminal consonants.

The first attempt was alphabetic: the inherent vowel didn’t exist, and if you wanted a vowel, you needed to put a diacritic.  The second thing they tried was to make the reader figure out if there was a vowel there or not.  The third thing — and what they stuck with — was to use a pulli, which is just a different name for virama.

There is evidence that literacy in Tamil-Brahmi was higher than Ashokan  Brahmi.  A strong bardic/poetic tradition, local rule (in the local language, instead of a foreign elite’s language imposed upon them), and extensive seafaring trade from their long coastline all contributed to making literacy more common.

Links: Wikipedia Tamil-Brahmi, Wikipedia Tamil history, Ponniyin Selvan I, II, and III, Saigan

Update: Ooops, I said that all the previous posts were Southern Brahmi scripts, while I meant to say Northern.  Fixed.

Update2: Found out more about how Tamil-Brahmi handled terminal consonants.

Posted in Abugida, first in its area, Rating: 4 "Huh, interesting!" | 2 Comments

Lepcha — ~1700 AD, India

Lepcha "flan"

The Lepcha script was developed either by prince Phyagdor Namgyal or by the scholar Thikúng Men Salóng sometime around 1700 AD.  Although it was pretty clearly derived from Tibetan, which is written left-to-right, early Lepcha was written vertically, probably from Chinese influence.

To write the language vertically, they rotated everything 90 degrees. After not very much time, they switched to left-to-right like Tibetan, but declined to change the orientation of the glyphs.  The the Lepcha glyphs look very similar to Tibetan glyphs, but rotated 90 degrees.

Note that the Lepcha glyphs are rotated 90 degrees counter-clockwise, while you would think that going from left-to-right to top-to-bottom, they would rotate them by 90 degrees clockwise.  I have no explanation for that: I don’t know if they wrote bottom-to-top or did some strange gymnastics.

In Lepcha, there are some syllables that end with a consonant, meaning a vowel needs to be killed.  Unlike Tibetan, which kills vowels by using consonant conjuncts, Lepcha uses diacritics to mark final consonants.  (My sources say that this is due to the funky double-rotation that happened, but don’t explain how.) The glyph adorning this post is made up of two pieces: the “fla” on the bottom, and the twirly little “n” diacritic on the top.

Links: Wikipedia, Omniglot, Ancient Scripts,, Promotora Española de Lingüística (in Spanish, with nice comparison table)

Posted in Abugida, Rating: 4 "Huh, interesting!" | Leave a comment

Tani Lipi — 2001 AD, India

A gentleman named Tony Koyu designed the Tani Lipi script in 2001 for the Tani group of languages in the far northeast of India.  It is explicitly designed to unify the many Tani tribes.  There is pretty fierce debate over whether this writing system is a good thing or a bad thing.

There is very little information available about Tani Lipi.  I can’t even figure out if it is a syllabic alphabet or an abugida, alas.

Links: Unicode introduction

Posted in inventor known, National pride, Rating: 1 "Dull, only here for completeness" | 3 Comments

Varang Chiti — 1963 AD?, India

Varang Chiti "ha"

The Ho language is a minority language in northeastern India, but in India, minority languages can have a large number of speakers: there are over a million Ho speakers.

At some point in the 20th century, a gentleman named Lako Bodra developed or discovered the Varang Chiti writing system.  Bodra asserted that the Varang Chiti script was invented in the 13th century by Dhawan Turi, and that it was revealed to Bodra in a dream.  He modernized it, and released it to the world.  Note that this is not an uncommon story: I will describe more cases in later posts where an ancient script was revealed to someone in modern times who popularized it.

Like Santali, which I mentioned in yesterday’s post on Ol Chiki, Ho is an Austro-Asiatic language, somewhat related to Santali.

Varang Chiti is an abugida, but a somewhat unusual one.  Most Brahmi-derived writing systems have a default vowel attached to consonants that can be changed with diacritics or killed with a virama.  Varang Chiti is similar, but while most Brahmi-derived writing systems have the same default vowel for all of their consonant+vowel syllables, Varang Chiti has a different vowel for different syllables.

Links: Wikipedia, Omniglot, Swarthmore, LisIndia

Posted in Abugida, inventor known, National pride, Rating: 4 "Huh, interesting!", revealed in a dream | Leave a comment